diff --git a/CMakeLists.txt b/CMakeLists.txt
index 8b94016..7faddc1 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1,5 +1,6 @@
 cmake_minimum_required(VERSION 3.7)
 
+
 option(VERSION_STRING "Project version string" "0.0.0")
 
 project(antkeeper VERSION ${VERSION_STRING} LANGUAGES CXX)
diff --git a/src/color/aces.hpp b/src/color/aces.hpp
new file mode 100644
index 0000000..a127ca1
--- /dev/null
+++ b/src/color/aces.hpp
@@ -0,0 +1,91 @@
+/*
+ * Copyright (C) 2021  Christopher J. Howard
+ *
+ * This file is part of Antkeeper source code.
+ *
+ * Antkeeper source code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * Antkeeper source code is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Antkeeper source code.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef ANTKEEPER_COLOR_ACES_HPP
+#define ANTKEEPER_COLOR_ACES_HPP
+
+#include "color/rgb.hpp"
+#include "math/math.hpp"
+
+namespace color {
+
+/// ACES color spaces.
+namespace aces {
+
+/// CIE xy chromaticity coordinates of the ACES white point (~D60).
+template <class T>
+constexpr math::vector2<T> white_point = {T{0.32168}, {0.33767}};
+
+/// ACES AP0 color space.
+template <class T>
+constexpr rgb::color_space<T> ap0
+(
+	{T{0.7347}, T{ 0.2653}},
+	{T{0.0000}, T{ 1.0000}},
+	{T{0.0001}, T{-0.0770}},
+	aces::white_point<T>,
+	nullptr,
+	nullptr
+);
+
+/// ACES AP1 color space.
+template <class T>
+constexpr rgb::color_space<T> ap1
+(
+	{T{0.713}, T{0.293}},
+	{T{0.165}, T{0.830}},
+	{T{0.128}, T{0.044}},
+	aces::white_point<T>,
+	nullptr,
+	nullptr
+);
+
+/**
+ * Constructs a saturation adjustment matrix.
+ *
+ * @param s Saturation adjustment factor.
+ * @param to_y Color space to CIE XYZ luminance vector.
+ *
+ * @return Saturation adjustment matrix.
+ */
+template <class T>
+constexpr math::matrix<T, 3, 3> adjust_saturation(T s, const math::vector3<T>& to_y)
+{
+	const math::vector3<T> v = to_y * (T{1} - s);
+	return math::matrix<T, 3, 3>
+	{
+		v[0] + s, v[0],     v[0],
+		v[1],     v[1] + s, v[1],
+		v[2],     v[2],     v[2] + s
+	};
+}
+
+/// ACES AP1 RRT saturation adjustment matrix.
+template <class T>
+constexpr math::matrix<T, 3, 3> ap1_rrt_sat = aces::adjust_saturation(T{0.96}, ap1<T>.to_y);
+
+/// ACES AP1 ODT saturation adjustment matrix.
+template <class T>
+constexpr math::matrix<T, 3, 3> ap1_odt_sat = aces::adjust_saturation(T{0.93}, ap1<T>.to_y);
+
+} // namespace aces
+
+} // namespace color
+
+#endif // ANTKEEPER_COLOR_ACES_HPP
diff --git a/src/color/acescg.hpp b/src/color/acescg.hpp
deleted file mode 100644
index 1b50f34..0000000
--- a/src/color/acescg.hpp
+++ /dev/null
@@ -1,99 +0,0 @@
-/*
- * Copyright (C) 2021  Christopher J. Howard
- *
- * This file is part of Antkeeper source code.
- *
- * Antkeeper source code is free software: you can redistribute it and/or modify
- * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation, either version 3 of the License, or
- * (at your option) any later version.
- *
- * Antkeeper source code is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
- * GNU General Public License for more details.
- *
- * You should have received a copy of the GNU General Public License
- * along with Antkeeper source code.  If not, see <http://www.gnu.org/licenses/>.
- */
-
-#ifndef ANTKEEPER_COLOR_ACESCG_HPP
-#define ANTKEEPER_COLOR_ACESCG_HPP
-
-#include "math/math.hpp"
-
-namespace color {
-
-/// Functions which operate in the ACEScg colorspace.
-namespace acescg {
-
-/// CIE chromaticity coordinates of the ACEScg white point (~D60).
-template <class T>
-constexpr math::vector2<T> whitepoint = {0.32168, 0.33767};
-
-/**
- * Calculates the luminance of an ACEScg color.
- *
- * @param x ACEScg color.
- * @return return Luminance of @p x.
- */
-template <class T>
-T luminance(const math::vector3<T>& x);
-
-/**
- * Transforms an ACEScg color into the linear sRGB colorspace using the Bradford chromatic adaption transform.
- *
- * @param x ACEScg color.
- * @return return Linear sRGB color.
- *
- * @see https://www.colour-science.org/apps/
- */
-template <class T>
-math::vector3<T> to_srgb(const math::vector3<T>& x);
-
-/**
- * Transforms an ACEScg color into the CIE XYZ colorspace.
- *
- * @param x ACEScg color.
- * @return return CIE XYZ color.
- */
-template <class T>
-math::vector3<T> to_xyz(const math::vector3<T>& x);
-
-template <class T>
-T luminance(const math::vector3<T>& x)
-{
-	static const math::vector3<T> luma = {0.272228716780914, 0.674081765811148, 0.053689517407937};
-	return math::dot(x, luma);
-}
-
-template <class T>
-math::vector3<T> to_srgb(const math::vector3<T>& x)
-{
-	static const math::matrix3<T> acescg_to_srgb
-	{{
-		{ 1.704858676289160, -0.130076824208823, -0.023964072927574},
-		{-0.621716021885330,  1.140735774822504, -0.128975508299318},
-		{-0.083299371729058, -0.010559801677511,  1.153014018916862}
-	}};
-	
-	return acescg_to_srgb * x;
-}
-
-template <class T>
-math::vector3<T> to_xyz(const math::vector3<T>& x)
-{
-	static const math::matrix3<T> acescg_to_xyz
-	{{
-		{0.662454181108505, 0.272228716780914, -0.005574649490394},
-		{0.134004206456433, 0.674081765811148,  0.004060733528983},
-		{0.156187687004908, 0.053689517407937,  1.010339100312997}
-	}};
-	
-	return acescg_to_xyz * x;
-}
-
-} // namespace acescg
-} // namespace color
-
-#endif // ANTKEEPER_COLOR_ACESCG_HPP
diff --git a/src/color/cat.hpp b/src/color/cat.hpp
new file mode 100644
index 0000000..3b56d03
--- /dev/null
+++ b/src/color/cat.hpp
@@ -0,0 +1,106 @@
+/*
+ * Copyright (C) 2021  Christopher J. Howard
+ *
+ * This file is part of Antkeeper source code.
+ *
+ * Antkeeper source code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * Antkeeper source code is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Antkeeper source code.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef ANTKEEPER_COLOR_CAT_HPP
+#define ANTKEEPER_COLOR_CAT_HPP
+
+#include "math/math.hpp"
+
+namespace color {
+
+/// Chromatic adaption transforms (CAT).
+namespace cat {
+
+/**
+ * Bradford cone response matrix.
+ *
+ * @see Specification ICC.1:2010 (Profile version 4.3.0.0). Image technology colour management — Architecture, profile format, and data structure, Annex E.3, pp. 102.
+ * @see http://www.brucelindbloom.com/index.html?Eqn_ChromAdapt.html
+ */
+template <class T>
+constexpr math::matrix<T, 3, 3> bradford =
+{
+	 0.8951, -0.7502,  0.0389,
+	 0.2664,  1.7135, -0.0685,
+	-0.1614,  0.0367,  1.0296
+};
+
+/**
+ * von Kries cone response matrix.
+ *
+ * @see http://www.brucelindbloom.com/index.html?Eqn_ChromAdapt.html
+ */
+template <class T>
+constexpr math::matrix<T, 3, 3> von_kries =
+{
+	 0.40024, -0.22630, 0.00000
+	 0.70760,  1.16532, 0.00000
+	-0.08081,  0.04570, 0.91822
+};
+
+/**
+ * XYZ scaling cone response matrix.
+ *
+ * @see http://www.brucelindbloom.com/index.html?Eqn_ChromAdapt.html
+ */
+template <class T>
+constexpr math::matrix<T, 3, 3> xyz_scaling =
+{
+	T{1}, T{0}. T{0},
+	T{0}, T{1}. T{0},
+	T{0}, T{0}. T{1}
+};
+
+/**
+ * Constructs a chromatic adaptation transform (CAT) matrix.
+ *
+ * @param w0 CIE xy chromaticity coordinates of the source illuminant.
+ * @param w1 CIE xy chromaticity coordinates of the destination illuminant.
+ * @param cone_response Cone response matrix.
+ *
+ * @return CAT matrix.
+ *
+ * @see Specification ICC.1:2010 (Profile version 4.3.0.0). Image technology colour management — Architecture, profile format, and data structure, Annex E.3, pp. 102.
+ * @see http://www.brucelindbloom.com/index.html?Eqn_ChromAdapt.html
+ */
+template <class T>
+constexpr math::matrix<T, 3, 3> matrix(const math::vector2<T>& w0, const math::vector2<T>& w1, const math::matrix<T, 3, 3>& cone_response = bradford<T>)
+{
+	// Convert CIE xy chromaticity coordinates to CIE XYZ colors
+	const math::vector3<T> w0_xyz = {w0[0] / w0[1], T{1}, (T{1} - w0[0] - w0[1]) / w0[1]};
+	const math::vector3<T> w1_xyz = {w1[0] / w1[1], T{1}, (T{1} - w1[0] - w1[1]) / w1[1]};
+	
+	// Calculate cone response of CIE XYZ colors
+	const math::vector3<T> w0_cone_response = cone_response * w0_xyz;
+	const math::vector3<T> w1_cone_response = cone_response * w1_xyz;
+	
+	const math::matrix<T, 3, 3> scale =
+	{
+		w1_cone_response[0] / w0_cone_response[0], T{0}, T{0},
+		T{0}, w1_cone_response[1] / w0_cone_response[1], T{0},
+		T{0}, T{0}, w1_cone_response[2] / w0_cone_response[2],
+	};
+	
+	return math::inverse(cone_response) * scale * cone_response;
+}
+
+} // namespace cat
+} // namespace color
+
+#endif // ANTKEEPER_COLOR_CAT_HPP
diff --git a/src/color/cct.hpp b/src/color/cct.hpp
index c9d676c..e9f3eb5 100644
--- a/src/color/cct.hpp
+++ b/src/color/cct.hpp
@@ -38,7 +38,15 @@ namespace cct {
  * @see Krystek, M. (1985), An algorithm to calculate correlated colour temperature. Color Res. Appl., 10: 38-40.
  */
 template <class T>
-math::vector2<T> to_ucs(T t);
+math::vector2<T> to_ucs(T t)
+{
+	const T tt = t * t;
+	return math::vector2<T>
+	{
+		(T{0.860117757} + T{1.54118254e-4} * t + T{1.28641212e-7} * tt) / (T{1} + T{8.42420235e-4} * t + T{7.08145163e-7} * tt),
+		(T{0.317398726} + T{4.22806245e-5} * t + T{4.20481691e-8} * tt) / (T{1} - T{2.89741816e-5} * t + T{1.61456053e-7} * tt)
+	};
+}
 
 /**
  * Calculates CIE xyY colorspace chromaticity coordinates given a correlated color temperature using Krystek's algorithm.
@@ -47,7 +55,10 @@ math::vector2<T> to_ucs(T t);
  * @return CIE xyY color with `Y = 1`.
  */
 template <class T>
-math::vector3<T> to_xyy(T t);
+math::vector3<T> to_xyy(T t)
+{
+	return ucs::to_xyy(to_ucs(t), T{1});
+}
 
 /**
  * Calculates CIE XYZ colorspace chromaticity coordinates given a correlated color temperature using Krystek's algorithm.
@@ -55,27 +66,6 @@ math::vector3<T> to_xyy(T t);
  * @param t Correlated color temperature, in Kelvin.
  * @return CIE XYZ color with `Y = 1`.
  */
-template <class T>
-math::vector3<T> to_xyz(T t);
-
-template <class T>
-math::vector2<T> to_ucs(T t)
-{
-	// Approximate the Planckian locus in CIE 1960 UCS colorspace (Krystek's algorithm)
-	const T tt = t * t;
-	return math::vector2<T>
-		{
-			(T(0.860117757) + T(1.54118254e-4) * t + T(1.28641212e-7) * tt) / (T(1.0) + T(8.42420235e-4) * t + T(7.08145163e-7) * tt),
-			(T(0.317398726) + T(4.22806245e-5) * t + T(4.20481691e-8) * tt) / (T(1.0) - T(2.89741816e-5) * t + T(1.61456053e-7) * tt)
-		};
-}
-
-template <class T>
-math::vector3<T> to_xyy(T t)
-{
-	return ucs::to_xyy(to_ucs(t), T(1.0));
-}
-
 template <class T>
 math::vector3<T> to_xyz(T t)
 {
diff --git a/src/color/color.hpp b/src/color/color.hpp
index b5bcf10..8be5a20 100644
--- a/src/color/color.hpp
+++ b/src/color/color.hpp
@@ -23,12 +23,15 @@
 /// Color manipulation functions.
 namespace color {}
 
-#include "acescg.hpp"
+#include "aces.hpp"
 #include "cct.hpp"
 #include "index.hpp"
+#include "rgb.hpp"
 #include "srgb.hpp"
 #include "ucs.hpp"
 #include "xyy.hpp"
 #include "xyz.hpp"
+#include "cat.hpp"
+#include "illuminant.hpp"
 
 #endif // ANTKEEPER_COLOR_HPP
diff --git a/src/color/illuminant.hpp b/src/color/illuminant.hpp
new file mode 100644
index 0000000..0b134ec
--- /dev/null
+++ b/src/color/illuminant.hpp
@@ -0,0 +1,152 @@
+/*
+ * Copyright (C) 2021  Christopher J. Howard
+ *
+ * This file is part of Antkeeper source code.
+ *
+ * Antkeeper source code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * Antkeeper source code is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Antkeeper source code.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef ANTKEEPER_COLOR_ILLUMINANT_HPP
+#define ANTKEEPER_COLOR_ILLUMINANT_HPP
+
+#include "math/math.hpp"
+
+namespace color {
+
+/**
+ * CIE standard illuminants.
+ *
+ * @see https://en.wikipedia.org/wiki/Standard_illuminant
+ */
+namespace illuminant {
+
+/// CIE 1931 2 Degree Standard Observer illuminants.
+namespace deg2 {
+
+	template <class T>
+	constexpr math::vector2<T> a    = {T{0.44757}, T{0.40745}};
+	template <class T>
+	constexpr math::vector2<T> b    = {T{0.34842}, T{0.35161}};
+	template <class T>
+	constexpr math::vector2<T> c    = {T{0.31006}, T{0.31616}};
+	template <class T>
+	constexpr math::vector2<T> d50  = {T{0.34567}, T{0.35850}};
+	template <class T>
+	constexpr math::vector2<T> d55  = {T{0.33242}, T{0.34743}};
+	template <class T>
+	constexpr math::vector2<T> d65  = {T{0.31271}, T{0.32902}};
+	template <class T>
+	constexpr math::vector2<T> d75  = {T{0.29902}, T{0.31485}};
+	template <class T>
+	constexpr math::vector2<T> d93  = {T{0.28315}, T{0.29711}};
+	template <class T>
+	constexpr math::vector2<T> e    = {T{0.33333}, T{0.33333}};
+	template <class T>
+	constexpr math::vector2<T> f1   = {T{0.31310}, T{0.33727}};
+	template <class T>
+	constexpr math::vector2<T> f2   = {T{0.37208}, T{0.37529}};
+	template <class T>
+	constexpr math::vector2<T> f3   = {T{0.40910}, T{0.39430}};
+	template <class T>
+	constexpr math::vector2<T> f4   = {T{0.44018}, T{0.40329}};
+	template <class T>
+	constexpr math::vector2<T> f5   = {T{0.31379}, T{0.34531}};
+	template <class T>
+	constexpr math::vector2<T> f6   = {T{0.37790}, T{0.38835}};
+	template <class T>
+	constexpr math::vector2<T> f7   = {T{0.31292}, T{0.32933}};
+	template <class T>
+	constexpr math::vector2<T> f8   = {T{0.34588}, T{0.35875}};
+	template <class T>
+	constexpr math::vector2<T> f9   = {T{0.37417}, T{0.37281}};
+	template <class T>
+	constexpr math::vector2<T> f10  = {T{0.34609}, T{0.35986}};
+	template <class T>
+	constexpr math::vector2<T> f11  = {T{0.38052}, T{0.37713}};
+	template <class T>
+	constexpr math::vector2<T> f12  = {T{0.43695}, T{0.40441}};
+	template <class T>
+	constexpr math::vector2<T> led_b1   = {T{0.4560}, T{0.4078}};
+	template <class T>
+	constexpr math::vector2<T> led_b2   = {T{0.4357}, T{0.4012}};
+	template <class T>
+	constexpr math::vector2<T> led_b3   = {T{0.3756}, T{0.3723}};
+	template <class T>
+	constexpr math::vector2<T> led_b4   = {T{0.3422}, T{0.3502}};
+	template <class T>
+	constexpr math::vector2<T> led_b5   = {T{0.3118}, T{0.3236}};
+	template <class T>
+	constexpr math::vector2<T> led_bh1  = {T{0.4474}, T{0.4066}};
+	template <class T>
+	constexpr math::vector2<T> led_rgb1 = {T{0.4557}, T{0.4211}};
+	template <class T>
+	constexpr math::vector2<T> led_v1   = {T{0.4560}, T{0.4548}};
+	template <class T>
+	constexpr math::vector2<T> led_v2   = {T{0.3781}, T{0.3775}};
+
+} // deg2
+
+/// CIE 1964 10 Degree Standard Observer illuminants.
+namespace deg10 {
+
+	template <class T>
+	constexpr math::vector2<T> a   = {T{0.45117}, T{0.40594}};
+	template <class T>
+	constexpr math::vector2<T> b   = {T{0.34980}, T{0.35270}};
+	template <class T>
+	constexpr math::vector2<T> c   = {T{0.31039}, T{0.31905}};
+	template <class T>
+	constexpr math::vector2<T> d50 = {T{0.34773}, T{0.35952}};
+	template <class T>
+	constexpr math::vector2<T> d55 = {T{0.33411}, T{0.34877}};
+	template <class T>
+	constexpr math::vector2<T> d65 = {T{0.31382}, T{0.33100}};
+	template <class T>
+	constexpr math::vector2<T> d75 = {T{0.29968}, T{0.31740}};
+	template <class T>
+	constexpr math::vector2<T> d93 = {T{0.28327}, T{0.30043}};
+	template <class T>
+	constexpr math::vector2<T> e   = {T{0.33333}, T{0.33333}};
+	template <class T>
+	constexpr math::vector2<T> f1  = {T{0.31811}, T{0.33559}};
+	template <class T>
+	constexpr math::vector2<T> f2  = {T{0.37925}, T{0.36733}};
+	template <class T>
+	constexpr math::vector2<T> f3  = {T{0.41761}, T{0.38324}};
+	template <class T>
+	constexpr math::vector2<T> f4  = {T{0.44920}, T{0.39074}};
+	template <class T>
+	constexpr math::vector2<T> f5  = {T{0.31975}, T{0.34246}};
+	template <class T>
+	constexpr math::vector2<T> f6  = {T{0.38660}, T{0.37847}};
+	template <class T>
+	constexpr math::vector2<T> f7  = {T{0.31569}, T{0.32960}};
+	template <class T>
+	constexpr math::vector2<T> f8  = {T{0.34902}, T{0.35939}};
+	template <class T>
+	constexpr math::vector2<T> f9  = {T{0.37829}, T{0.37045}};
+	template <class T>
+	constexpr math::vector2<T> f10 = {T{0.35090}, T{0.35444}};
+	template <class T>
+	constexpr math::vector2<T> f11 = {T{0.38541}, T{0.37123}};
+	template <class T>
+	constexpr math::vector2<T> f12 = {T{0.44256}, T{0.39717}};
+	
+} // namespace deg10
+
+} // namespace illuminant
+
+} // namespace color
+
+#endif // ANTKEEPER_COLOR_ILLUMINANT_HPP
diff --git a/src/color/index.hpp b/src/color/index.hpp
index 6033035..ff7b552 100644
--- a/src/color/index.hpp
+++ b/src/color/index.hpp
@@ -33,13 +33,10 @@ namespace index {
  *
  * @see Ballesteros, F. J. (2012). "New insights into black bodies". EPL 97 (2012) 34008.
  */
-template <class T>
-T bv_to_cct(T bv);
-
 template <class T>
 T bv_to_cct(T bv)
 {
-	return T(4600.0) * (T(1.0) / (T(0.92) * bv + T(1.7)) + T(1.0) / (T(0.92) * bv + T(0.62)));
+	return T{4600} * (T{1} / (T{0.92} * bv + T{1.7}) + T{1} / (T{0.92} * bv + T{0.62}));
 }
 
 } // namespace index
diff --git a/src/color/rgb.hpp b/src/color/rgb.hpp
new file mode 100644
index 0000000..4e83a7f
--- /dev/null
+++ b/src/color/rgb.hpp
@@ -0,0 +1,156 @@
+/*
+ * Copyright (C) 2021  Christopher J. Howard
+ *
+ * This file is part of Antkeeper source code.
+ *
+ * Antkeeper source code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * Antkeeper source code is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Antkeeper source code.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef ANTKEEPER_COLOR_RGB_HPP
+#define ANTKEEPER_COLOR_RGB_HPP
+
+#include "color/cat.hpp"
+#include "math/math.hpp"
+
+namespace color {
+
+/// Functions which operate on various RGB colorspaces.
+namespace rgb {
+
+/**
+ * Constructs a matrix which transforms an RGB color into a CIE XYZ color.
+ *
+ * @param r CIE xy chromaticity coordinates of the red primary.
+ * @param g CIE xy chromaticity coordinates of the green primary.
+ * @param b CIE xy chromaticity coordinates of the blue primary.
+ * @param w CIE xy chromaticity coordinates of the white point.
+ *
+ * @return Matrix which transforms an RGB color into a CIE XYZ color.
+ *
+ * @see https://www.ryanjuckett.com/rgb-color-space-conversion/
+ * @see https://mina86.com/2019/srgb-xyz-matrix/
+ */
+template <class T>
+constexpr math::matrix<T, 3, 3> to_xyz(const math::vector2<T>& r, const math::vector2<T>& g, const math::vector2<T>& b, const math::vector2<T>& w)
+{
+	const math::matrix<T, 3, 3> m = 
+	{
+		r[0], r[1], T{1} - (r[0] + r[1]),
+		g[0], g[1], T{1} - (g[0] + g[1]),
+		b[0], b[1], T{1} - (b[0] + b[1])
+	};
+	
+	const math::vector3<T> scale = math::inverse(m) * math::vector3<T>{w[0] / w[1], T{1}, (T{1} - (w[0] + w[1])) / w[1]};
+	
+	return math::matrix<T, 3, 3>
+	{
+		m[0][0] * scale[0], m[0][1] * scale[0], m[0][2] * scale[0],
+		m[1][0] * scale[1], m[1][1] * scale[1], m[1][2] * scale[1],
+		m[2][0] * scale[2], m[2][1] * scale[2], m[2][2] * scale[2],
+	};
+}
+
+/**
+ * RGB color space.
+ */
+template <class T>
+struct color_space
+{
+	/// Transfer function function pointer type.
+	typedef math::vector3<T> (*transfer_function_type)(const math::vector3<T>&);
+	
+	/// CIE xy chromaticity coordinates of the red primary.
+	const math::vector2<T> r;
+	
+	/// CIE xy chromaticity coordinates of the green primary.
+	const math::vector2<T> g;
+	
+	/// CIE xy chromaticity coordinates of the blue primary.
+	const math::vector2<T> b;
+	
+	/// CIE xy chromaticity coordinates of the white point.
+	const math::vector2<T> w;
+	
+	/// Function pointer to the electro-optical transfer function.
+	const transfer_function_type eotf;
+	
+	/// Function pointer to the inverse electro-optical transfer function.
+	const transfer_function_type inverse_eotf;
+	
+	/// Matrix which transforms an RGB color to a CIE XYZ color.
+	const math::matrix3x3<T> to_xyz;
+	
+	/// Matrix which transforms a CIE XYZ color to an RGB color.
+	const math::matrix3x3<T> from_xyz;
+	
+	/// Vector which gives the luminance of an RGB color via dot product.
+	const math::vector3<T> to_y;
+	
+	/**
+	 * Constructs an RGB color space.
+	 *
+	 * @param r CIE xy chromaticity coordinates of the red primary.
+	 * @param g CIE xy chromaticity coordinates of the green primary.
+	 * @param b CIE xy chromaticity coordinates of the blue primary.
+	 * @param w CIE xy chromaticity coordinates of the white point.
+	 */
+	constexpr color_space(const math::vector2<T>& r, const math::vector2<T>& g, const math::vector2<T>& b, const math::vector2<T>& w, transfer_function_type eotf, transfer_function_type inverse_eotf);
+	
+	/**
+	 * Measures the luminance of a linear RGB color.
+	 *
+	 * @param x Linear RGB color.
+	 * @return return Luminance of @p x.
+	 */
+	constexpr T luminance(const math::vector3<T>& x) const;
+};
+
+template <class T>
+constexpr color_space<T>::color_space(const math::vector2<T>& r, const math::vector2<T>& g, const math::vector2<T>& b, const math::vector2<T>& w, transfer_function_type eotf, transfer_function_type inverse_eotf):
+	r(r),
+	g(g),
+	b(b),
+	w(w),
+	eotf(eotf),
+	inverse_eotf(inverse_eotf),
+	to_xyz(color::rgb::to_xyz<T>(r, g, b, w)),
+	from_xyz(math::inverse(to_xyz)),
+	to_y{to_xyz[0][1], to_xyz[1][1], to_xyz[2][1]}
+{}
+
+template <class T>
+constexpr T color_space<T>::luminance(const math::vector3<T>& x) const
+{
+	return math::dot(x, to_y);
+}
+
+/**
+ * Constructs a matrix which transforms a color from one RGB color space to another RGB color space.
+ *
+ * @param s0 Source color space.
+ * @param s1 Destination color space.
+ * @param cone_response Chromatic adaptation transform cone response matrix.
+ * 
+ * @return Color space transformation matrix.
+ */
+template <class T>
+constexpr math::matrix3x3<T> to_rgb(const color_space<T>& s0, const color_space<T>& s1, const math::matrix3x3<T>& cone_response = color::cat::bradford<T>)
+{
+	return s1.from_xyz * color::cat::matrix(s0.w, s1.w, cone_response) * s0.to_xyz;
+}
+
+} // namespace rgb
+} // namespace color
+
+#endif // ANTKEEPER_COLOR_RGB_HPP
diff --git a/src/color/srgb.hpp b/src/color/srgb.hpp
index 03561be..a43e429 100644
--- a/src/color/srgb.hpp
+++ b/src/color/srgb.hpp
@@ -20,149 +20,71 @@
 #ifndef ANTKEEPER_COLOR_SRGB_HPP
 #define ANTKEEPER_COLOR_SRGB_HPP
 
+#include "color/rgb.hpp"
+#include "color/illuminant.hpp"
 #include "math/math.hpp"
 #include <cmath>
 
 namespace color {
 
-/// Functions which operate in the sRGB colorspace.
-namespace srgb {
-
-/// CIE chromaticity coordinates of the sRGB white point (D65)
-template <class T>
-constexpr math::vector2<T> whitepoint = {0.31271, 0.32902};
-
 /**
- * Performs the sRGB Electro-Optical Transfer Function (EOTF), also known as the sRGB decoding function.
+ * sRGB electro-optical transfer function (EOTF), also known as the sRGB decoding function.
  *
  * @param v sRGB electrical signal (gamma-encoded sRGB).
+ *
  * @return Corresponding luminance of the signal (linear sRGB).
  */
-/// @{
-float eotf(float v);
-double eotf(double v);
 template <class T>
-math::vector3<T> eotf(const math::vector3<T>& v);
-/// @}
+math::vector3<T> srgb_eotf(const math::vector3<T>& v)
+{
+	auto f = [](T x) -> T
+	{
+		return x < T{0.04045} ? x / T{12.92} : std::pow((x + T{0.055}) / T{1.055}, T{2.4});
+	};
+	
+	return math::vector3<T>
+	{
+		f(v[0]),
+		f(v[1]),
+		f(v[2])
+	};
+}
 
 /**
- * Performs the sRGB inverse Electro-Optical Transfer Function (EOTF), also known as the sRGB encoding function.
+ * sRGB inverse electro-optical transfer function (EOTF), also known as the sRGB encoding function.
  *
  * @param l sRGB luminance (linear sRGB).
- * @return Corresponding electrical signal (gamma-encoded sRGB).
- */
-/// @{
-float eotf_inverse(float v);
-double eotf_inverse(double v);
-template <class T>
-math::vector3<T> eotf_inverse(const math::vector3<T>& l);
-/// @}
-
-/**
- * Calculates the luminance of a linear sRGB color.
  *
- * @param x Linear sRGB color.
- * @return return Luminance of @p x.
- */
-template <class T>
-T luminance(const math::vector3<T>& x);
-
-/**
- * Transforms a linear sRGB color into the ACEScg colorspace using the Bradford chromatic adaption transform.
- *
- * @param x Linear sRGB color.
- * @return ACEScg color.
- *
- * @see https://www.colour-science.org/apps/
- */
-template <class T>
-math::vector3<T> to_acescg(const math::vector3<T>& x);
-
-/**
- * Transforms a linear sRGB color into the CIE XYZ colorspace.
- *
- * @param x Linear sRGB color.
- * @return CIE XYZ color.
+ * @return Corresponding electrical signal (gamma-encoded sRGB).
  */
 template <class T>
-math::vector3<T> to_xyz(const math::vector3<T>& x);
-
-inline float eotf(float v)
-{
-	return (v < 0.04045f) ? (v / 12.92f) : std::pow((v + 0.055f) / 1.055f, 2.4f);
-}
-
-inline double eotf(double v)
+math::vector3<T> srgb_inverse_eotf(const math::vector3<T>& l)
 {
-	return (v < 0.04045) ? (v / 12.92) : std::pow((v + 0.055) / 1.055, 2.4);
-}
-
-template <class T>
-math::vector3<T> eotf(const math::vector3<T>& v)
-{
-	return math::vector3<T>
-		{
-			eotf(v[0]),
-			eotf(v[1]),
-			eotf(v[2])
-		};
-}
-
-inline float eotf_inverse(float l)
-{
-	return (l <= 0.0031308f) ? (l * 12.92f) : (std::pow(l, 1.0f / 2.4f) * 1.055f - 0.055f);
-}
-
-inline double eotf_inverse(double l)
-{
-	return (l <= 0.0031308) ? (l * 12.92) : (std::pow(l, 1.0 / 2.4) * 1.055 - 0.055);
-}
-
-template <class T>
-math::vector3<T> eotf_inverse(const math::vector3<T>& l)
-{
-	return math::vector3<T>
-		{
-			eotf_inverse(l[0]),
-			eotf_inverse(l[1]),
-			eotf_inverse(l[2])
-		};
-}
-
-template <class T>
-T luminance(const math::vector3<T>& x)
-{
-	static const math::vector3<T> luma = {0.212639005871510, 0.715168678767756, 0.072192315360734};
-	return math::dot(x, luma);
-}
-
-template <class T>
-math::vector3<T> to_acescg(const math::vector3<T>& x)
-{
-	static const math::matrix3<T> srgb_to_acescg
-	{{
-		{0.613132422390542, 0.070124380833917, 0.020587657528185},
-		{0.339538015799666, 0.916394011313573, 0.109574571610682},
-		{0.047416696048269, 0.013451523958235, 0.869785404035327}
-	}};
+	auto f = [](T x) -> T
+	{
+		return x <= T{0.0031308} ? x * T{12.92} : std::pow(x, T{1} / T{2.4}) * T{1.055} - T{0.055};
+	};
 	
-	return srgb_to_acescg * x;
+	return math::vector3<T>
+	{
+		f(l[0]),
+		f(l[1]),
+		f(l[2])
+	};
 }
 
+/// sRGB color space.
 template <class T>
-math::vector3<T> to_xyz(const math::vector3<T>& x)
-{
-	static const math::matrix3<T> srgb_to_xyz
-	{{
-		{0.412390799265959, 0.212639005871510, 0.019330818715592},
-		{0.357584339383878, 0.715168678767756, 0.119194779794626},
-		{0.180480788401834, 0.072192315360734, 0.950532152249661}
-	}};
-	
-	return srgb_to_xyz * x;
-}
+constexpr rgb::color_space<T> srgb
+(
+	{T{0.64}, T{0.33}},
+	{T{0.30}, T{0.60}},
+	{T{0.15}, T{0.06}},
+	color::illuminant::deg2::d65<T>,
+	&srgb_eotf<T>,
+	&srgb_inverse_eotf
+);
 
-} // namespace srgb
 } // namespace color
 
 #endif // ANTKEEPER_COLOR_SRGB_HPP
diff --git a/src/color/ucs.hpp b/src/color/ucs.hpp
index a19819f..7b6048b 100644
--- a/src/color/ucs.hpp
+++ b/src/color/ucs.hpp
@@ -31,23 +31,14 @@ namespace ucs {
  * Transforms CIE 1960 UCS chromaticity coordinates into the CIE xyY colorspace.
  *
  * @param uv CIE 1960 UCS chromaticity coordinates.
- * @param luminance Luminance or `Y` value of the resulting xyY color.
+ * @param y Luminance or `Y` value of the resulting xyY color.
  * @return CIE xyY color.
  */
 template <class T>
-math::vector3<T> to_xyy(const math::vector2<T>& uv, T luminance = T(1.0));
-
-template <class T>
-math::vector3<T> to_xyy(const math::vector2<T>& uv, T luminance)
+constexpr math::vector3<T> to_xyy(const math::vector2<T>& uv, T y = T{1})
 {
-	const T inverse_denom = T(1.0) / (T(2.0) * uv[0] - T(8.0) * uv[1] + T(4.0));
-	
-	return math::vector3<T>
-		{
-			(T(3.0) * uv[0]) * inverse_denom,
-			(T(2.0) * uv[1]) * inverse_denom,
-			luminance
-		};
+	const T d = T{1} / (T{2} * uv[0] - T{8} * uv[1] + T{4});
+	return math::vector3<T>{(T{3} * uv[0]) * d, (T{2} * uv[1]) * d, y};
 }
 
 } // namespace ucs
diff --git a/src/color/xyy.hpp b/src/color/xyy.hpp
index f3e9d4d..975ffb2 100644
--- a/src/color/xyy.hpp
+++ b/src/color/xyy.hpp
@@ -34,7 +34,10 @@ namespace xyy {
  * @return return Luminance of @p x.
  */
 template <class T>
-T luminance(const math::vector3<T>& x);
+constexpr inline T luminance(const math::vector3<T>& x)
+{
+	return x[2];
+}
 
 /**
  * Transforms a CIE xyY color into the CIE 1960 UCS colorspace.
@@ -43,7 +46,11 @@ T luminance(const math::vector3<T>& x);
  * @return CIE 1960 UCS color.
  */
 template <class T>
-math::vector2<T> to_ucs(const math::vector3<T>& x);
+constexpr math::vector2<T> to_ucs(const math::vector3<T>& x)
+{
+	const T d = T({1} / (T{-2} * x[0] + T{12} * x[1] + T{3}));
+	return math::vector2<T>{(T{4} * x[0]) * d, (T{6} * x[1]) * d};
+}
 
 /**
  * Transforms a CIE xyY color into the CIE XYZ colorspace.
@@ -52,35 +59,9 @@ math::vector2<T> to_ucs(const math::vector3<T>& x);
  * @return CIE XYZ color.
  */
 template <class T>
-math::vector3<T> to_xyz(const math::vector3<T>& x);
-
-template <class T>
-inline T luminance(const math::vector3<T>& x)
-{
-	return x[2];
-}
-
-template <class T>
-math::vector2<T> to_ucs(const math::vector3<T>& x)
-{
-	const T inverse_denom = T(1.0) / (T(-2.0) * x[0] + T(12.0) * x[1] + T(3.0));
-	
-	return math::vector2<T>
-		{
-			(T(4.0) * x[0]) * inverse_denom,
-			(T(6.0) * x[1]) * inverse_denom
-		};
-}
-
-template <class T>
-math::vector3<T> to_xyz(const math::vector3<T>& x)
+constexpr math::vector3<T> to_xyz(const math::vector3<T>& x)
 {
-	return math::vector3<T>
-		{
-			(x[0] * x[2]) / x[1],
-			x[2],
-			((T(1.0) - x[0] - x[1]) * x[2]) / x[1]
-		};
+	return math::vector3<T>{(x[0] * x[2]) / x[1], x[2], ((T{1} - x[0] - x[1]) * x[2]) / x[1]};
 }
 
 } // namespace xyy
diff --git a/src/color/xyz.hpp b/src/color/xyz.hpp
index 47a8108..279f55d 100644
--- a/src/color/xyz.hpp
+++ b/src/color/xyz.hpp
@@ -25,7 +25,7 @@
 namespace color {
 
 /**
- * Functions which operate in the CIE XYZ colorspace.
+ * Functions which operate in the CIE XYZ color space.
  *
  * @see https://en.wikipedia.org/wiki/CIE_1931_color_space
  */
@@ -38,22 +38,23 @@ namespace xyz {
  * @return return Luminance of @p x.
  */
 template <class T>
-T luminance(const math::vector3<T>& x);
+constexpr inline T luminance(const math::vector3<T>& x)
+{
+	return x[1];
+}
 
 /**
- * Fitted piecewise gaussian approximation to the CIE 1931 standard observer color matching function.
+ * Transforms a CIE XYZ color into the CIE xyY color space.
  *
- * @param lambda Wavelength of light, in nanometers.
- * @return Matching CIE XYZ color.
- *
- * @see match_x(T)
- * @see match_y(T)
- * @see match_z(T)
- *
- * @see Wyman, C., Sloan, P.J., & Shirley, P. (2013). Simple Analytic Approximations to the CIE XYZ Color Matching Functions.
+ * @param x CIE XYZ color.
+ * @return CIE xyY color.
  */
 template <class T>
-math::vector3<T> match(T lambda);
+constexpr math::vector3<T> to_xyy(const math::vector3<T>& x)
+{
+	const T sum = x[0] + x[1] + x[2];
+	return math::vector3<T>{x[0] / sum, x[1] / sum, x[1]}
+}
 
 /**
  * CIE 1931 standard observer color matching function for the X tristimulus value.
@@ -63,98 +64,6 @@ math::vector3<T> match(T lambda);
  *
  * @see match(T)
  */
-template <class T>
-T match_x(T lambda);
-
-/**
- * CIE 1931 standard observer color matching function for the Y tristimulus value.
- *
- * @param lambda Wavelength of light, in nanometers.
- * @return Matching Y tristimulus value.
- *
- * @see match(T)
- */
-template <class T>
-T match_y(T lambda);
-
-/**
- * CIE 1931 standard observer color matching function for the Z tristimulus value.
- *
- * @param lambda Wavelength of light, in nanometers.
- * @return Matching Z tristimulus value.
- *
- * @see match(T)
- */
-template <class T>
-T match_z(T lambda);
-
-/**
- * Transforms a CIE XYZ color into the ACEScg colorspace.
- *
- * @param x CIE XYZ color.
- * @return ACEScg color.
- */
-template <class T>
-math::vector3<T> to_acescg(const math::vector3<T>& x);
-
-/**
- * Transforms a CIE XYZ color into the linear sRGB colorspace.
- *
- * @param x CIE XYZ color.
- * @return Linear sRGB color.
- */
-template <class T>
-math::vector3<T> to_srgb(const math::vector3<T>& x);
-
-/**
- * Transforms a CIE XYZ color into the CIE xyY colorspace.
- *
- * @param x CIE XYZ color.
- * @return CIE xyY color.
- */
-template <class T>
-math::vector3<T> to_xyy(const math::vector3<T>& x);
-
-/// Chromatic Adaptation Transforms (CATs).
-namespace cat {
-
-	/**
-	 * Transforms a CIE XYZ color with an ACES (~D60) illuminant to a CIE XYZ color with a D65 illuminant using the Bradford chromatic adaption transform.
-	 *
-	 * @param x CIE XYZ color with an ACES (~D60) illuminant.
-	 * @return CIE XYZ color with a D65 illuminant.
-	 */
-	template <class T>
-	math::vector3<T> aces_to_d65(const math::vector3<T>& x);
-
-	/**
-	 * Transforms a CIE XYZ color with a D65 illuminant to a CIE XYZ color with an ACES (~D60) illuminant using the Bradford chromatic adaption transform.
-	 *
-	 * @param x CIE XYZ color with a D65 illuminant.
-	 * @return CIE XYZ color with an ACES (~D60) illuminant.
-	 */
-	template <class T>
-	math::vector3<T> d65_to_aces(const math::vector3<T>& x);
-
-} // namespace cat
-
-template <class T>
-inline T luminance(const math::vector3<T>& x)
-{
-	return x[1];
-}
-
-template <class T>
-math::vector3<T> match(T lambda)
-{
-	return math::vector3<T>
-	{
-		match_x<T>(lambda),
-		match_y<T>(lambda),
-		match_z<T>(lambda)
-	};
-}
-
 template <class T>
 T match_x(T lambda)
 {
@@ -169,6 +78,14 @@ T match_x(T lambda)
 	return x0 + x1 + x2;
 }
 
+/**
+ * CIE 1931 standard observer color matching function for the Y tristimulus value.
+ *
+ * @param lambda Wavelength of light, in nanometers.
+ * @return Matching Y tristimulus value.
+ *
+ * @see match(T)
+ */
 template <class T>
 T match_y(T lambda)
 {
@@ -181,6 +98,14 @@ T match_y(T lambda)
 	return y0 + y1;
 }
 
+/**
+ * CIE 1931 standard observer color matching function for the Z tristimulus value.
+ *
+ * @param lambda Wavelength of light, in nanometers.
+ * @return Matching Z tristimulus value.
+ *
+ * @see match(T)
+ */
 template <class T>
 T match_z(T lambda)
 {
@@ -193,74 +118,28 @@ T match_z(T lambda)
 	return z0 + z1;
 }
 
+/**
+ * Fitted piecewise gaussian approximation to the CIE 1931 standard observer color matching function.
+ *
+ * @param lambda Wavelength of light, in nanometers.
+ * @return Matching CIE XYZ color.
+ *
+ * @see match_x(T)
+ * @see match_y(T)
+ * @see match_z(T)
+ *
+ * @see Wyman, C., Sloan, P.J., & Shirley, P. (2013). Simple Analytic Approximations to the CIE XYZ Color Matching Functions.
+ */
 template <class T>
-math::vector3<T> to_acescg(const math::vector3<T>& x)
-{
-	static const math::matrix3<T> xyz_to_acescg
-	{{
-		{ 1.641023379694326, -0.663662858722983,  0.011721894328375},
-		{-0.324803294184790,  1.615331591657338, -0.008284441996237},
-		{-0.236424695237612,  0.016756347685530,  0.988394858539022}
-	}};
-	
-	return xyz_to_acescg * x;
-}
-
-template <class T>
-math::vector3<T> to_srgb(const math::vector3<T>& x)
-{
-	static const math::matrix3<T> xyz_to_srgb
-	{{
-		{ 3.240969941904523, -0.969243636280880,  0.055630079696994},
-		{-1.537383177570094,  1.875967501507721, -0.203976958888977},
-		{-0.498610760293003,  0.041555057407176,  1.056971514242879}
-	}};
-	
-	return xyz_to_srgb * x;
-}
-
-template <class T>
-math::vector3<T> to_xyy(const math::vector3<T>& x)
+math::vector3<T> match(T lambda)
 {
-	const T sum = x[0] + x[1] + x[2];
-	
 	return math::vector3<T>
-		{
-			x[0] / sum
-			x[1] / sum,
-			x[1]
-		};
-}
-
-namespace cat {
-
-	template <class T>
-	math::vector3<T> aces_to_d65(const math::vector3<T>& x)
 	{
-		static const math::matrix3<T> cat_aces_to_d65
-		{{
-			{ 0.987225397551079, -0.007603864790602,  0.003066041131217},
-			{-0.006114800968213,  1.001874800208719, -0.005084242870792},
-			{ 0.015926393295811,  0.005322023893623,  1.081519155692042}
-		}};
-		
-		return cat_aces_to_d65 * x;
-	}
-
-	template <class T>
-	math::vector3<T> d65_to_aces(const math::vector3<T>& x)
-	{
-		static const math::matrix3<T> cat_d65_to_aces
-		{{
-			{ 1.013033366661459,  0.007703617525351, -0.002835673220262},
-			{ 0.006107049021859,  0.998150224406968,  0.004675010768250},
-			{-0.014947927269674, -0.005025218608126,  0.924644077051100}
-		}};
-		
-		return cat_d65_to_aces * x;
-	}
-
-} // namespace cat
+		match_x<T>(lambda),
+		match_y<T>(lambda),
+		match_z<T>(lambda)
+	};
+}
 
 } // namespace xyz
 } // namespace color
diff --git a/src/config.hpp.in b/src/config.hpp.in
index f97f258..a02bd2a 100644
--- a/src/config.hpp.in
+++ b/src/config.hpp.in
@@ -44,7 +44,7 @@ constexpr float menu_fade_out_duration = 0.125f;
 constexpr float menu_mouseover_padding = 0.1f;
 
 /// Opacity of the menu background.
-constexpr float menu_bg_opacity = 2.0f / 3.0f;
+constexpr float menu_bg_opacity = 2.0f / 4.0f;
 
 /// RGBA color of active menu items.
 constexpr float4 menu_active_color{1.0f, 1.0f, 1.0f, 1.0f};
@@ -66,7 +66,7 @@ constexpr float nuptial_flight_fade_in_duration = 5.0f;
 
 #define MATERIAL_PASS_MAX_AMBIENT_LIGHT_COUNT 1
 #define MATERIAL_PASS_MAX_POINT_LIGHT_COUNT 1
-#define MATERIAL_PASS_MAX_DIRECTIONAL_LIGHT_COUNT 2
+#define MATERIAL_PASS_MAX_DIRECTIONAL_LIGHT_COUNT 3
 #define MATERIAL_PASS_MAX_SPOTLIGHT_COUNT 1
 #define MATERIAL_PASS_MAX_BONE_COUNT 64
 #define TERRAIN_PATCH_SIZE 200.0f
diff --git a/src/entity/components/atmosphere.hpp b/src/entity/components/atmosphere.hpp
index 50e28ca..340b21f 100644
--- a/src/entity/components/atmosphere.hpp
+++ b/src/entity/components/atmosphere.hpp
@@ -34,32 +34,35 @@ struct atmosphere
 	/// Index of refraction of air at sea level.
 	double index_of_refraction;
 	
-	/// Molar concentration of air at sea level, in mol/m-3.
-	double air_concentration;
+	/// Molar concentration of Rayleigh particles at sea level, in mol/m-3.
+	double rayleigh_concentration;
 	
 	/// Scale height of the exponential distribution of Rayleigh particles, in meters.
 	double rayleigh_scale_height;
 	
-	/// Molecular density of Rayleigh particles at sea level.
-	double rayleigh_density;
-	
 	/// (Dependent) Rayleigh scattering coefficients.
 	double3 rayleigh_scattering;
 	
+	/// Molar concentration of Mie particles at sea level, in mol/m-3.
+	double mie_concentration;
+	
 	/// Scale height of the exponential distribution of Mie particles, in meters.
 	double mie_scale_height;
 	
-	/// Molecular density of Mie particles at sea level.
-	double mie_density;
+	/// Mie phase function anisotropy factor.
+	double mie_anisotropy;
+	
+	/// Mie single-scattering albedo.
+	double mie_albedo;
 	
 	/// (Dependent) Mie scattering coefficient.
 	double mie_scattering;
 	
-	/// (Dependent) Mie absorption coefficient.
-	double mie_absorption;
+	/// (Dependent) Mie extinction coefficient.
+	double mie_extinction;
 	
-	/// Mie phase function anisotropy factor.
-	double mie_anisotropy;
+	/// Concentration of ozone in the atmosphere, unitless.
+	double ozone_concentration;
 	
 	/// Elevation of the lower limit of the triangular distribution of ozone particles, in meters.
 	double ozone_lower_limit;
@@ -70,14 +73,11 @@ struct atmosphere
 	/// Elevation of the mode of the triangular distribution of ozone particles, in meters.
 	double ozone_mode;
 	
-	/// Concentration of ozone in the atmosphere, unitless.
-	double ozone_concentration;
-	
 	/// (Dependent) Ozone absorption coefficients.
 	double3 ozone_absorption;
 	
-	/// Airglow, in lux.
-	double airglow;
+	/// Airglow illuminance, in lux.
+	double3 airglow_illuminance;
 };
 
 } // namespace component
diff --git a/src/entity/components/observer.hpp b/src/entity/components/observer.hpp
index 8b8bff3..aadb678 100644
--- a/src/entity/components/observer.hpp
+++ b/src/entity/components/observer.hpp
@@ -21,20 +21,26 @@
 #define ANTKEEPER_ENTITY_COMPONENT_OBSERVER_HPP
 
 #include "entity/id.hpp"
-#include "utility/fundamental-types.hpp"
-#include "scene/camera.hpp"
 
 namespace entity {
 namespace component {
 
-/// Observer
+/**
+ * 
+ */
 struct observer
 {
+	/// Entity ID of a celestial body to which the observer position is relative.
 	entity::id reference_body_eid;
+	
+	/// Elevation of the observer, in radians.
 	double elevation;
+	
+	/// Latitude of the observer, in radians.
 	double latitude;
+	
+	/// Longitude of the observer, in radians.
 	double longitude;
-	scene::camera* camera;
 };
 
 } // namespace component
diff --git a/src/entity/systems/astronomy.cpp b/src/entity/systems/astronomy.cpp
index 1f92388..9085e2e 100644
--- a/src/entity/systems/astronomy.cpp
+++ b/src/entity/systems/astronomy.cpp
@@ -27,6 +27,7 @@
 #include "color/color.hpp"
 #include "physics/orbit/orbit.hpp"
 #include "physics/time/ut1.hpp"
+#include "physics/time/jd.hpp"
 #include "physics/light/photometry.hpp"
 #include "physics/light/luminosity.hpp"
 #include "physics/light/refraction.hpp"
@@ -40,107 +41,107 @@
 namespace entity {
 namespace system {
 
-template <class T>
-math::vector3<T> transmittance(T depth_r, T depth_m, T depth_o, const math::vector3<T>& extinction_r, T extinction_m, const math::vector3<T>& extinction_o)
-{
-	math::vector3<T> transmittance_r = extinction_r * depth_r;
-	math::vector3<T> transmittance_m = math::vector3<T>{extinction_m, extinction_m, extinction_m} * depth_m;
-	math::vector3<T> transmittance_o = extinction_o * depth_o;
-	
-	math::vector3<T> t = transmittance_r + transmittance_m + transmittance_o;
-	t.x = std::exp(-t.x);
-	t.y = std::exp(-t.y);
-	t.z = std::exp(-t.z);
-	
-	return t;
-}
-
 astronomy::astronomy(entity::registry& registry):
 	updatable(registry),
-	time(0.0),
+	time_days(0.0),
+	time_centuries(0.0),
 	time_scale(1.0),
-	reference_entity(entt::null),
-	observer_location{0, 0, 0},
+	observer_eid(entt::null),
+	reference_body_eid(entt::null),
+	transmittance_samples(0),
 	sun_light(nullptr),
 	sky_light(nullptr),
 	moon_light(nullptr),
-	camera(nullptr),
+	bounce_light(nullptr),
+	bounce_albedo{0, 0, 0},
 	sky_pass(nullptr),
-	exposure_offset(0.0),
-	starlight_illuminance(0.0)
+	starlight_illuminance{0, 0, 0}
 {
-	// Construct transformation which transforms coordinates from ENU to EUS
+	// Construct ENU to EUS transformation
 	enu_to_eus = math::transformation::se3<double>
 	{
 		{0, 0, 0},
 		math::quaternion<double>::rotate_x(-math::half_pi<double>)
 	};
 	
-	registry.on_construct<entity::component::celestial_body>().connect<&astronomy::on_celestial_body_construct>(this);
-	registry.on_replace<entity::component::celestial_body>().connect<&astronomy::on_celestial_body_replace>(this);
+	registry.on_construct<entity::component::observer>().connect<&astronomy::on_observer_modified>(this);
+	registry.on_replace<entity::component::observer>().connect<&astronomy::on_observer_modified>(this);
+	registry.on_destroy<entity::component::observer>().connect<&astronomy::on_observer_destroyed>(this);
+	registry.on_construct<entity::component::celestial_body>().connect<&astronomy::on_celestial_body_modified>(this);
+	registry.on_replace<entity::component::celestial_body>().connect<&astronomy::on_celestial_body_modified>(this);
+	registry.on_destroy<entity::component::celestial_body>().connect<&astronomy::on_celestial_body_destroyed>(this);
+	registry.on_construct<entity::component::orbit>().connect<&astronomy::on_orbit_modified>(this);
+	registry.on_replace<entity::component::orbit>().connect<&astronomy::on_orbit_modified>(this);
+	registry.on_destroy<entity::component::orbit>().connect<&astronomy::on_orbit_destroyed>(this);
+	registry.on_construct<entity::component::atmosphere>().connect<&astronomy::on_atmosphere_modified>(this);
+	registry.on_replace<entity::component::atmosphere>().connect<&astronomy::on_atmosphere_modified>(this);
+	registry.on_destroy<entity::component::atmosphere>().connect<&astronomy::on_atmosphere_destroyed>(this);
 }
 
 astronomy::~astronomy()
 {
-	registry.on_construct<entity::component::celestial_body>().disconnect<&astronomy::on_celestial_body_construct>(this);
-	registry.on_replace<entity::component::celestial_body>().disconnect<&astronomy::on_celestial_body_replace>(this);
+	registry.on_construct<entity::component::observer>().disconnect<&astronomy::on_observer_modified>(this);
+	registry.on_replace<entity::component::observer>().disconnect<&astronomy::on_observer_modified>(this);
+	registry.on_destroy<entity::component::observer>().disconnect<&astronomy::on_observer_destroyed>(this);
+	registry.on_construct<entity::component::celestial_body>().disconnect<&astronomy::on_celestial_body_modified>(this);
+	registry.on_replace<entity::component::celestial_body>().disconnect<&astronomy::on_celestial_body_modified>(this);
+	registry.on_destroy<entity::component::celestial_body>().disconnect<&astronomy::on_celestial_body_destroyed>(this);
+	registry.on_construct<entity::component::orbit>().disconnect<&astronomy::on_orbit_modified>(this);
+	registry.on_replace<entity::component::orbit>().disconnect<&astronomy::on_orbit_modified>(this);
+	registry.on_destroy<entity::component::orbit>().disconnect<&astronomy::on_orbit_destroyed>(this);
+	registry.on_construct<entity::component::atmosphere>().disconnect<&astronomy::on_atmosphere_modified>(this);
+	registry.on_replace<entity::component::atmosphere>().disconnect<&astronomy::on_atmosphere_modified>(this);
+	registry.on_destroy<entity::component::atmosphere>().disconnect<&astronomy::on_atmosphere_destroyed>(this);
 }
 
 void astronomy::update(double t, double dt)
 {
-	double total_illuminance = 0.0;
-	double sky_light_illuminance = 0.0;
+	double3 sky_light_illuminance = {0.0, 0.0, 0.0};
 	
 	// Add scaled timestep to current time
-	set_time(time + dt * time_scale);
+	set_time(time_days + dt * time_scale);
 	
-	// Abort if no reference body
-	if (reference_entity == entt::null)
+	// Abort if no valid observer entity or reference body entity
+	if (observer_eid == entt::null || reference_body_eid == entt::null)
 		return;
-
-	// Abort if either reference body or orbit have not been set
-	if (!registry.has<entity::component::orbit>(reference_entity) || !registry.has<entity::component::celestial_body>(reference_entity))
-		return;
-	const entity::component::orbit& reference_orbit = registry.get<entity::component::orbit>(reference_entity);
-	const entity::component::celestial_body& reference_body = registry.get<entity::component::celestial_body>(reference_entity);
 	
-	math::transformation::se3<double> icrf_to_bci{{-reference_orbit.position}, math::quaternion<double>::identity};
+	// Get pointer to observer component
+	const auto observer = registry.try_get<component::observer>(observer_eid);
 	
-	const double days_from_epoch = time;
-	const double centuries_from_epoch = days_from_epoch / 36525.0;
+	// Abort if no observer component
+	if (!observer)
+		return;
 	
-	// Evaluate reference body orientation polynomials
-	const double reference_body_pole_ra = math::polynomial::horner(reference_body.pole_ra.begin(), reference_body.pole_ra.end(), centuries_from_epoch);
-	const double reference_body_pole_dec = math::polynomial::horner(reference_body.pole_dec.begin(), reference_body.pole_dec.end(), centuries_from_epoch);
-	const double reference_body_prime_meridian = math::polynomial::horner(reference_body.prime_meridian.begin(), reference_body.prime_meridian.end(), days_from_epoch);
+	// Get pointers to reference body components
+	const auto
+	[
+		reference_body,
+		reference_orbit,
+		reference_atmosphere
+	] = registry.try_get<component::celestial_body, component::orbit, component::atmosphere>(reference_body_eid);
 	
-	// Construct transformation from the ICRF frame to the reference body BCBF frame	
-	icrf_to_bcbf = physics::orbit::frame::bci::to_bcbf
-	(
-		reference_body_pole_ra,
-		reference_body_pole_dec,
-		reference_body_prime_meridian
-	);
-	icrf_to_bcbf.t = icrf_to_bcbf.r * -reference_orbit.position;
+	// Abort if no reference body or reference orbit
+	if (!reference_body || !reference_orbit)
+		return;
 	
-	icrf_to_enu = icrf_to_bcbf * bcbf_to_enu;
-	icrf_to_eus = icrf_to_enu * enu_to_eus;
+	// Update ICRF to EUS transformation
+	update_icrf_to_eus(*reference_body, *reference_orbit);
 	
 	// Set the transform component translations of orbiting bodies to their topocentric positions
 	registry.view<component::celestial_body, component::orbit, component::transform>().each(
 	[&](entity::id entity_id, const auto& body, const auto& orbit, auto& transform)
 	{
-		// Skip reference entity
-		if (entity_id == this->reference_entity)
+		// Skip reference body entity
+		if (entity_id == reference_body_eid)
 			return;
 		
 		// Transform orbital Cartesian position (r) from the ICRF frame to the EUS frame
 		const double3 r_eus = icrf_to_eus * orbit.position;
 		
 		// Evaluate body orientation polynomials
-		const double body_pole_ra = math::polynomial::horner(body.pole_ra.begin(), body.pole_ra.end(), centuries_from_epoch);
-		const double body_pole_dec = math::polynomial::horner(body.pole_dec.begin(), body.pole_dec.end(), centuries_from_epoch);
-		const double body_prime_meridian = math::polynomial::horner(body.prime_meridian.begin(), body.prime_meridian.end(), days_from_epoch);
+		const double body_pole_ra = math::polynomial::horner(body.pole_ra.begin(), body.pole_ra.end(), time_centuries);
+		const double body_pole_dec = math::polynomial::horner(body.pole_dec.begin(), body.pole_dec.end(), time_centuries);
+		const double body_prime_meridian = math::polynomial::horner(body.prime_meridian.begin(), body.prime_meridian.end(), time_days);
 		
 		// Determine body orientation in the ICRF frame
 		math::quaternion<double> rotation_icrf = physics::orbit::frame::bcbf::to_bci
@@ -160,133 +161,80 @@ void astronomy::update(double t, double dt)
 			transform.local.rotation = math::type_cast<float>(rotation_eus);
 			transform.local.scale = {1.0f, 1.0f, 1.0f};
 		}
-		
-		/*
-		if (orbit.parent != entt::null)
-		{
-			// RA-DEC
-			const double3 r_bci = icrf_to_bci * orbit.position;
-			double3 r_bci_spherical = physics::orbit::frame::bci::spherical(r_bci);
-			if (r_bci_spherical.z < 0.0)
-				r_bci_spherical.z += math::two_pi<double>;
-			const double dec = math::degrees(r_bci_spherical.y);
-			const double ra = math::degrees(r_bci_spherical.z);
-			
-			// AZ-EL
-			const double3 r_enu = icrf_to_enu * orbit.position;
-			double3 r_enu_spherical = physics::orbit::frame::enu::spherical(r_enu);
-			if (r_enu_spherical.z < 0.0)
-				r_enu_spherical.z += math::two_pi<double>;
-			const double el = math::degrees(r_enu_spherical.y);
-			const double az = math::degrees(r_enu_spherical.z);
-			
-			std::cout << "t: " << this->time << "; ra: " << ra << "; dec: " << dec << std::endl;
-			std::cout << "t: " << this->time << "; az: " << az << "; el: " << el << std::endl;
-		}
-		*/
 	});
 	
+	constexpr double3 bounce_normal = {0, 1, 0};
+	double3 bounce_illuminance = {0, 0, 0};
+	
 	// Update blackbody lighting
 	registry.view<component::celestial_body, component::orbit, component::blackbody>().each(
-	[&](entity::id entity_id, const auto& body, const auto& orbit, const auto& blackbody)
+	[&](entity::id entity_id, const auto& blackbody_body, const auto& blackbody_orbit, const auto& blackbody)
 	{
-		// Calculate blackbody inertial basis
-		//double3 blackbody_forward_icrf = math::normalize(reference_orbit.icrf_position - orbit.icrf_position);
-		double3 blackbody_up_icrf = {0, 0, 1};
+		// Transform blackbody position from ICRF frame to EUS frame
+		const double3 blackbody_position_eus = icrf_to_eus * blackbody_orbit.position;
 		
-		// Transform blackbody ICRF position and basis into EUS frame
-		double3 blackbody_position_eus = icrf_to_eus * orbit.position;
-		double3 blackbody_position_enu = icrf_to_enu * orbit.position;
-		double3 blackbody_forward_eus = math::normalize(-blackbody_position_eus);
-		double3 blackbody_up_eus = icrf_to_eus.r * blackbody_up_icrf;
+		// Measure distance and direction, in EUS frame, from observer to blackbody
+		const double observer_blackbody_distance = math::length(blackbody_position_eus);
+		const double3 observer_blackbody_direction_eus = blackbody_position_eus / observer_blackbody_distance;
 		
-		// Calculate distance from observer to blackbody
-		double blackbody_distance = math::length(blackbody_position_eus) - body.radius;
+		// Measure blackbody solid angle as seen by observer
+		const double observer_blackbody_angular_radius = astro::angular_radius(blackbody_body.radius, observer_blackbody_distance);
+		const double observer_blackbody_solid_angle = geom::solid_angle::cone(observer_blackbody_angular_radius);
 		
-		// Calculate blackbody solid angle
-		const double blackbody_angular_radius = astro::angular_radius(body.radius, blackbody_distance);
-		const double blackbody_solid_angle = geom::solid_angle::cone(blackbody_angular_radius);
+		// Calculate illuminance from blackbody reaching observer
+		const double3 observer_blackbody_illuminance = blackbody.luminance * observer_blackbody_solid_angle;
 		
-		// Calculate blackbody illuminance
-		const double3 blackbody_illuminance = blackbody.luminance * blackbody_solid_angle;
-		
-		// Init atmospheric transmittance
-		double3 atmospheric_transmittance = {1.0, 1.0, 1.0};
-		
-		// Get atmosphere component of reference body (if any)
-		if (this->registry.has<entity::component::atmosphere>(reference_entity))
+		// Calculate illuminance from blackbody reaching observer after atmospheric extinction
+		double3 observer_blackbody_transmitted_illuminance = observer_blackbody_illuminance;
+		if (reference_atmosphere)
 		{
-			const entity::component::atmosphere& reference_atmosphere = registry.get<entity::component::atmosphere>(reference_entity);
-
-			// Altitude of observer in meters
-			geom::ray<double> sample_ray;
-			sample_ray.origin = {0, reference_body.radius + observer_location[0], 0};
-			sample_ray.direction = math::normalize(blackbody_position_eus);
-			
-			geom::sphere<double> exosphere;
-			exosphere.center = {0, 0, 0};
-			exosphere.radius = reference_body.radius + reference_atmosphere.upper_limit;
-			
-			auto intersection_result = geom::ray_sphere_intersection(sample_ray, exosphere);
+			// Construct ray at observer pointing towards the blackbody
+			const geom::ray<double> ray = {{0, 0, 0}, observer_blackbody_direction_eus};
 			
-			if (std::get<0>(intersection_result))
-			{
-				double3 sample_start = sample_ray.origin;
-				double3 sample_end = sample_ray.extrapolate(std::get<2>(intersection_result));
-				
-				double optical_depth_r = physics::gas::atmosphere::optical_depth_exp(sample_start, sample_end, reference_body.radius, reference_atmosphere.rayleigh_scale_height, 16);
-				double optical_depth_m = physics::gas::atmosphere::optical_depth_exp(sample_start, sample_end, reference_body.radius, reference_atmosphere.mie_scale_height, 16);
-				double optical_depth_o = physics::gas::atmosphere::optical_depth_tri(sample_start, sample_end, reference_body.radius, reference_atmosphere.ozone_lower_limit, reference_atmosphere.ozone_upper_limit, reference_atmosphere.ozone_mode, 16);
-				
-				atmospheric_transmittance = transmittance(optical_depth_r, optical_depth_m, optical_depth_o, reference_atmosphere.rayleigh_scattering, reference_atmosphere.mie_scattering + reference_atmosphere.mie_absorption, reference_atmosphere.ozone_absorption);
-			}
+			// Integrate atmospheric spectral transmittance factor between observer and blackbody
+			const double3 transmittance = integrate_transmittance(*observer, *reference_body, *reference_atmosphere, ray);
 			
-			// Add airglow to sky light illuminance
-			sky_light_illuminance += reference_atmosphere.airglow;
+			// Attenuate illuminance from blackbody reaching observer by spectral transmittance factor
+			observer_blackbody_transmitted_illuminance *= transmittance;
 		}
 		
-		// Blackbody illuminance transmitted through the atmosphere
-		const double3 transmitted_blackbody_illuminance = blackbody_illuminance * atmospheric_transmittance;
-		
-		// Add atmosphere-transmitted blackbody illuminance to total illuminance
-		total_illuminance += (transmitted_blackbody_illuminance.x + transmitted_blackbody_illuminance.y + transmitted_blackbody_illuminance.z) / 3.0;
-		
 		// Update sun light
 		if (sun_light != nullptr)
 		{
-			sun_light->set_translation(math::normalize(math::type_cast<float>(blackbody_position_eus)));
+			const double3 blackbody_up_eus = icrf_to_eus.r * double3{0, 0, 1};
 			sun_light->set_rotation
 			(
 				math::look_rotation
 				(
-					math::type_cast<float>(blackbody_forward_eus),
+					math::type_cast<float>(-observer_blackbody_direction_eus),
 					math::type_cast<float>(blackbody_up_eus)
 				)
 			);
 			
-			sun_light->set_color(math::type_cast<float>(transmitted_blackbody_illuminance));	
+			sun_light->set_color(math::type_cast<float>(observer_blackbody_transmitted_illuminance));
+			
+			// Bounce sun light
+			bounce_illuminance += std::max(0.0, math::dot(bounce_normal, -observer_blackbody_direction_eus)) * observer_blackbody_transmitted_illuminance * bounce_albedo;
 		}
 		
 		// Update sky light
 		if (sky_light != nullptr)
 		{
 			// Calculate sky illuminance
-			double3 blackbody_position_enu_spherical = physics::orbit::frame::enu::spherical(blackbody_position_enu);
+			double3 blackbody_position_enu_spherical = physics::orbit::frame::enu::spherical(enu_to_eus.inverse() * blackbody_position_eus);
 			const double sky_illuminance = 25000.0 * std::max<double>(0.0, std::sin(blackbody_position_enu_spherical.y));
 			
 			// Add sky illuminance to sky light illuminance
-			sky_light_illuminance += sky_illuminance;
+			sky_light_illuminance += {sky_illuminance, sky_illuminance, sky_illuminance};
 			
 			// Add starlight illuminance to sky light illuminance
 			sky_light_illuminance += starlight_illuminance;
 			
-			// Add sky light illuminance to total illuminance
-			total_illuminance += sky_light_illuminance;
-			
-			//std::cout << "sky light illum: " << sky_light_illuminance << std::endl;
-			
 			// Update sky light
-			sky_light->set_color(float3{1.0f, 1.0f, 1.0f} * static_cast<float>(sky_light_illuminance));
+			sky_light->set_color(math::type_cast<float>(sky_light_illuminance));
+			
+			// Bounce sky light
+			bounce_illuminance += sky_light_illuminance * bounce_albedo;
 		}
 		
 		// Upload blackbody params to sky pass
@@ -294,143 +242,109 @@ void astronomy::update(double t, double dt)
 		{
 			this->sky_pass->set_sun_position(math::type_cast<float>(blackbody_position_eus));
 			this->sky_pass->set_sun_luminance(math::type_cast<float>(blackbody.luminance));
-			this->sky_pass->set_sun_illuminance(math::type_cast<float>(blackbody_illuminance));
-			this->sky_pass->set_sun_angular_radius(static_cast<float>(blackbody_angular_radius));
+			this->sky_pass->set_sun_illuminance(math::type_cast<float>(observer_blackbody_illuminance), math::type_cast<float>(observer_blackbody_transmitted_illuminance));
+			this->sky_pass->set_sun_angular_radius(static_cast<float>(observer_blackbody_angular_radius));
 		}
 		
-		const double3& blackbody_icrf_position = orbit.position;
-		
 		// Update diffuse reflectors
 		this->registry.view<component::celestial_body, component::orbit, component::diffuse_reflector, component::transform>().each(
 		[&](entity::id entity_id, const auto& reflector_body, const auto& reflector_orbit, const auto& reflector, const auto& transform)
 		{
-			// Calculate distance to blackbody and direction of incoming light
-			double3 blackbody_light_direction_icrf = reflector_orbit.position - blackbody_icrf_position;
-			double blackbody_distance = math::length(blackbody_light_direction_icrf);
-			blackbody_light_direction_icrf = blackbody_light_direction_icrf / blackbody_distance;
+			// Transform reflector position from ICRF frame to EUS frame
+			const double3 reflector_position_eus = icrf_to_eus * reflector_orbit.position;
+			
+			// Measure distance and direction, in EUS frame, from observer to reflector
+			const double observer_reflector_distance = math::length(reflector_position_eus);
+			const double3 observer_reflector_direction_eus = reflector_position_eus / observer_reflector_distance;
 			
-			// Transform blackbody light direction from the ICRF frame to the EUS frame
-			double3 blackbody_light_direction_eus = icrf_to_eus.r * blackbody_light_direction_icrf;
+			// Measure distance and direction, in EUS frame, from reflector to blackbody
+			double3 reflector_blackbody_direction_eus = blackbody_position_eus - reflector_position_eus;
+			const double reflector_blackbody_distance = math::length(reflector_blackbody_direction_eus);
+			reflector_blackbody_direction_eus /= reflector_blackbody_distance;
 			
-			// Calculate blackbody solid angle
-			const double blackbody_angular_radius = astro::angular_radius(body.radius, blackbody_distance);
-			const double blackbody_solid_angle = geom::solid_angle::cone(blackbody_angular_radius);
+			// Measure blackbody solid angle as seen by reflector
+			const double reflector_blackbody_angular_radius = astro::angular_radius(blackbody_body.radius, reflector_blackbody_distance);
+			const double reflector_blackbody_solid_angle = geom::solid_angle::cone(reflector_blackbody_angular_radius);
 			
-			// Calculate blackbody illuminance
+			// Calculate blackbody illuminance reaching reflector
+			const double3 reflector_blackbody_illuminance = blackbody.luminance * reflector_blackbody_solid_angle;
 			
-			double3 view_direction_icrf = reflector_orbit.position - reference_orbit.position;
-			const double reflector_distance = math::length(view_direction_icrf);
-			view_direction_icrf = view_direction_icrf / reflector_distance;
+			// Measure reflector solid angle as seen by observer
+			const double observer_reflector_angular_radius = astro::angular_radius(reflector_body.radius, observer_reflector_distance);
+			const double observer_reflector_solid_angle = geom::solid_angle::cone(observer_reflector_angular_radius);
 			
-			const double3 sunlight_illuminance = blackbody.luminance * blackbody_solid_angle;
-			const double reflector_angular_radius = astro::angular_radius(reflector_body.radius, reflector_distance);
-			const double reflector_solid_angle = geom::solid_angle::cone(reflector_angular_radius);
-			const double reflector_phase_factor = dot(view_direction_icrf, blackbody_light_direction_icrf) * 0.5 + 0.5;
+			// Determine phase factor of reflector as seen by observer
+			const double observer_reflector_phase_factor = dot(observer_reflector_direction_eus, -reflector_blackbody_direction_eus) * 0.5 + 0.5;
 			
-			const double3 planet_luminance = (sunlight_illuminance * reference_body.albedo) / math::pi<double>;
-			const double planet_angular_radius = astro::angular_radius(reference_body.radius, reflector_distance);
-			const double planet_solid_angle = geom::solid_angle::cone(planet_angular_radius);
-			const double planet_phase_factor = math::dot(-view_direction_icrf, math::normalize(reference_orbit.position - blackbody_icrf_position)) * 0.5 + 0.5;
-			const double3 planetlight_illuminance = planet_luminance * planet_solid_angle * planet_phase_factor;
-			double3 planetlight_direction_eus = math::normalize(icrf_to_eus.r * view_direction_icrf);
+			// Measure observer reference body solid angle as seen by reflector
+			const double reflector_observer_angular_radius = astro::angular_radius(reference_body->radius, observer_reflector_distance);
+			const double reflector_observer_solid_angle = geom::solid_angle::cone(reflector_observer_angular_radius);
 			
-			const double3 reflected_sunlight_luminance = (sunlight_illuminance * reflector.albedo) / math::pi<double>;
-			const double3 reflected_sunlight_illuminance = reflected_sunlight_luminance * reflector_solid_angle * reflector_phase_factor;
-			const double3 reflected_planetlight_luminance = (planetlight_illuminance * reflector.albedo) / math::pi<double>;
-			const double3 reflected_planetlight_illuminance = reflected_planetlight_luminance * reflector_solid_angle;
+			// Determine phase factor of observer reference body as by reflector
+			const double reflector_observer_phase_factor = dot(-observer_reflector_direction_eus, -observer_blackbody_direction_eus) * 0.5 + 0.5;
 			
-			/*
-			std::cout << "reflected sunlight illuminance: " << reflected_sunlight_illuminance << std::endl;
-			std::cout << "planetlight illuminance: " << planetlight_illuminance << std::endl;
-			std::cout << "planet luminance: " << planet_luminance << std::endl;
-			std::cout << "reflected planetlight luminance: " << reflected_planetlight_luminance << std::endl;
-			std::cout << "reflected planetlight illuminance: " << reflected_planetlight_illuminance << std::endl;
-			std::cout << "reflector phase: " << reflector_phase_factor << std::endl;
-			std::cout << "planet phase: " << planet_phase_factor << std::endl;
-			*/
+			// Calculate spectral transmittance between observer and reflector factor due to atmospheric extinction
+			double3 observer_reflector_transmittance = {1, 1, 1};
+			if (reference_atmosphere)
+			{
+				const geom::ray<double> ray = {{0, 0, 0}, observer_reflector_direction_eus};
+				observer_reflector_transmittance = integrate_transmittance(*observer, *reference_body, *reference_atmosphere, ray);
+			}
 			
+			// Measure luminance of observer reference body as seen by reflector
+			const double3 reflector_observer_luminance = observer_blackbody_illuminance * reference_body->albedo * observer_reflector_transmittance * reflector_observer_phase_factor * math::inverse_pi<double>;
+			
+			// Measure illuminance from observer reference body reaching reflector
+			const double3 reflector_observer_illuminance = reflector_observer_luminance * reflector_observer_solid_angle;
+			
+			// Measure luminance of reflector as seen by observer
+			const double3 observer_reflector_luminance = (reflector_blackbody_illuminance * observer_reflector_phase_factor + reflector_observer_illuminance) * reflector.albedo * observer_reflector_transmittance * math::inverse_pi<double>;
+			
+			// Measure illuminance from reflector reaching observer
+			const double3 observer_reflector_illuminance = observer_reflector_luminance * observer_reflector_solid_angle;
 			
 			if (this->sky_pass)
 			{
 				this->sky_pass->set_moon_position(transform.local.translation);
 				this->sky_pass->set_moon_rotation(transform.local.rotation);
-				this->sky_pass->set_moon_angular_radius(static_cast<float>(reflector_angular_radius));
-				this->sky_pass->set_moon_sunlight_direction(math::type_cast<float>(blackbody_light_direction_eus));
-				this->sky_pass->set_moon_sunlight_illuminance(math::type_cast<float>(sunlight_illuminance));
-				this->sky_pass->set_moon_planetlight_direction(math::type_cast<float>(planetlight_direction_eus));
-				this->sky_pass->set_moon_planetlight_illuminance(math::type_cast<float>(planetlight_illuminance));
+				this->sky_pass->set_moon_angular_radius(static_cast<float>(observer_reflector_angular_radius));
+				this->sky_pass->set_moon_sunlight_direction(math::type_cast<float>(-reflector_blackbody_direction_eus));
+				this->sky_pass->set_moon_sunlight_illuminance(math::type_cast<float>(reflector_blackbody_illuminance * observer_reflector_transmittance));
+				this->sky_pass->set_moon_planetlight_direction(math::type_cast<float>(observer_reflector_direction_eus));
+				this->sky_pass->set_moon_planetlight_illuminance(math::type_cast<float>(reflector_observer_illuminance * observer_reflector_transmittance));
+				this->sky_pass->set_moon_illuminance(math::type_cast<float>(observer_reflector_illuminance / observer_reflector_transmittance), math::type_cast<float>(observer_reflector_illuminance));
 			}
 			
 			if (this->moon_light)
 			{
-				float3 reflector_up_eus = math::type_cast<float>(icrf_to_eus.r * double3{0, 0, 1});
-				
-				double3 reflected_illuminance = reflected_sunlight_illuminance + reflected_planetlight_illuminance;
-				//reflected_illuminance *= std::max<double>(0.0, std::sin(transform.local.translation.y));
-				
-				total_illuminance += (reflected_illuminance.x + reflected_illuminance.y + reflected_illuminance.z) / 3.0;
+				const float3 reflector_up_eus = math::type_cast<float>(icrf_to_eus.r * double3{0, 0, 1});
 				
-				this->moon_light->set_color(math::type_cast<float>(reflected_illuminance));
+				this->moon_light->set_color(math::type_cast<float>(observer_reflector_illuminance));
 				this->moon_light->set_rotation
 				(
 					math::look_rotation
 					(
-						math::normalize(-transform.local.translation),
+						math::type_cast<float>(-observer_reflector_direction_eus),
 						reflector_up_eus
 					)
 				);
+				
+				// Bounce moon light
+				bounce_illuminance += std::max(0.0, math::dot(bounce_normal, -observer_reflector_direction_eus)) * observer_reflector_illuminance * bounce_albedo;
 			}
-			
-			/*
-			std::cout << "moon: sun solid angle: " << blackbody_solid_angle << std::endl;
-			std::cout << "moon: sun illuminance: " << blackbody_illuminance << std::endl;
-			std::cout << "moon: moon luminance: " << reflector_luminance << std::endl;
-			std::cout << "sun brightness: " << sun_brightness << std::endl;
-			std::cout << "vega brightness: " << vega_brightness << std::endl;
-			std::cout << "earth: moon distance: " << reflector_distance << std::endl;
-			std::cout << "earth: moon solid angle: " << reflector_solid_angle << std::endl;
-			std::cout << "earth: moon phase: " << reflector_phase << std::endl;
-			std::cout << "earth: moon phase angle: " << math::degrees(reflector_phase_angle) << std::endl;
-			std::cout << "earth: moon illum %: " << reflector_illumination_factor * 100.0 << std::endl;
-			std::cout << "earth: moon illuminance: " << reflector_illuminance << std::endl;
-			std::cout << "earth: moon phase-modulated illuminance: " << reflector_illuminance * reflector_illumination_factor << std::endl;
-			*/
 		});
 	});
 	
-	// Update sky pass topocentric frame
-	if (sky_pass != nullptr)
+	if (bounce_light)
 	{
-		// Upload topocentric frame to sky pass
-		sky_pass->set_icrf_to_eus
-		(
-			math::transformation::se3<float>
-			{
-				math::type_cast<float>(icrf_to_eus.t),
-				math::type_cast<float>(icrf_to_eus.r)
-			}
-		);
-		
-		// Upload observer elevation to sky pass
-		sky_pass->set_observer_elevation(observer_location[0]);
-		
-		sky_pass->set_planet_radius(static_cast<float>(reference_body.radius));
-	}
-	
-	// Auto-exposure
-	if (camera)
-	{
-		const double calibration = 250.0;
-		const double ev100 = std::log2((total_illuminance * 100.0) / calibration);
-		//std::cout << "LUX: " << total_illuminance << std::endl;
-		//std::cout << "EV100: " << ev100 << std::endl;
-		//camera->set_exposure(exposure_offset);
+		bounce_light->set_color(math::type_cast<float>(bounce_illuminance));
 	}
 }
 
-void astronomy::set_time(double time)
+void astronomy::set_time(double t)
 {
-	this->time = time;
+	time_days = t;
+	time_centuries = time_days * physics::time::jd::centuries_per_day<double>;
 }
 
 void astronomy::set_time_scale(double scale)
@@ -438,16 +352,22 @@ void astronomy::set_time_scale(double scale)
 	time_scale = scale;
 }
 
-void astronomy::set_reference_body(entity::id entity_id)
+void astronomy::set_observer(entity::id eid)
 {
-	reference_entity = entity_id;
-	update_bcbf_to_enu();
+	if (observer_eid != eid)
+	{
+		observer_eid = eid;
+		
+		if (observer_eid != entt::null)
+			observer_modified();
+		else
+			reference_body_eid = entt::null;
+	}
 }
 
-void astronomy::set_observer_location(const double3& location)
+void astronomy::set_transmittance_samples(std::size_t samples)
 {
-	observer_location = location;
-	update_bcbf_to_enu();
+	transmittance_samples = samples;
 }
 
 void astronomy::set_sun_light(scene::directional_light* light)
@@ -465,17 +385,17 @@ void astronomy::set_moon_light(scene::directional_light* light)
 	moon_light = light;
 }
 
-void astronomy::set_camera(scene::camera* camera)
+void astronomy::set_bounce_light(scene::directional_light* light)
 {
-	this->camera = camera;
+	bounce_light = light;
 }
 
-void astronomy::set_exposure_offset(float offset)
+void astronomy::set_bounce_albedo(const double3& albedo)
 {
-	exposure_offset = offset;
+	bounce_albedo = albedo;
 }
 
-void astronomy::set_starlight_illuminance(double illuminance)
+void astronomy::set_starlight_illuminance(const double3& illuminance)
 {
 	starlight_illuminance = illuminance;
 }
@@ -483,37 +403,225 @@ void astronomy::set_starlight_illuminance(double illuminance)
 void astronomy::set_sky_pass(::render::sky_pass* pass)
 {
 	this->sky_pass = pass;
+	
+	if (sky_pass)
+	{
+		if (observer_eid != entt::null)
+		{
+			// Get pointer to observer
+			const auto observer = registry.try_get<component::observer>(reference_body_eid);
+			
+			sky_pass->set_observer_elevation(static_cast<float>(observer->elevation));
+		}
+		
+		if (reference_body_eid != entt::null)
+		{
+			// Get pointer to reference celestial body
+			const auto reference_body = registry.try_get<component::celestial_body>(reference_body_eid);
+			
+			if (reference_body)
+				sky_pass->set_planet_radius(static_cast<float>(reference_body->radius));
+			else
+				sky_pass->set_planet_radius(0.0f);
+		}
+	}
 }
 
-void astronomy::on_celestial_body_construct(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& celestial_body)
+void astronomy::on_observer_modified(entity::registry& registry, entity::id entity_id, entity::component::observer& component)
 {
-	if (entity_id == reference_entity)
-		update_bcbf_to_enu();
+	if (entity_id == observer_eid)
+		observer_modified();
 }
 
-void astronomy::on_celestial_body_replace(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& celestial_body)
+void astronomy::on_observer_destroyed(entity::registry& registry, entity::id entity_id)
 {
-	if (entity_id == reference_entity)
-		update_bcbf_to_enu();
+	if (entity_id == observer_eid)
+		observer_modified();
 }
 
-void astronomy::update_bcbf_to_enu()
+void astronomy::on_celestial_body_modified(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& component)
 {
-	double radial_distance = observer_location[0];
+	if (entity_id == reference_body_eid)
+		reference_body_modified();
+}
+
+void astronomy::on_celestial_body_destroyed(entity::registry& registry, entity::id entity_id)
+{
+	if (entity_id == reference_body_eid)
+		reference_body_modified();
+}
+
+void astronomy::on_orbit_modified(entity::registry& registry, entity::id entity_id, entity::component::orbit& component)
+{
+	if (entity_id == reference_body_eid)
+		reference_orbit_modified();
+}
+
+void astronomy::on_orbit_destroyed(entity::registry& registry, entity::id entity_id)
+{
+	if (entity_id == reference_body_eid)
+		reference_orbit_modified();
+}
+
+void astronomy::on_atmosphere_modified(entity::registry& registry, entity::id entity_id, entity::component::atmosphere& component)
+{
+	if (entity_id == reference_body_eid)
+		reference_atmosphere_modified();
+}
+
+void astronomy::on_atmosphere_destroyed(entity::registry& registry, entity::id entity_id)
+{
+	if (entity_id == reference_body_eid)
+		reference_atmosphere_modified();
+}
+
+void astronomy::observer_modified()
+{
+	// Get pointer to observer component
+	const auto observer = registry.try_get<component::observer>(observer_eid);
 	
-	if (reference_entity)
+	if (observer)
 	{
-		if (registry.has<entity::component::celestial_body>(reference_entity))
-			radial_distance += registry.get<entity::component::celestial_body>(reference_entity).radius;
+		if (reference_body_eid != observer->reference_body_eid)
+		{
+			// Reference body changed
+			reference_body_eid = observer->reference_body_eid;
+			reference_body_modified();
+			reference_orbit_modified();
+			reference_atmosphere_modified();
+		}
+		
+		if (reference_body_eid != entt::null)
+		{
+			// Get pointer to reference celestial body
+			const auto reference_body = registry.try_get<component::celestial_body>(reference_body_eid);
+			
+			// Update BCBF to EUS transformation
+			if (reference_body)
+				update_bcbf_to_eus(*observer, *reference_body);
+		}
+	
+		// Upload observer elevation to sky pass
+		if (sky_pass)
+			sky_pass->set_observer_elevation(static_cast<float>(observer->elevation));
+	}
+}
+
+void astronomy::reference_body_modified()
+{
+	// Get pointer to reference celestial body
+	const auto reference_body = registry.try_get<component::celestial_body>(reference_body_eid);
+	
+	if (reference_body)
+	{
+		// Get pointer to observer
+		const auto observer = registry.try_get<component::observer>(observer_eid);
+		
+		// Update BCBF to EUS transformation
+		if (observer)
+			update_bcbf_to_eus(*observer, *reference_body);
 	}
 	
-	// Construct reference frame which transforms coordinates from a BCBF frame to a horizontal frame
-	bcbf_to_enu = physics::orbit::frame::bcbf::to_enu
+	// Update reference celestial body-related sky pass parameters
+	if (sky_pass)
+	{
+		if (reference_body)
+			sky_pass->set_planet_radius(static_cast<float>(reference_body->radius));
+		else
+			sky_pass->set_planet_radius(0.0f);
+	}
+}
+
+void astronomy::reference_orbit_modified()
+{
+	
+}
+
+void astronomy::reference_atmosphere_modified()
+{
+
+}
+
+void astronomy::update_bcbf_to_eus(const entity::component::observer& observer, const entity::component::celestial_body& body)
+{
+	// Construct BCBF to EUS transformation
+	bcbf_to_eus = physics::orbit::frame::bcbf::to_enu
 	(
-		radial_distance,
-		observer_location[1],
-		observer_location[2]
+		body.radius + observer.elevation,
+		observer.latitude,
+		observer.longitude
+	) * enu_to_eus;
+}
+
+void astronomy::update_icrf_to_eus(const entity::component::celestial_body& body, const entity::component::orbit& orbit)
+{
+	// Evaluate reference body orientation polynomials
+	const double body_pole_ra = math::polynomial::horner(body.pole_ra.begin(), body.pole_ra.end(), time_centuries);
+	const double body_pole_dec = math::polynomial::horner(body.pole_dec.begin(), body.pole_dec.end(), time_centuries);
+	const double body_prime_meridian = math::polynomial::horner(body.prime_meridian.begin(), body.prime_meridian.end(), time_days);
+	
+	// Construct ICRF frame to BCBF transformation
+	math::transformation::se3<double> icrf_to_bcbf = physics::orbit::frame::bci::to_bcbf
+	(
+		body_pole_ra,
+		body_pole_dec,
+		body_prime_meridian
 	);
+	icrf_to_bcbf.t = icrf_to_bcbf.r * -orbit.position;
+	
+	/// Construct ICRF to EUS transformation
+	icrf_to_eus = icrf_to_bcbf * bcbf_to_eus;
+	
+	// Pass ICRF to EUS transformation to sky pass
+	if (sky_pass)
+	{
+		// Upload topocentric frame to sky pass
+		sky_pass->set_icrf_to_eus
+		(
+			math::transformation::se3<float>
+			{
+				math::type_cast<float>(icrf_to_eus.t),
+				math::type_cast<float>(icrf_to_eus.r)
+			}
+		);
+	}
+}
+
+double3 astronomy::integrate_transmittance(const entity::component::observer& observer, const entity::component::celestial_body& body, const entity::component::atmosphere& atmosphere, geom::ray<double> ray) const
+{
+	double3 transmittance = {1, 1, 1};
+	
+	// Make ray height relative to center of reference body
+	ray.origin.y += body.radius + observer.elevation;
+	
+	// Construct sphere representing upper limit of the atmosphere
+	geom::sphere<double> atmosphere_sphere;
+	atmosphere_sphere.center = {0, 0, 0};
+	atmosphere_sphere.radius = body.radius + atmosphere.upper_limit;
+	
+	// Check for intersection between the ray and atmosphere
+	auto intersection = geom::ray_sphere_intersection(ray, atmosphere_sphere);
+	if (std::get<0>(intersection))
+	{
+		// Get point of intersection
+		const double3 intersection_point = ray.extrapolate(std::get<2>(intersection));
+		
+		// Integrate optical of Rayleigh, Mie, and ozone particles
+		const double optical_depth_r = physics::gas::atmosphere::optical_depth_exp(ray.origin, intersection_point, body.radius, atmosphere.rayleigh_scale_height, transmittance_samples);
+		const double optical_depth_m = physics::gas::atmosphere::optical_depth_exp(ray.origin, intersection_point, body.radius, atmosphere.mie_scale_height, transmittance_samples);
+		const double optical_depth_o = physics::gas::atmosphere::optical_depth_tri(ray.origin, intersection_point, body.radius, atmosphere.ozone_lower_limit, atmosphere.ozone_upper_limit, atmosphere.ozone_mode, transmittance_samples);
+		
+		// Calculate transmittance factor due to scattering and absorption
+		const double3 extinction_r = atmosphere.rayleigh_scattering * optical_depth_r;
+		const double extinction_m = atmosphere.mie_extinction * optical_depth_m;
+		const double3 extinction_o = atmosphere.ozone_absorption * optical_depth_o;
+		transmittance = extinction_r + double3{extinction_m, extinction_m, extinction_m} + extinction_o;
+		transmittance.x = std::exp(-transmittance.x);
+		transmittance.y = std::exp(-transmittance.y);
+		transmittance.z = std::exp(-transmittance.z);
+	}
+	
+	return transmittance;
 }
 
 } // namespace system
diff --git a/src/entity/systems/astronomy.hpp b/src/entity/systems/astronomy.hpp
index 83599f4..db08a94 100644
--- a/src/entity/systems/astronomy.hpp
+++ b/src/entity/systems/astronomy.hpp
@@ -24,19 +24,20 @@
 #include "entity/id.hpp"
 #include "scene/directional-light.hpp"
 #include "scene/ambient-light.hpp"
-#include "scene/camera.hpp"
 #include "utility/fundamental-types.hpp"
 #include "math/se3.hpp"
 #include "render/passes/sky-pass.hpp"
+#include "entity/components/observer.hpp"
 #include "entity/components/atmosphere.hpp"
 #include "entity/components/celestial-body.hpp"
 #include "entity/components/orbit.hpp"
+#include "geom/ray.hpp"
 
 namespace entity {
 namespace system {
 
 /**
- * Calculates apparent properties of celestial bodies relative to an observer.
+ * Calculates apparent properties of celestial bodies as seen by an observer.
  */
 class astronomy:
 	public updatable
@@ -46,7 +47,7 @@ public:
 	~astronomy();
 	
 	/**
-	 * Scales then adds the timestep `dt` to the current time, then recalculates the positions of celestial bodies.
+	 * Adds the timestep `dt`, scaled by set time scale, to the current time, then calculates apparent properties of celestial bodies as seen by an observer.
 	 *
 	 * @param t Time, in seconds.
 	 * @param dt Delta time, in seconds.
@@ -56,9 +57,9 @@ public:
 	/**
 	 * Sets the current time.
 	 *
-	 * @param time Time, in days.
+	 * @param t Time since epoch, in days.
 	 */
-	void set_time(double time);
+	void set_time(double t);
 	
 	/**
 	 * Sets the factor by which the timestep `dt` will be scaled before being added to the current time.
@@ -68,57 +69,99 @@ public:
 	void set_time_scale(double scale);
 	
 	/**
-	 * Sets the reference body entity, from which observations are taking place.
+	 * Sets the observer entity.
 	 *
-	 * @param entity Entity of the reference body.
+	 * @param eid Entity ID of the observer.
 	 */
-	void set_reference_body(entity::id entity_id);
+	void set_observer(entity::id eid);
 	
 	/**
-	 * Sets the location of the observer using spherical coordinates in BCBF space.
+	 * Sets the number of samples to take when integrating atmospheric transmittance.
 	 *
-	 * @param location Spherical coordinates of the observer, in reference body BCBF space, in the ISO order of altitude (meters), latitude (radians), and longitude (radians).
+	 * @param samples Number of integration samples.
 	 */
-	void set_observer_location(const double3& location);
+	void set_transmittance_samples(std::size_t samples);
 	
 	void set_sun_light(scene::directional_light* light);
 	void set_sky_light(scene::ambient_light* light);
 	void set_moon_light(scene::directional_light* light);
-	void set_camera(scene::camera* camera);
-	void set_exposure_offset(float offset);
+	void set_bounce_light(scene::directional_light* light);
+	void set_bounce_albedo(const double3& albedo);
+	void set_starlight_illuminance(const double3& illuminance);
+	void set_sky_pass(::render::sky_pass* pass);
 	
-	void set_starlight_illuminance(double illuminance);
+private:
+	void on_observer_modified(entity::registry& registry, entity::id entity_id, entity::component::observer& component);
+	void on_observer_destroyed(entity::registry& registry, entity::id entity_id);
+	void on_celestial_body_modified(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& component);
+	void on_celestial_body_destroyed(entity::registry& registry, entity::id entity_id);
+	void on_orbit_modified(entity::registry& registry, entity::id entity_id, entity::component::orbit& component);
+	void on_orbit_destroyed(entity::registry& registry, entity::id entity_id);
+	void on_atmosphere_modified(entity::registry& registry, entity::id entity_id, entity::component::atmosphere& component);
+	void on_atmosphere_destroyed(entity::registry& registry, entity::id entity_id);
 	
-	inline float get_exposure_offset() const { return exposure_offset; };
+	/// Called each time the observer is modified.
+	void observer_modified();
 	
-	void set_sky_pass(::render::sky_pass* pass);
+	/// Called each time the celestial body of the reference body is modified.
+	void reference_body_modified();
 	
-private:
-	void on_celestial_body_construct(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& celestial_body);
-	void on_celestial_body_replace(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& celestial_body);
+	/// Called each time the orbit of the reference body is modified.
+	void reference_orbit_modified();
 	
-	void update_bcbf_to_enu();
+	/// Called each time the atmosphere of the reference body is modified.
+	void reference_atmosphere_modified();
 
-	double time;
+	/// Updates the BCBF to EUS transformation.
+	void update_bcbf_to_eus(const entity::component::observer& observer, const entity::component::celestial_body& body);
+	
+	/// Updates the ICRF to EUS transformation.
+	void update_icrf_to_eus(const entity::component::celestial_body& body, const entity::component::orbit& orbit);
+	
+	/**
+	 * Integrates a transmittance factor due to atmospheric extinction along a ray.
+	 *
+	 * @param ray Ray to cast, in the EUS frame.
+	 * @param samples Number of samples to integrate.
+	 *
+	 * @return Spectral transmittance factor.
+	 */
+	double3 integrate_transmittance(const entity::component::observer& observer, const entity::component::celestial_body& body, const entity::component::atmosphere& atmosphere, geom::ray<double> ray) const;
+	
+	/// Time since epoch, in days.
+	double time_days;
+	
+	/// Time since epoch, in centuries.
+	double time_centuries;
+	
+	/// Time scale.
 	double time_scale;
 	
-	entity::id reference_entity;
+	/// Number of transmittance integration samples.
+	std::size_t transmittance_samples;
+	
+	/// Entity ID of the observer.
+	entity::id observer_eid;
 	
-	double3 observer_location;
+	/// Entity ID of the reference body.
+	entity::id reference_body_eid;
 	
-	math::transformation::se3<double> icrf_to_bcbf;
-	math::transformation::se3<double> bcbf_to_enu;
-	math::transformation::se3<double> icrf_to_enu;
+	/// ENU to EUS transformation.
 	math::transformation::se3<double> enu_to_eus;
+	
+	/// BCBF to EUS transformation.
+	math::transformation::se3<double> bcbf_to_eus;
+	
+	/// ICRF to EUS tranformation.
 	math::transformation::se3<double> icrf_to_eus;
 	
 	scene::directional_light* sun_light;
 	scene::ambient_light* sky_light;
 	scene::directional_light* moon_light;
-	scene::camera* camera;
+	scene::directional_light* bounce_light;
+	double3 bounce_albedo;
 	::render::sky_pass* sky_pass;
-	float exposure_offset;
-	double starlight_illuminance;
+	double3 starlight_illuminance;
 };
 
 } // namespace system
diff --git a/src/entity/systems/atmosphere.cpp b/src/entity/systems/atmosphere.cpp
index 78ac0e1..4a16002 100644
--- a/src/entity/systems/atmosphere.cpp
+++ b/src/entity/systems/atmosphere.cpp
@@ -21,7 +21,6 @@
 #include "physics/gas/atmosphere.hpp"
 #include "physics/gas/ozone.hpp"
 #include "physics/number-density.hpp"
-#include "color/srgb.hpp"
 
 namespace entity {
 namespace system {
@@ -44,12 +43,15 @@ void atmosphere::update(double t, double dt)
 void atmosphere::set_rgb_wavelengths(const double3& wavelengths)
 {
 	rgb_wavelengths = wavelengths;
-	atmosphere_modified();
-}
-
-void atmosphere::set_rgb_ozone_cross_sections(const double3& cross_sections)
-{
-	rgb_ozone_cross_sections = cross_sections;
+	
+	// Update ozone cross sections
+	rgb_ozone_cross_sections =
+	{
+		physics::gas::ozone::cross_section_293k<double>(wavelengths.x * 1e9),
+		physics::gas::ozone::cross_section_293k<double>(wavelengths.y * 1e9),
+		physics::gas::ozone::cross_section_293k<double>(wavelengths.z * 1e9)
+	};
+	
 	atmosphere_modified();
 }
 
@@ -67,37 +69,31 @@ void atmosphere::atmosphere_modified()
 	// Get atmosphere component of the entity
 	entity::component::atmosphere& component = *atmosphere_component;
 	
-	// Calculate Rayleigh scattering coefficients for sRGB wavelengths
-	const double rayleigh_polarization = physics::gas::atmosphere::polarization(component.index_of_refraction, component.rayleigh_density);
-	const double3 rayleigh_scattering_srgb =
+	// Calculate Rayleigh scattering coefficients
+	const double rayleigh_density = physics::number_density(component.rayleigh_concentration);
+	const double rayleigh_polarization = physics::gas::atmosphere::polarization(component.index_of_refraction, rayleigh_density);
+	component.rayleigh_scattering =
 	{
-		physics::gas::atmosphere::scattering_rayleigh(rgb_wavelengths.x, component.rayleigh_density, rayleigh_polarization),
-		physics::gas::atmosphere::scattering_rayleigh(rgb_wavelengths.y, component.rayleigh_density, rayleigh_polarization),
-		physics::gas::atmosphere::scattering_rayleigh(rgb_wavelengths.z, component.rayleigh_density, rayleigh_polarization)
+		physics::gas::atmosphere::scattering(rayleigh_density, rayleigh_polarization, rgb_wavelengths.x),
+		physics::gas::atmosphere::scattering(rayleigh_density, rayleigh_polarization, rgb_wavelengths.y),
+		physics::gas::atmosphere::scattering(rayleigh_density, rayleigh_polarization, rgb_wavelengths.z)
 	};
 	
-	// Transform Rayleigh scattering coefficients from sRGB to ACEScg
-	component.rayleigh_scattering = color::srgb::to_acescg(rayleigh_scattering_srgb);
-	
-	// Calculate Mie scattering coefficient
-	const double mie_polarization = physics::gas::atmosphere::polarization(component.index_of_refraction, component.mie_density);
-	component.mie_scattering = physics::gas::atmosphere::scattering_mie(component.mie_density, mie_polarization);
+	// Calculate Mie scattering and extinction coefficients
+	const double mie_density = physics::number_density(component.mie_concentration);
+	const double mie_polarization = physics::gas::atmosphere::polarization(component.index_of_refraction, mie_density);
+	component.mie_scattering = physics::gas::atmosphere::scattering(mie_density, mie_polarization);
+	component.mie_extinction = physics::gas::atmosphere::extinction(component.mie_scattering, component.mie_albedo);
 	
-	// Calculate Mie absorption coefficient
-	component.mie_absorption = physics::gas::atmosphere::absorption_mie(component.mie_scattering);
-	
-	// Calculate ozone absorption coefficients for sRGB wavelengths
-	const double air_number_density = physics::number_density(component.air_concentration);
-	const double3 ozone_absorption_srgb =
+	// Calculate ozone absorption coefficients
+	const double ozone_density = physics::number_density(component.ozone_concentration);
+	component.ozone_absorption =
 	{
-		physics::gas::ozone::absorption(rgb_ozone_cross_sections.x, air_number_density, component.ozone_concentration),
-		physics::gas::ozone::absorption(rgb_ozone_cross_sections.y, air_number_density, component.ozone_concentration),
-		physics::gas::ozone::absorption(rgb_ozone_cross_sections.z, air_number_density, component.ozone_concentration)
+		physics::gas::ozone::absorption(rgb_ozone_cross_sections.x, ozone_density),
+		physics::gas::ozone::absorption(rgb_ozone_cross_sections.y, ozone_density),
+		physics::gas::ozone::absorption(rgb_ozone_cross_sections.z, ozone_density)
 	};
 	
-	// Transform ozone absorption coefficients from sRGB to ACEScg
-	component.ozone_absorption = color::srgb::to_acescg(ozone_absorption_srgb);
-	
 	// Pass atmosphere parameters to sky pass
 	update_sky_pass();
 }
@@ -111,8 +107,9 @@ void atmosphere::update_sky_pass()
 	
 	sky_pass->set_atmosphere_upper_limit(static_cast<float>(component.upper_limit));
 	sky_pass->set_rayleigh_parameters(static_cast<float>(component.rayleigh_scale_height), math::type_cast<float>(component.rayleigh_scattering));
-	sky_pass->set_mie_parameters(static_cast<float>(component.mie_scale_height), static_cast<float>(component.mie_scattering), static_cast<float>(component.mie_absorption), static_cast<float>(component.mie_anisotropy));
+	sky_pass->set_mie_parameters(static_cast<float>(component.mie_scale_height), static_cast<float>(component.mie_scattering), static_cast<float>(component.mie_extinction), static_cast<float>(component.mie_anisotropy));
 	sky_pass->set_ozone_parameters(static_cast<float>(component.ozone_lower_limit), static_cast<float>(component.ozone_upper_limit), static_cast<float>(component.ozone_mode), math::type_cast<float>(component.ozone_absorption));
+	sky_pass->set_airglow_illuminance(math::type_cast<float>(component.airglow_illuminance));
 }
 
 void atmosphere::on_atmosphere_construct(entity::registry& registry, entity::id entity_id, entity::component::atmosphere& component)
diff --git a/src/entity/systems/atmosphere.hpp b/src/entity/systems/atmosphere.hpp
index 1557200..2ca28e7 100644
--- a/src/entity/systems/atmosphere.hpp
+++ b/src/entity/systems/atmosphere.hpp
@@ -47,13 +47,6 @@ public:
 	 */
 	void set_rgb_wavelengths(const double3& wavelengths);
 	
-	/**
-	 * Sets ozone cross sections for red, green, and blue wavelengths.
-	 *
-	 * @param cross_sections Vector containing the ozone cross sections of red (x), green (y), and blue (z) light, in m-2/molecule.
-	 */
-	void set_rgb_ozone_cross_sections(const double3& cross_sections);
-	
 	void set_sky_pass(::render::sky_pass* pass);
 	
 private:
diff --git a/src/entity/systems/blackbody.cpp b/src/entity/systems/blackbody.cpp
index 5eb4cd4..7ba8fd7 100644
--- a/src/entity/systems/blackbody.cpp
+++ b/src/entity/systems/blackbody.cpp
@@ -27,7 +27,8 @@ namespace entity {
 namespace system {
 
 blackbody::blackbody(entity::registry& registry):
-	updatable(registry)
+	updatable(registry),
+	illuminant(color::illuminant::deg2::d50<double>)
 {
 	// Construct a range of sample wavelengths in the visible spectrum
 	visible_wavelengths_nm.resize(780 - 280);
@@ -43,6 +44,11 @@ blackbody::blackbody(entity::registry& registry):
 void blackbody::update(double t, double dt)
 {}
 
+void blackbody::set_illuminant(const math::vector2<double>& illuminant)
+{
+	this->illuminant = illuminant;
+}
+
 void blackbody::update_luminance(entity::id entity_id)
 {
 	// Abort if entity has no blackbody component
@@ -62,8 +68,11 @@ void blackbody::update_luminance(entity::id entity_id)
 	// Get celestial body component of the entity
 	const component::celestial_body& celestial_body = registry.get<component::celestial_body>(entity_id);
 	
+	// Construct chromatic adaptation transform
+	const double3x3 cat = color::cat::matrix(illuminant, color::aces::white_point<double>);
+	
 	// Construct a lambda function which calculates the blackbody's RGB luminance of a given wavelength
-	auto rgb_luminance = [blackbody](double wavelength_nm) -> double3
+	auto rgb_luminance = [blackbody, cat](double wavelength_nm) -> double3
 	{
 		// Convert wavelength from nanometers to meters
 		const double wavelength_m = wavelength_nm * 1e-9;
@@ -71,8 +80,9 @@ void blackbody::update_luminance(entity::id entity_id)
 		// Calculate the spectral intensity of the wavelength
 		const double spectral_radiance = physics::light::blackbody::spectral_radiance<double>(blackbody.temperature, wavelength_m);
 		
+		
 		// Calculate the ACEScg color of the wavelength using CIE color matching functions
-		double3 spectral_color = color::xyz::to_acescg(color::xyz::match(wavelength_nm));
+		double3 spectral_color = color::aces::ap1<double>.from_xyz * cat * color::xyz::match(wavelength_nm);
 		
 		// Scale the spectral color by spectral intensity
 		return spectral_color * spectral_radiance * 1e-9 * physics::light::max_luminous_efficacy<double>;
diff --git a/src/entity/systems/blackbody.hpp b/src/entity/systems/blackbody.hpp
index 7512398..b38eb75 100644
--- a/src/entity/systems/blackbody.hpp
+++ b/src/entity/systems/blackbody.hpp
@@ -41,6 +41,13 @@ public:
 	
 	virtual void update(double t, double dt);
 	
+	/**
+	 * Sets the blackbody illuminant.
+	 *
+	 * @param illuminant CIE chromaticity coordinates of an illuminant.
+	 */
+	void set_illuminant(const math::vector2<double>& illuminant);
+	
 private:
 	void update_luminance(entity::id entity_id);
 	
@@ -50,6 +57,7 @@ private:
 	void on_celestial_body_construct(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& celestial_body);
 	void on_celestial_body_replace(entity::registry& registry, entity::id entity_id, entity::component::celestial_body& celestial_body);
 	
+	math::vector2<double> illuminant;
 	std::vector<double> visible_wavelengths_nm;
 };
 
diff --git a/src/entity/systems/terrain.cpp b/src/entity/systems/terrain.cpp
index 0eb9ee7..0664242 100644
--- a/src/entity/systems/terrain.cpp
+++ b/src/entity/systems/terrain.cpp
@@ -72,6 +72,7 @@ terrain::~terrain()
 
 void terrain::update(double t, double dt)
 {
+	/*
 	// Refine the level of detail of each terrain quadsphere
 	registry.view<component::terrain, component::celestial_body>().each(
 	[&](entity::id terrain_eid, const auto& terrain_component, const auto& terrain_body)
@@ -238,6 +239,7 @@ void terrain::update(double t, double dt)
 			}
 		}
 	});
+	*/
 }
 
 void terrain::set_patch_subdivisions(std::uint8_t n)
diff --git a/src/game/ant/swarm.cpp b/src/game/ant/swarm.cpp
new file mode 100644
index 0000000..79b6c6b
--- /dev/null
+++ b/src/game/ant/swarm.cpp
@@ -0,0 +1,147 @@
+/*
+ * Copyright (C) 2021  Christopher J. Howard
+ *
+ * This file is part of Antkeeper source code.
+ *
+ * Antkeeper source code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * Antkeeper source code is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Antkeeper source code.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "game/ant/swarm.hpp"
+#include "entity/components/transform.hpp"
+#include "entity/components/steering.hpp"
+#include "entity/components/model.hpp"
+#include "resources/resource-manager.hpp"
+#include "config.hpp"
+#include <cmath>
+#include <random>
+
+namespace game {
+namespace ant {
+
+/**
+ * Generates a random point in a unit sphere.
+ *
+ * @see https://math.stackexchange.com/questions/87230/picking-random-points-in-the-volume-of-sphere-with-uniform-probability/87238#87238
+ */
+template <class T, class Generator>
+static math::vector3<T> sphere_random(Generator& rng)
+{
+	const std::uniform_real_distribution<T> distribution(T{-1}, T{1});
+	
+	const T x = distribution(rng);
+	const T y = distribution(rng);
+	const T z = distribution(rng);
+	
+	const T cbrt_u = std::cbrt(distribution(rng));
+	
+	const T scale = (T{1} / std::sqrt(x * x + y * y + z * z)) * cbrt_u;
+	
+	return {x * scale, y * scale, z * scale};
+}
+
+void create_swarm(game::context& ctx)
+{
+	// Determine swarm properties
+	const float3 swarm_center = {0, 100, 0};
+	const float swarm_radius = 25.0f;
+	const std::size_t male_count = 50;
+	const std::size_t queen_count = 50;
+	const std::size_t alate_count = male_count + queen_count;
+	
+	const float3 male_scale = {0.5, 0.5, 0.5};
+	const float3 queen_scale = {1, 1, 1};
+	
+	// Init male model component
+	entity::component::model male_model;
+	male_model.render_model = ctx.resource_manager->load<render::model>("male-boid.mdl");
+	male_model.instance_count = 0;
+	male_model.layers = 1;
+	
+	// Init queen model component
+	entity::component::model queen_model;
+	queen_model.render_model = ctx.resource_manager->load<render::model>("queen-boid.mdl");
+	queen_model.instance_count = 0;
+	queen_model.layers = 1;
+	
+	// Init transform component
+	entity::component::transform transform;
+	transform.local = math::transform<float>::identity;
+	transform.world = transform.local;
+	transform.warp = true;
+	
+	// Init steering component
+	entity::component::steering steering;
+	steering.agent.mass = 1.0f;
+	steering.agent.velocity = {0, 0, 0};
+	steering.agent.acceleration = {0, 0, 0};
+	steering.agent.max_force = 4.0f;
+	steering.agent.max_speed = 5.0f;
+	steering.agent.max_speed_squared = steering.agent.max_speed * steering.agent.max_speed;
+	steering.agent.orientation = math::quaternion<float>::identity;
+	steering.agent.forward = steering.agent.orientation * config::global_forward;
+	steering.agent.up = steering.agent.orientation * config::global_up;
+	steering.wander_weight = 1.0f;
+	steering.wander_noise = math::radians(2000.0f);
+	steering.wander_distance = 10.0f;
+	steering.wander_radius = 8.0f;
+	steering.wander_angle = 0.0f;
+	steering.wander_angle2 = 0.0f;
+	steering.seek_weight = 0.2f;
+	steering.seek_target = {0, 100, 0};
+	steering.flee_weight = 0.0f;
+	steering.sum_weights = steering.wander_weight + steering.seek_weight + steering.flee_weight;
+	
+	// Construct and seed random number generator
+	std::random_device seed;
+    std::mt19937 rng(seed());
+	
+	// Create alates
+	for (std::size_t i = 0; i < alate_count; ++i)
+	{
+		// Generate random position in swarm sphere
+		steering.agent.position = swarm_center + sphere_random<float>(rng) * swarm_radius;
+		transform.local.translation = steering.agent.position;
+		
+		entity::id alate_eid = ctx.entity_registry->create();
+		ctx.entity_registry->assign<entity::component::steering>(alate_eid, steering);
+		
+		if (i < male_count)
+		{
+			// Create male
+			ctx.entity_registry->assign<entity::component::model>(alate_eid, male_model);
+			
+			transform.local.scale = male_scale;
+			transform.world = transform.local;
+			ctx.entity_registry->assign<entity::component::transform>(alate_eid, transform);
+		}
+		else
+		{
+			// Create queen
+			ctx.entity_registry->assign<entity::component::model>(alate_eid, queen_model);
+			
+			transform.local.scale = queen_scale;
+			transform.world = transform.local;
+			ctx.entity_registry->assign<entity::component::transform>(alate_eid, transform);
+		}
+	}
+}
+
+void destroy_swarm(game::context& ctx)
+{
+	auto view = ctx.entity_registry->view<entity::component::steering>();
+	ctx.entity_registry->destroy(view.begin(), view.end());
+}
+
+} // namespace ant
+} // namespace game
diff --git a/src/game/ant/swarm.hpp b/src/game/ant/swarm.hpp
new file mode 100644
index 0000000..3b60400
--- /dev/null
+++ b/src/game/ant/swarm.hpp
@@ -0,0 +1,34 @@
+/*
+ * Copyright (C) 2021  Christopher J. Howard
+ *
+ * This file is part of Antkeeper source code.
+ *
+ * Antkeeper source code is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * Antkeeper source code is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with Antkeeper source code.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef ANTKEEPER_GAME_ANT_SWARM_HPP
+#define ANTKEEPER_GAME_ANT_SWARM_HPP
+
+#include "game/context.hpp"
+
+namespace game {
+namespace ant {
+
+void create_swarm(game::context& ctx);
+void destroy_swarm(game::context& ctx);
+
+} // namespace ant
+} // namespace game
+
+#endif // ANTKEEPER_GAME_ANT_SWARM_HPP
diff --git a/src/game/context.hpp b/src/game/context.hpp
index 91c572c..d4e7488 100644
--- a/src/game/context.hpp
+++ b/src/game/context.hpp
@@ -295,9 +295,7 @@ struct context
 	entity::system::orbit* orbit_system;
 	entity::system::proteome* proteome_system;
 	
-	double elevation;
-	double latitude;
-	double longitude;
+	double3 rgb_wavelengths;
 };
 
 } // namespace game
diff --git a/src/game/controls.cpp b/src/game/controls.cpp
index 8db64ac..bb127d2 100644
--- a/src/game/controls.cpp
+++ b/src/game/controls.cpp
@@ -660,4 +660,14 @@ void apply_gamepad_calibration(input::gamepad& gamepad, const json& calibration)
 	}
 }
 
+void enable_orbit_controls(game::context& ctx)
+{
+
+}
+
+void disable_orbit_controls(game::context& ctx)
+{
+	
+}
+
 } // namespace game
diff --git a/src/game/controls.hpp b/src/game/controls.hpp
index 37f817c..6479df9 100644
--- a/src/game/controls.hpp
+++ b/src/game/controls.hpp
@@ -87,6 +87,9 @@ bool save_gamepad_calibration(const game::context& ctx, const input::gamepad& ga
  */
 void apply_gamepad_calibration(input::gamepad& gamepad, const json& calibration);
 
+void enable_orbit_controls(game::context& ctx);
+void disable_orbit_controls(game::context& ctx);
+
 } // namespace game
 
 #endif // ANTKEEPER_GAME_CONTROLS_HPP
diff --git a/src/game/load.cpp b/src/game/load.cpp
index c5d6000..89a60d0 100644
--- a/src/game/load.cpp
+++ b/src/game/load.cpp
@@ -86,6 +86,19 @@ void biome(game::context& ctx, const std::filesystem::path& path)
 			{
 				ctx.logger->warning("Biome terrain material undefined");
 			}
+			
+			// Setup lighting
+			double3 terrain_albedo = {0, 0, 0};
+			if (terrain->contains("albedo"))
+			{
+				const auto& albedo_element = (*terrain)["albedo"];
+				
+				terrain_albedo[0] = albedo_element[0].get<double>();
+				terrain_albedo[1]= albedo_element[1].get<double>();
+				terrain_albedo[2] = albedo_element[2].get<double>();
+				
+			}
+			ctx.astronomy_system->set_bounce_albedo(terrain_albedo);
 		}
 		else
 		{
diff --git a/src/game/state/boot.cpp b/src/game/state/boot.cpp
index 39d33f4..0f88ef3 100644
--- a/src/game/state/boot.cpp
+++ b/src/game/state/boot.cpp
@@ -93,7 +93,6 @@
 #include "game/graphics.hpp"
 #include "utility/timestamp.hpp"
 #include "color/color.hpp"
-#include "physics/gas/ozone.hpp"
 #include <cxxopts.hpp>
 #include <entt/entt.hpp>
 #include <filesystem>
@@ -834,20 +833,6 @@ void boot::setup_systems()
 	const auto& viewport_dimensions = ctx.app->get_viewport_dimensions();
 	float4 viewport = {0.0f, 0.0f, static_cast<float>(viewport_dimensions[0]), static_cast<float>(viewport_dimensions[1])};
 	
-	// RGB wavelengths
-	const double3 rgb_wavelengths_nm = {680, 550, 440};
-	const double3 rgb_wavelengths_m = rgb_wavelengths_nm * 1e-9;
-	
-	// Ozone cross sections
-	double3 rgb_ozone_cross_sections_293k =
-	{
-		physics::gas::ozone::cross_section_680nm_293k<double>,
-		physics::gas::ozone::cross_section_550nm_293k<double>,
-		physics::gas::ozone::cross_section_440nm_293k<double>
-	};
-	
-	rgb_ozone_cross_sections_293k = rgb_ozone_cross_sections_293k;
-	
 	// Setup terrain system
 	ctx.terrain_system = new entity::system::terrain(*ctx.entity_registry);
 	ctx.terrain_system->set_patch_subdivisions(30);
@@ -904,15 +889,19 @@ void boot::setup_systems()
 	
 	// Setup blackbody system
 	ctx.blackbody_system = new entity::system::blackbody(*ctx.entity_registry);
+	ctx.blackbody_system->set_illuminant(color::illuminant::deg2::d55<double>);
+	
+	// RGB wavelengths for atmospheric scatteering
+	ctx.rgb_wavelengths = {645, 575, 440};
 	
 	// Setup atmosphere system
 	ctx.atmosphere_system = new entity::system::atmosphere(*ctx.entity_registry);
-	ctx.atmosphere_system->set_rgb_wavelengths(rgb_wavelengths_m);
-	ctx.atmosphere_system->set_rgb_ozone_cross_sections(rgb_ozone_cross_sections_293k);
+	ctx.atmosphere_system->set_rgb_wavelengths(ctx.rgb_wavelengths * 1e-9);
 	ctx.atmosphere_system->set_sky_pass(ctx.sky_pass);
 	
 	// Setup astronomy system
 	ctx.astronomy_system = new entity::system::astronomy(*ctx.entity_registry);
+	ctx.astronomy_system->set_transmittance_samples(16);
 	ctx.astronomy_system->set_sky_pass(ctx.sky_pass);
 	
 	// Setup proteome system
diff --git a/src/game/state/main-menu.cpp b/src/game/state/main-menu.cpp
index 62e67d9..e8a60b1 100644
--- a/src/game/state/main-menu.cpp
+++ b/src/game/state/main-menu.cpp
@@ -21,8 +21,13 @@
 #include "game/state/options-menu.hpp"
 #include "game/state/extras-menu.hpp"
 #include "game/state/nuptial-flight.hpp"
+#include "game/world.hpp"
+#include "game/load.hpp"
 #include "game/menu.hpp"
+#include "game/ant/swarm.hpp"
 #include "render/passes/clear-pass.hpp"
+#include "render/passes/ground-pass.hpp"
+#include "render/passes/sky-pass.hpp"
 #include "resources/resource-manager.hpp"
 #include "render/model.hpp"
 #include "animation/animation.hpp"
@@ -31,6 +36,10 @@
 #include "animation/ease.hpp"
 #include "application.hpp"
 #include "config.hpp"
+#include "physics/light/exposure.hpp"
+#include "entity/components/model.hpp"
+#include "entity/components/steering.hpp"
+#include "entity/components/transform.hpp"
 #include <limits>
 
 namespace game {
@@ -233,6 +242,40 @@ main_menu::main_menu(game::context& ctx, bool fade_in):
 		game::menu::fade_in(ctx, nullptr);
 	}
 	
+	if (ctx.entities.find("earth") == ctx.entities.end())
+	{
+		game::world::cosmogenesis(ctx);
+		game::world::create_observer(ctx);
+		game::load::biome(ctx, "debug.bio");
+	}
+	
+	// Set world time
+	game::world::set_time(ctx, 2022, 6, 21, 12, 0, 0.0);
+	
+	// Set world time scale
+	game::world::set_time_scale(ctx, 0.0);
+	
+	ctx.surface_camera->set_active(true);
+	const float ev100_sunny16 = physics::light::ev::from_settings(16.0f, 1.0f / 100.0f, 100.0f);
+	ctx.surface_camera->set_exposure(15.5f);
+	
+	ctx.surface_camera->look_at({-55, 36, 10}, {0, 100, 0}, {0, 1, 0});
+	ctx.surface_camera->update_tweens();
+	
+	// Setup and enable sky and ground passes
+	ctx.sky_pass->set_enabled(true);
+	ctx.ground_pass->set_enabled(true);
+	
+	// Disable UI color clear
+	ctx.ui_clear_pass->set_cleared_buffers(false, true, false);
+	
+	// Create mating swarm
+	game::ant::create_swarm(ctx);
+	
+	if (!ctx.menu_bg_billboard->is_active())
+		game::menu::fade_in_bg(ctx);
+		
+	
 	ctx.logger->pop_task(EXIT_SUCCESS);
 }
 
@@ -253,6 +296,9 @@ main_menu::~main_menu()
 	// Destruct title text
 	ctx.ui_scene->remove_object(&title_text);
 	
+	// Destroy swarm
+	game::ant::destroy_swarm(ctx);
+	
 	ctx.logger->pop_task(EXIT_SUCCESS);
 }
 
diff --git a/src/game/state/nuptial-flight.cpp b/src/game/state/nuptial-flight.cpp
index 3e1cd4d..ce4723c 100644
--- a/src/game/state/nuptial-flight.cpp
+++ b/src/game/state/nuptial-flight.cpp
@@ -19,9 +19,11 @@
 
 #include "game/state/nuptial-flight.hpp"
 #include "game/state/pause-menu.hpp"
+#include "game/ant/swarm.hpp"
 #include "entity/archetype.hpp"
 #include "entity/systems/camera.hpp"
 #include "entity/systems/astronomy.hpp"
+#include "entity/systems/atmosphere.hpp"
 #include "entity/components/locomotion.hpp"
 #include "entity/components/transform.hpp"
 #include "entity/components/terrain.hpp"
@@ -45,6 +47,8 @@
 #include "game/ant/breed.hpp"
 #include "game/ant/morphogenesis.hpp"
 #include "math/interpolation.hpp"
+#include "physics/light/exposure.hpp"
+#include "color/color.hpp"
 #include <iostream>
 
 using namespace game::ant;
@@ -92,61 +96,20 @@ nuptial_flight::nuptial_flight(game::context& ctx):
 	ctx.ui_clear_pass->set_cleared_buffers(false, true, false);
 	
 	// Create world if not yet created
-	if (ctx.entities.find("planet") == ctx.entities.end())
+	if (ctx.entities.find("earth") == ctx.entities.end())
 	{
 		// Create cosmos
 		game::world::cosmogenesis(ctx);
 		
-		// Create boids
-		for (int i = 0; i < 100; ++i)
-		{
-			entity::id boid_eid = ctx.entity_registry->create();
-			
-			// Create transform component
-			entity::component::transform transform;
-			transform.local = math::transform<float>::identity;
-			transform.world = math::transform<float>::identity;
-			transform.warp = true;
-			ctx.entity_registry->assign<entity::component::transform>(boid_eid, transform);
-			
-			// Create model component
-			entity::component::model model;
-			model.render_model = ctx.resource_manager->load<render::model>("boid.mdl");
-			model.instance_count = 0;
-			model.layers = 1;
-			ctx.entity_registry->assign<entity::component::model>(boid_eid, model);
-			
-			// Create steering component
-			entity::component::steering steering;
-			steering.agent.mass = 1.0f;
-			steering.agent.position = {0, 100, 0};
-			steering.agent.velocity = {0, 0, 0};
-			steering.agent.acceleration = {0, 0, 0};
-			steering.agent.max_force = 4.0f;
-			steering.agent.max_speed = 5.0f;
-			steering.agent.max_speed_squared = steering.agent.max_speed * steering.agent.max_speed;
-			steering.agent.orientation = math::quaternion<float>::identity;
-			steering.agent.forward = steering.agent.orientation * config::global_forward;
-			steering.agent.up = steering.agent.orientation * config::global_up;
-			steering.wander_weight = 1.0f;
-			steering.wander_noise = math::radians(2000.0f);
-			steering.wander_distance = 10.0f;
-			steering.wander_radius = 8.0f;
-			steering.wander_angle = 0.0f;
-			steering.wander_angle2 = 0.0f;
-			steering.seek_weight = 0.2f;
-			steering.seek_target = {0, 100, 0};
-			steering.flee_weight = 0.0f;
-			steering.sum_weights = steering.wander_weight + steering.seek_weight + steering.flee_weight;
-			ctx.entity_registry->assign<entity::component::steering>(boid_eid, steering);
-		}
+		// Create observer
+		game::world::create_observer(ctx);
 	}
 	
 	// Load biome
 	game::load::biome(ctx, "desert-scrub.bio");
 	
 	// Set world time
-	game::world::set_time(ctx, 2022, 6, 21, 12, 0, 0.0);
+	game::world::set_time(ctx, 2022, 10, 9, 12, 0, 0.0);
 	
 	// Set world time scale
 	game::world::set_time_scale(ctx, 60.0);
@@ -162,6 +125,13 @@ nuptial_flight::nuptial_flight(game::context& ctx):
 		entity::command::warp_to(*ctx.entity_registry, color_checker_eid, {0, 0, -10});
 	}
 	
+	// Create diffuse spheres
+	{
+		entity::archetype* diffuse_spheres_archetype = ctx.resource_manager->load<entity::archetype>("diffuse-spheres.ent");
+		auto diffuse_spheres_eid = diffuse_spheres_archetype->create(*ctx.entity_registry);
+		entity::command::warp_to(*ctx.entity_registry, diffuse_spheres_eid, {0, 0, -20});
+	}
+	
 	// Create ruler
 	{
 		entity::archetype* ruler_10cm_archetype = ctx.resource_manager->load<entity::archetype>("ruler-10cm.ent");
@@ -169,20 +139,6 @@ nuptial_flight::nuptial_flight(game::context& ctx):
 		entity::command::warp_to(*ctx.entity_registry, ruler_10cm_eid, {0, 0, 10});
 	}
 	
-	// Create cocoon
-	{
-		entity::archetype* cocoon_archetype = ctx.resource_manager->load<entity::archetype>("cocoon.ent");
-		auto cocoon_eid = cocoon_archetype->create(*ctx.entity_registry);
-		entity::command::warp_to(*ctx.entity_registry, cocoon_eid, {-10, 0, 10});
-	}
-	
-	// Create larva
-	{
-		entity::archetype* larva_archetype = ctx.resource_manager->load<entity::archetype>("long-neck-larva.ent");
-		auto larva_eid = larva_archetype->create(*ctx.entity_registry);
-		entity::command::warp_to(*ctx.entity_registry, larva_eid, {-13, 0, 10});
-	}
-	
 	// Create keeper if not yet created
 	if (ctx.entities.find("keeper") == ctx.entities.end())
 	{
@@ -190,33 +146,6 @@ nuptial_flight::nuptial_flight(game::context& ctx):
 		ctx.entities["keeper"] = keeper_eid;
 	}
 	
-	// Create ant if not created
-	if (ctx.entities.find("ant") == ctx.entities.end())
-	{
-		auto boid_eid = ctx.entity_registry->create();
-		
-		entity::component::model model;
-		model.render_model = worker_model;//ctx.resource_manager->load<render::model>("ant-test.mdl");
-		model.instance_count = 0;
-		model.layers = 1;
-		ctx.entity_registry->assign<entity::component::model>(boid_eid, model);
-		
-		entity::component::transform transform;
-		transform.local = math::transform<float>::identity;
-		transform.world = math::transform<float>::identity;
-		transform.warp = true;
-		ctx.entity_registry->assign<entity::component::transform>(boid_eid, transform);
-		
-		entity::component::locomotion locomotion;
-		locomotion.yaw = 0.0f;
-		ctx.entity_registry->assign<entity::component::locomotion>(boid_eid, locomotion);
-		
-		entity::command::warp_to(*ctx.entity_registry, boid_eid, {0, 1, 0});
-
-		// Set target ant
-		ctx.entities["ant"] = boid_eid;
-	}
-	
 	// Start as ant-keeper
 	is_keeper = true;
 	
@@ -322,10 +251,12 @@ void nuptial_flight::setup_camera()
 		constraint_stack.head = spring_constraint_eid;
 		ctx.entity_registry->assign<entity::component::constraint_stack>(camera_eid, constraint_stack);
 	}
-
-	ctx.surface_camera->set_exposure(15.5f);
 	
-	ctx.astronomy_system->set_camera(ctx.surface_camera);
+	game::ant::create_swarm(ctx);
+	
+	const float ev100_sunny16 = physics::light::ev::from_settings(16.0f, 1.0f / 100.0f, 100.0f);
+	const float ev100_looney11 = physics::light::ev::from_settings(11.0f, 1.0f / 100.0f, 100.0f);
+	ctx.surface_camera->set_exposure(ev100_sunny16);
 }
 
 void nuptial_flight::enable_keeper_controls()
@@ -687,6 +618,76 @@ void nuptial_flight::enable_keeper_controls()
 			ctx.logger->log("EV100: " + std::to_string(ctx.surface_camera->get_exposure()));
 		}
 	);
+	
+	const float wavelength_speed = 20.0 * (1.0 / 60.0);
+	ctx.controls["dec_red"]->set_active_callback
+	(
+		[&ctx = this->ctx, wavelength_speed](float)
+		{
+			ctx.rgb_wavelengths.x -= wavelength_speed;
+			ctx.atmosphere_system->set_rgb_wavelengths(ctx.rgb_wavelengths * 1e-9);
+			std::stringstream stream;
+			stream << ctx.rgb_wavelengths;
+			ctx.logger->log("wavelengths: " + stream.str());
+		}
+	);
+	ctx.controls["inc_red"]->set_active_callback
+	(
+		[&ctx = this->ctx, wavelength_speed](float)
+		{
+			ctx.rgb_wavelengths.x += wavelength_speed;
+			ctx.atmosphere_system->set_rgb_wavelengths(ctx.rgb_wavelengths * 1e-9);
+			std::stringstream stream;
+			stream << ctx.rgb_wavelengths;
+			ctx.logger->log("wavelengths: " + stream.str());
+		}
+	);
+	
+	ctx.controls["dec_green"]->set_active_callback
+	(
+		[&ctx = this->ctx, wavelength_speed](float)
+		{
+			ctx.rgb_wavelengths.y -= wavelength_speed;
+			ctx.atmosphere_system->set_rgb_wavelengths(ctx.rgb_wavelengths * 1e-9);
+			std::stringstream stream;
+			stream << ctx.rgb_wavelengths;
+			ctx.logger->log("wavelengths: " + stream.str());
+		}
+	);
+	ctx.controls["inc_green"]->set_active_callback
+	(
+		[&ctx = this->ctx, wavelength_speed](float)
+		{
+			ctx.rgb_wavelengths.y += wavelength_speed;
+			ctx.atmosphere_system->set_rgb_wavelengths(ctx.rgb_wavelengths * 1e-9);
+			std::stringstream stream;
+			stream << ctx.rgb_wavelengths;
+			ctx.logger->log("wavelengths: " + stream.str());
+		}
+	);
+	
+	ctx.controls["dec_blue"]->set_active_callback
+	(
+		[&ctx = this->ctx, wavelength_speed](float)
+		{
+			ctx.rgb_wavelengths.z -= wavelength_speed;
+			ctx.atmosphere_system->set_rgb_wavelengths(ctx.rgb_wavelengths * 1e-9);
+			std::stringstream stream;
+			stream << ctx.rgb_wavelengths;
+			ctx.logger->log("wavelengths: " + stream.str());
+		}
+	);
+	ctx.controls["inc_blue"]->set_active_callback
+	(
+		[&ctx = this->ctx, wavelength_speed](float)
+		{
+			ctx.rgb_wavelengths.z += wavelength_speed;
+			ctx.atmosphere_system->set_rgb_wavelengths(ctx.rgb_wavelengths * 1e-9);
+			std::stringstream stream;
+			stream << ctx.rgb_wavelengths;
+			ctx.logger->log("wavelengths: " + stream.str());
+		}
+	);
 }
 
 void nuptial_flight::disable_keeper_controls()
diff --git a/src/game/tools.cpp b/src/game/tools.cpp
index ad80635..14e794e 100644
--- a/src/game/tools.cpp
+++ b/src/game/tools.cpp
@@ -90,8 +90,8 @@ entity::id build_time_tool(game::context& ctx)
 	// Setup tool active calback
 	tool.active = [&ctx]()
 	{
-		auto [mouse_x, mouse_y] = ctx.app->get_mouse()->get_current_position();
-		auto [window_w, window_h] = ctx.app->get_viewport_dimensions();
+		//auto mouse = ctx.app->get_mouse()->get_current_position();
+		//auto window = ctx.app->get_viewport_dimensions();
 		
 		entity::id planet_eid = ctx.entities["planet"];
 		entity::component::celestial_body body = ctx.entity_registry->get<entity::component::celestial_body>(planet_eid);
diff --git a/src/game/world.cpp b/src/game/world.cpp
index a463cc6..b9964fe 100644
--- a/src/game/world.cpp
+++ b/src/game/world.cpp
@@ -27,6 +27,7 @@
 #include "entity/components/orbit.hpp"
 #include "entity/components/terrain.hpp"
 #include "entity/components/transform.hpp"
+#include "entity/components/observer.hpp"
 #include "entity/systems/astronomy.hpp"
 #include "entity/systems/orbit.hpp"
 #include "entity/commands.hpp"
@@ -52,6 +53,7 @@
 #include "gl/texture-wrapping.hpp"
 #include "gl/texture-filter.hpp"
 #include "render/material-flags.hpp"
+#include "geom/solid-angle.hpp"
 #include "config.hpp"
 #include <iostream>
 
@@ -88,15 +90,79 @@ void cosmogenesis(game::context& ctx)
 	ctx.logger->pop_task(EXIT_SUCCESS);
 }
 
-void set_location(game::context& ctx, double elevation, double latitude, double longitude)
+void create_observer(game::context& ctx)
 {
-	// Update context location
-	ctx.elevation = elevation;
-	ctx.latitude = latitude;
-	ctx.longitude = longitude;
+	ctx.logger->push_task("Creating observer");
 	
-	// Pass location to astronomy system
-	ctx.astronomy_system->set_observer_location({elevation, latitude, longitude});
+	try
+	{
+		// Create observer entity
+		entity::id observer_eid = ctx.entity_registry->create();
+		ctx.entities["observer"] = observer_eid;
+		
+		// Construct observer component
+		entity::component::observer observer;
+		
+		// Set observer reference body
+		if (auto it = ctx.entities.find("earth"); it != ctx.entities.end())
+			observer.reference_body_eid = it->second;
+		else
+			observer.reference_body_eid = entt::null;
+		
+		// Set observer location
+		observer.elevation = 0.0;
+		observer.latitude = 0.0;
+		observer.longitude = 0.0;
+		
+		// Assign observer component to observer entity
+		ctx.entity_registry->assign<entity::component::observer>(observer_eid, observer);
+		
+		// Set astronomy system observer
+		ctx.astronomy_system->set_observer(observer_eid);
+	}
+	catch (const std::exception&)
+	{
+		ctx.logger->pop_task(EXIT_FAILURE);
+		return;
+	}
+	
+	ctx.logger->pop_task(EXIT_SUCCESS);
+}
+
+void set_location(game::context& ctx, double elevation, double latitude, double longitude)
+{
+	if (auto it = ctx.entities.find("observer"); it != ctx.entities.end())
+	{
+		entity::id observer_eid = it->second;
+		
+		if (observer_eid != entt::null)
+		{
+			// Get pointer to observer component
+			const auto observer = ctx.entity_registry->try_get<entity::component::observer>(observer_eid);
+			
+			// Set observer location
+			if (observer)
+			{
+				observer->elevation = elevation;
+				observer->latitude = latitude;
+				observer->longitude = longitude;
+				ctx.entity_registry->replace<entity::component::observer>(observer_eid, *observer);
+			}
+
+			/*
+			ctx.entity_registry->patch<entity::component::observer>
+			(
+				observer_eid,
+				[&](auto& component)
+				{
+					component.elevation = elevation;
+					component.latitude = latitude;
+					component.longitude = longitude;
+				}
+			);
+			*/
+		}
+	}
 }
 
 void set_time(game::context& ctx, double t)
@@ -117,28 +183,36 @@ void set_time(game::context& ctx, double t)
 
 void set_time(game::context& ctx, int year, int month, int day, int hour, int minute, double second)
 {
-	const double utc_offset = physics::time::utc::offset<double>(ctx.longitude);
+	double longitude = 0.0;
+	
+	// Get longitude of observer
+	if (auto it = ctx.entities.find("observer"); it != ctx.entities.end())
+	{
+		entity::id observer_eid = it->second;
+		if (observer_eid != entt::null)
+		{
+			const auto observer = ctx.entity_registry->try_get<entity::component::observer>(observer_eid);
+			if (observer)
+				longitude = observer->longitude;
+		}
+	}
+	
+	// Calculate UTC offset at longitude
+	const double utc_offset = physics::time::utc::offset<double>(longitude);
+	
+	// Convert time from Gregorian to UT1
 	const double t = physics::time::gregorian::to_ut1<double>(year, month, day, hour, minute, second, utc_offset);
+	
 	set_time(ctx, t);
 }
 
 void set_time_scale(game::context& ctx, double scale)
 {
-	ctx.logger->push_task("Setting time scale to " + std::to_string(scale));
-	try
-	{
-		// Convert time scale from seconds to days
-		const double astronomical_scale = scale / physics::time::seconds_per_day<double>;
-		
-		ctx.orbit_system->set_time_scale(astronomical_scale);
-		ctx.astronomy_system->set_time_scale(astronomical_scale);
-	}
-	catch (const std::exception&)
-	{
-		ctx.logger->pop_task(EXIT_FAILURE);
-		return;
-	}
-	ctx.logger->pop_task(EXIT_SUCCESS);
+	// Convert time scale from seconds to days
+	const double astronomical_scale = scale / physics::time::seconds_per_day<double>;
+	
+	ctx.orbit_system->set_time_scale(astronomical_scale);
+	ctx.astronomy_system->set_time_scale(astronomical_scale);
 }
 
 void load_ephemeris(game::context& ctx)
@@ -208,7 +282,7 @@ void create_stars(game::context& ctx)
 	float* star_vertex = star_vertex_data;
 	
 	// Init starlight illuminance
-	double starlight_illuminance = 0.0;
+	double3 starlight_illuminance = {0, 0, 0};
 	
 	// Build star catalog vertex data
 	for (std::size_t i = 1; i < star_catalog->size(); ++i)
@@ -239,9 +313,6 @@ void create_stars(game::context& ctx)
 		// Convert ICRF coordinates from spherical to Cartesian
 		double3 position = physics::orbit::frame::bci::cartesian(double3{1.0, dec, ra});
 		
-		// Convert apparent magnitude to brightness factor relative to a 0th magnitude star
-		double brightness = physics::light::vmag::to_brightness(vmag);
-		
 		// Convert color index to color temperature
 		double cct = color::index::bv_to_cct(bv);
 		
@@ -249,10 +320,10 @@ void create_stars(game::context& ctx)
 		double3 color_xyz = color::cct::to_xyz(cct);
 		
 		// Transform XYZ color to ACEScg colorspace
-		double3 color_acescg = color::xyz::to_acescg(color_xyz);
+		double3 color_acescg = color::aces::ap1<double>.from_xyz * color_xyz;
 		
-		// Scale color by relative brightness
-		color_acescg *= brightness;
+		// Convert apparent magnitude to brightness factor relative to a 0th magnitude star
+		double brightness = physics::light::vmag::to_brightness(vmag);
 		
 		// Build vertex
 		*(star_vertex++) = static_cast<float>(position.x);
@@ -263,8 +334,11 @@ void create_stars(game::context& ctx)
 		*(star_vertex++) = static_cast<float>(color_acescg.z);
 		*(star_vertex++) = static_cast<float>(brightness);
 		
-		// Convert apparent magnitude to illuminance and add to total starlight
-		starlight_illuminance += physics::light::vmag::to_illuminance(vmag);
+		// Calculate spectral illuminance
+		double3 illuminance = color_acescg * physics::light::vmag::to_illuminance(vmag);
+		
+		// Add spectral illuminance to total starlight illuminance
+		starlight_illuminance += illuminance;
 	}
 	
 	// Unload star catalog
@@ -347,14 +421,22 @@ void create_sun(game::context& ctx)
 		sky_light->set_color({0, 0, 0});
 		sky_light->update_tweens();
 		
+		// Create bounce directional light scene object
+		scene::directional_light* bounce_light = new scene::directional_light();
+		bounce_light->set_color({0, 0, 0});
+		bounce_light->look_at({0, 0, 0}, {0, 1, 0}, {1, 0, 0});
+		bounce_light->update_tweens();
+		
 		// Add sun light scene objects to surface scene
 		ctx.surface_scene->add_object(sun_light);
 		ctx.surface_scene->add_object(sky_light);
+		ctx.surface_scene->add_object(bounce_light);
 		
 		// Pass direct sun light scene object to shadow map pass and astronomy system
 		ctx.surface_shadow_map_pass->set_light(sun_light);
 		ctx.astronomy_system->set_sun_light(sun_light);
 		ctx.astronomy_system->set_sky_light(sky_light);
+		ctx.astronomy_system->set_bounce_light(bounce_light);
 	}
 	catch (const std::exception&)
 	{
@@ -415,9 +497,6 @@ void create_earth(game::context& ctx)
 		terrain.max_lod = 0;
 		terrain.patch_material = nullptr;
 		//ctx.entity_registry->assign<entity::component::terrain>(earth_eid, terrain);
-		
-		// Pass earth to astronomy system as reference body
-		ctx.astronomy_system->set_reference_body(earth_eid);
 	}
 	catch (const std::exception&)
 	{
diff --git a/src/game/world.hpp b/src/game/world.hpp
index 5cb705e..c76f99b 100644
--- a/src/game/world.hpp
+++ b/src/game/world.hpp
@@ -30,6 +30,9 @@ namespace world {
 /// Creates the cosmos.
 void cosmogenesis(game::context& ctx);
 
+/// Creates the observer.
+void create_observer(game::context& ctx);
+
 /**
  * Sets the location of the observer.
  *
diff --git a/src/geom/sphere.hpp b/src/geom/sphere.hpp
index ea4bb48..bbff112 100644
--- a/src/geom/sphere.hpp
+++ b/src/geom/sphere.hpp
@@ -69,7 +69,7 @@ template <class T>
 bool sphere<T>::intersects(const sphere<T>& sphere) const
 {
 	vector_type d = center - sphere.center;
-	float r = radius + sphere.radius;
+	T r = radius + sphere.radius;
 	return (math::dot(d, d) <= r * r);
 }
 
@@ -82,7 +82,7 @@ bool sphere<T>::intersects(const aabb<T>& aabb) const
 template <class T>
 bool sphere<T>::contains(const sphere<T>& sphere) const
 {
-	float containment_radius = radius - sphere.radius;
+	T containment_radius = radius - sphere.radius;
 	if (containment_radius < T(0))
 		return false;
 	
@@ -115,4 +115,3 @@ bool sphere<T>::contains(const vector_type& point) const
 } // namespace geom
 
 #endif // ANTKEEPER_GEOM_SPHERE_HPP
-
diff --git a/src/gl/texture-2d.cpp b/src/gl/texture-2d.cpp
index 5e8d55b..8f286f3 100644
--- a/src/gl/texture-2d.cpp
+++ b/src/gl/texture-2d.cpp
@@ -33,6 +33,11 @@ void texture_2d::resize(std::uint16_t width, std::uint16_t height, gl::pixel_typ
 	texture::resize(width, height, type, format, color_space, data);
 }
 
+void texture_2d::resize(std::uint16_t width, std::uint16_t height, const void* data)
+{
+	texture::resize(width, height, get_pixel_type(), get_pixel_format(), get_color_space(), data);
+}
+
 void texture_2d::set_wrapping(gl::texture_wrapping wrap_s, texture_wrapping wrap_t)
 {
 	texture::set_wrapping(wrap_s, wrap_t);
diff --git a/src/gl/texture-2d.hpp b/src/gl/texture-2d.hpp
index 4f5c824..4fa76a9 100644
--- a/src/gl/texture-2d.hpp
+++ b/src/gl/texture-2d.hpp
@@ -39,6 +39,15 @@ public:
 	/// @copydoc texture::resize(std::uint16_t, std::uint16_t, gl::pixel_type, gl::pixel_format, gl::color_space, const void*)
 	virtual void resize(std::uint16_t width, std::uint16_t height, gl::pixel_type type, gl::pixel_format format, gl::color_space color_space, const void* data);
 	
+	/**
+	 * Resizes the texture.
+	 *
+	 * @param width Texture width, in pixels.
+	 * @param height Texture height, in pixels.
+	 * @param data Pointer to pixel data.
+	 */
+	void resize(std::uint16_t width, std::uint16_t height, const void* data);
+	
 	/// @copydoc texture::set_wrapping(gl::texture_wrapping, gl::texture_wrapping)
 	virtual void set_wrapping(gl::texture_wrapping wrap_s, texture_wrapping wrap_t);
 };
diff --git a/src/math/matrix-functions.hpp b/src/math/matrix-functions.hpp
index d4edf5d..c335899 100644
--- a/src/math/matrix-functions.hpp
+++ b/src/math/matrix-functions.hpp
@@ -35,15 +35,15 @@ namespace math {
  * @return Sum of the two matrices.
  */
 template <class T>
-matrix<T, 2, 2> add(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
+constexpr matrix<T, 2, 2> add(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
 
 /// @copydoc add(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 3, 3> add(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
+constexpr matrix<T, 3, 3> add(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
 
 /// @copydoc add(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 4, 4> add(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
+constexpr matrix<T, 4, 4> add(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
 
 /**
  * Reinterprets data as an `NxM` matrix of type `T`.
@@ -54,7 +54,7 @@ matrix<T, 4, 4> add(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
  * @param data Data to reinterpret.
  */
 template <std::size_t N, std::size_t M, typename T>
-matrix<T, N, M>& as_matrix(T& data);
+constexpr matrix<T, N, M>& as_matrix(T& data);
 
 /**
  * Calculates the determinant of a matrix.
@@ -62,15 +62,15 @@ matrix<T, N, M>& as_matrix(T& data);
  * @param m Matrix of which to take the determinant. 
  */
 template <class T>
-T determinant(const matrix<T, 2, 2>& m);
+constexpr T determinant(const matrix<T, 2, 2>& m);
 
 /// @copydoc determinant(const matrix<T, 2, 2>&)
 template <class T>
-T determinant(const matrix<T, 3, 3>& m);
+constexpr T determinant(const matrix<T, 3, 3>& m);
 
 /// @copydoc determinant(const matrix<T, 2, 2>&)
 template <class T>
-T determinant(const matrix<T, 4, 4>& m);
+constexpr T determinant(const matrix<T, 4, 4>& m);
 
 /**
  * Calculates the inverse of a matrix.
@@ -78,15 +78,15 @@ T determinant(const matrix<T, 4, 4>& m);
  * @param m Matrix of which to take the inverse.
  */
 template <class T>
-matrix<T, 2, 2> inverse(const matrix<T, 2, 2>& m);
+constexpr matrix<T, 2, 2> inverse(const matrix<T, 2, 2>& m);
 
 /// @copydoc inverse(const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 3, 3> inverse(const matrix<T, 3, 3>& m);
+constexpr matrix<T, 3, 3> inverse(const matrix<T, 3, 3>& m);
 
 /// @copydoc inverse(const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 4, 4> inverse(const matrix<T, 4, 4>& m);
+constexpr matrix<T, 4, 4> inverse(const matrix<T, 4, 4>& m);
 
 /**
  * Performs a component-wise multiplication of two matrices.
@@ -95,15 +95,15 @@ matrix<T, 4, 4> inverse(const matrix<T, 4, 4>& m);
  * @param y Second matrix multiplicand.
  */
 template <class T>
-matrix<T, 2, 2> componentwise_mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
+constexpr matrix<T, 2, 2> componentwise_mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
 
 /// @copydoc componentwise_mul(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 3, 3> componentwise_mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
+constexpr matrix<T, 3, 3> componentwise_mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
 
 /// @copydoc componentwise_mul(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 4, 4> componentwise_mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
+constexpr matrix<T, 4, 4> componentwise_mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
 
 /**
  * Creates a viewing transformation matrix.
@@ -114,7 +114,7 @@ matrix<T, 4, 4> componentwise_mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4
  * @return Viewing transformation matrix.
  */
 template <class T>
-matrix<T, 4, 4> look_at(const vector<T, 3>& position, const vector<T, 3>& target, vector<T, 3> up);
+constexpr matrix<T, 4, 4> look_at(const vector<T, 3>& position, const vector<T, 3>& target, vector<T, 3> up);
 
 /**
  * Multiplies two matrices.
@@ -124,15 +124,15 @@ matrix<T, 4, 4> look_at(const vector<T, 3>& position, const vector<T, 3>& target
  * @return Product of the two matrices.
  */
 template <class T>
-matrix<T, 2, 2> mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
+constexpr matrix<T, 2, 2> mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
 
 /// @copydoc mul(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&);
 template <class T>
-matrix<T, 3, 3> mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
+constexpr matrix<T, 3, 3> mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
 
 /// @copydoc mul(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&);
 template <class T>
-matrix<T, 4, 4> mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
+constexpr matrix<T, 4, 4> mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
 
 /**
  * Multiplies a matrix by a scalar.
@@ -142,7 +142,7 @@ matrix<T, 4, 4> mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
  * @return Product of the matrix and the scalar..
  */
 template <class T, std::size_t N, std::size_t M>
-matrix<T, N, M> mul(const matrix<T, N, M>& m, T s);
+constexpr matrix<T, N, M> mul(const matrix<T, N, M>& m, T s);
 
 /**
  * Transforms a vector by a matrix.
@@ -152,15 +152,15 @@ matrix<T, N, M> mul(const matrix<T, N, M>& m, T s);
  * @return Transformed vector.
  */
 template <class T>
-vector<T, 2> mul(const matrix<T, 2, 2>& m, const vector<T, 2>& v);
+constexpr vector<T, 2> mul(const matrix<T, 2, 2>& m, const vector<T, 2>& v);
 
 /// @copydoc mul(const matrix<T, 2, 2>&, const vector<T, 2>&)
 template <class T>
-vector<T, 3> mul(const matrix<T, 3, 3>& m, const vector<T, 3>& v);
+constexpr vector<T, 3> mul(const matrix<T, 3, 3>& m, const vector<T, 3>& v);
 
 /// @copydoc mul(const matrix<T, 2, 2>&, const vector<T, 2>&)
 template <class T>
-vector<T, 4> mul(const matrix<T, 4, 4>& m, const vector<T, 4>& v);
+constexpr vector<T, 4> mul(const matrix<T, 4, 4>& m, const vector<T, 4>& v);
 
 /**
  * Creates an orthographic projection matrix which will transform the near and far clipping planes to `[-1, 1]`, respectively.
@@ -174,7 +174,7 @@ vector<T, 4> mul(const matrix<T, 4, 4>& m, const vector<T, 4>& v);
  * @return Orthographic projection matrix.
  */
 template <class T>
-matrix<T, 4, 4> ortho(T left, T right, T bottom, T top, T z_near, T z_far);
+constexpr matrix<T, 4, 4> ortho(T left, T right, T bottom, T top, T z_near, T z_far);
 
 /**
  * Creates an orthographic projection matrix which will transform the near and far clipping planes to `[0, 1]`, respectively.
@@ -188,7 +188,7 @@ matrix<T, 4, 4> ortho(T left, T right, T bottom, T top, T z_near, T z_far);
  * @return Orthographic projection matrix.
  */
 template <class T>
-matrix<T, 4, 4> ortho_half_z(T left, T right, T bottom, T top, T z_near, T z_far);
+constexpr matrix<T, 4, 4> ortho_half_z(T left, T right, T bottom, T top, T z_near, T z_far);
 
 /**
  * Calculates the outer product of a pair of vectors.
@@ -197,15 +197,15 @@ matrix<T, 4, 4> ortho_half_z(T left, T right, T bottom, T top, T z_near, T z_far
  * @param r Parameter to be treated as a row vector.
  */
 template <class T>
-matrix<T, 2, 2> outer_product(const vector<T, 2>& c, const vector<T, 2>& r);
+constexpr matrix<T, 2, 2> outer_product(const vector<T, 2>& c, const vector<T, 2>& r);
 
 /// @copydoc outer_product(const vector<T, 2>&, const vector<T, 2>&)
 template <class T>
-matrix<T, 3, 3> outer_product(const vector<T, 3>& c, const vector<T, 3>& r);
+constexpr matrix<T, 3, 3> outer_product(const vector<T, 3>& c, const vector<T, 3>& r);
 
 /// @copydoc outer_product(const vector<T, 2>&, const vector<T, 2>&)
 template <class T>
-matrix<T, 4, 4> outer_product(const vector<T, 4>& c, const vector<T, 4> r);
+constexpr matrix<T, 4, 4> outer_product(const vector<T, 4>& c, const vector<T, 4> r);
 
 /**
  * Creates a perspective projection matrix which will transform the near and far clipping planes to `[-1, 1]`, respectively.
@@ -238,7 +238,7 @@ matrix<T, 4, 4> perspective_half_z(T vertical_fov, T aspect_ratio, T z_near, T z
  * @return Resized matrix.
  */
 template <std::size_t N1, std::size_t M1, class T, std::size_t N0, std::size_t M0>
-matrix<T, N1, M1> resize(const matrix<T, N0, M0>& m);
+constexpr matrix<T, N1, M1> resize(const matrix<T, N0, M0>& m);
 
 /**
  * Constructs a rotation matrix.
@@ -285,7 +285,7 @@ matrix3<T> rotate_z(T angle);
  * @return Scaled matrix.
  */
 template <class T>
-matrix<T, 4, 4> scale(const matrix<T, 4, 4>& m, const vector<T, 3>& v);
+constexpr matrix<T, 4, 4> scale(const matrix<T, 4, 4>& m, const vector<T, 3>& v);
 
 /**
  * Subtracts a matrix from another matrix.
@@ -295,15 +295,15 @@ matrix<T, 4, 4> scale(const matrix<T, 4, 4>& m, const vector<T, 3>& v);
  * @return Difference between the two matrices.
  */
 template <class T>
-matrix<T, 2, 2> sub(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
+constexpr matrix<T, 2, 2> sub(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y);
 
 /// @copydoc sub(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 3, 3> sub(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
+constexpr matrix<T, 3, 3> sub(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y);
 
 /// @copydoc sub(const matrix<T, 2, 2>&, const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 4, 4> sub(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
+constexpr matrix<T, 4, 4> sub(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
 
 /**
  * Translates a matrix.
@@ -313,7 +313,7 @@ matrix<T, 4, 4> sub(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y);
  * @return Translated matrix.
  */
 template <class T>
-matrix<T, 4, 4> translate(const matrix<T, 4, 4>& m, const vector<T, 3>& v);
+constexpr matrix<T, 4, 4> translate(const matrix<T, 4, 4>& m, const vector<T, 3>& v);
 
 /**
  * Calculates the transpose of a matrix.
@@ -321,15 +321,15 @@ matrix<T, 4, 4> translate(const matrix<T, 4, 4>& m, const vector<T, 3>& v);
  * @param m Matrix of which to take the transpose.
  */
 template <class T>
-matrix<T, 2, 2> transpose(const matrix<T, 2, 2>& m);
+constexpr matrix<T, 2, 2> transpose(const matrix<T, 2, 2>& m);
 
 /// @copydoc transpose(const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 3, 3> transpose(const matrix<T, 3, 3>& m);
+constexpr matrix<T, 3, 3> transpose(const matrix<T, 3, 3>& m);
 
 /// @copydoc transpose(const matrix<T, 2, 2>&)
 template <class T>
-matrix<T, 4, 4> transpose(const matrix<T, 4, 4>& m);
+constexpr matrix<T, 4, 4> transpose(const matrix<T, 4, 4>& m);
 
 /**
  * Types casts each matrix element and returns a matrix of the casted type.
@@ -342,10 +342,10 @@ matrix<T, 4, 4> transpose(const matrix<T, 4, 4>& m);
  * @return Type-casted matrix.
  */
 template <class T2, class T1, std::size_t N, std::size_t M>
-matrix<T2, N, M> type_cast(const matrix<T1, N, M>& m);
+constexpr matrix<T2, N, M> type_cast(const matrix<T1, N, M>& m);
 
 template <class T>
-matrix<T, 2, 2> add(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
+constexpr matrix<T, 2, 2> add(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
 {
 	return
 		{{
@@ -355,7 +355,7 @@ matrix<T, 2, 2> add(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
 }
 
 template <class T>
-matrix<T, 3, 3> add(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
+constexpr matrix<T, 3, 3> add(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
 {
 	return
 		{{
@@ -366,7 +366,7 @@ matrix<T, 3, 3> add(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
 }
 
 template <class T>
-matrix<T, 4, 4> add(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
+constexpr matrix<T, 4, 4> add(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
 {
 	return
 		{{
@@ -378,20 +378,20 @@ matrix<T, 4, 4> add(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
 }
 
 template <std::size_t N, std::size_t M, typename T>
-inline matrix<T, N, M>& as_matrix(T& data)
+constexpr inline matrix<T, N, M>& as_matrix(T& data)
 {
 	static_assert(std::is_pod<matrix<T, N, M>>::value);
 	return reinterpret_cast<matrix<T, N, M>&>(data);
 }
 
 template <class T>
-T determinant(const matrix<T, 2, 2>& m)
+constexpr T determinant(const matrix<T, 2, 2>& m)
 {
 	return m[0][0] * m[1][1] - m[0][1] * m[1][0];
 }
 
 template <class T>
-T determinant(const matrix<T, 3, 3>& m)
+constexpr T determinant(const matrix<T, 3, 3>& m)
 {
 	return m[0][0] * m [1][1] * m[2][2] +
 		m[0][1] * m[1][2] * m[2][0] +
@@ -402,7 +402,7 @@ T determinant(const matrix<T, 3, 3>& m)
 }
 
 template <class T>
-T determinant(const matrix<T, 4, 4>& m)
+constexpr T determinant(const matrix<T, 4, 4>& m)
 {
 	return m[0][3] * m[1][2] * m[2][1] * m[3][0] - m[0][2] * m[1][3] * m[2][1] * m[3][0] -
 		m[0][3] * m[1][1] * m[2][2] * m[3][0] + m[0][1] * m[1][3] * m[2][2] * m[3][0] +
@@ -419,7 +419,7 @@ T determinant(const matrix<T, 4, 4>& m)
 }
 
 template <class T>
-matrix<T, 2, 2> inverse(const matrix<T, 2, 2>& m)
+constexpr matrix<T, 2, 2> inverse(const matrix<T, 2, 2>& m)
 {
 	static_assert(std::is_floating_point<T>::value);
 
@@ -432,7 +432,7 @@ matrix<T, 2, 2> inverse(const matrix<T, 2, 2>& m)
 }
 
 template <class T>
-matrix<T, 3, 3> inverse(const matrix<T, 3, 3>& m)
+constexpr matrix<T, 3, 3> inverse(const matrix<T, 3, 3>& m)
 {
 	static_assert(std::is_floating_point<T>::value);
 
@@ -461,7 +461,7 @@ matrix<T, 3, 3> inverse(const matrix<T, 3, 3>& m)
 }
 
 template <class T>
-matrix<T, 4, 4> inverse(const matrix<T, 4, 4>& m)
+constexpr matrix<T, 4, 4> inverse(const matrix<T, 4, 4>& m)
 {
 	static_assert(std::is_floating_point<T>::value);
 
@@ -500,7 +500,7 @@ matrix<T, 4, 4> inverse(const matrix<T, 4, 4>& m)
 }
 
 template <class T>
-matrix<T, 2, 2> componentwise_mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
+constexpr matrix<T, 2, 2> componentwise_mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
 {
 	return
 		{{
@@ -510,7 +510,7 @@ matrix<T, 2, 2> componentwise_mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2
 }
 
 template <class T>
-matrix<T, 3, 3> componentwise_mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
+constexpr matrix<T, 3, 3> componentwise_mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
 {
 	return
 		{{
@@ -521,7 +521,7 @@ matrix<T, 3, 3> componentwise_mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3
 }
 
 template <class T>
-matrix<T, 4, 4> componentwise_mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
+constexpr matrix<T, 4, 4> componentwise_mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
 {
 	return
 		{{
@@ -533,7 +533,7 @@ matrix<T, 4, 4> componentwise_mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4
 }
 
 template <class T>
-matrix<T, 4, 4> look_at(const vector<T, 3>& position, const vector<T, 3>& target, vector<T, 3> up)
+constexpr matrix<T, 4, 4> look_at(const vector<T, 3>& position, const vector<T, 3>& target, vector<T, 3> up)
 {
 	vector<T, 3> forward = normalize(sub(target, position));
 	vector<T, 3> right = normalize(cross(forward, up));
@@ -551,7 +551,7 @@ matrix<T, 4, 4> look_at(const vector<T, 3>& position, const vector<T, 3>& target
 }
 
 template <class T>
-matrix<T, 2, 2> mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
+constexpr matrix<T, 2, 2> mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
 {
 	return
 		{{
@@ -561,7 +561,7 @@ matrix<T, 2, 2> mul(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
 }
 
 template <class T>
-matrix<T, 3, 3> mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
+constexpr matrix<T, 3, 3> mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
 {
 	return
 		{{
@@ -572,7 +572,7 @@ matrix<T, 3, 3> mul(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
 }
 
 template <class T>
-matrix<T, 4, 4> mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
+constexpr matrix<T, 4, 4> mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
 {
 	return
 		{{
@@ -585,19 +585,19 @@ matrix<T, 4, 4> mul(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
 
 /// @private
 template <class T, std::size_t N, std::size_t M, std::size_t... I>
-inline matrix<T, N, M> mul(const matrix<T, N, M>& m, T s, std::index_sequence<I...>)
+constexpr inline matrix<T, N, M> mul(const matrix<T, N, M>& m, T s, std::index_sequence<I...>)
 {
 	return {{(m[I] * s)...}};
 }
 
 template <class T, std::size_t N, std::size_t M>
-inline matrix<T, N, M> mul(const matrix<T, N, M>& m, T s)
+constexpr inline matrix<T, N, M> mul(const matrix<T, N, M>& m, T s)
 {
 	return mul(m, s, std::make_index_sequence<N>{});
 }
 
 template <class T>
-vector<T, 2> mul(const matrix<T, 2, 2>& m, const vector<T, 2>& v)
+constexpr vector<T, 2> mul(const matrix<T, 2, 2>& m, const vector<T, 2>& v)
 {
 	return
 		{
@@ -607,7 +607,7 @@ vector<T, 2> mul(const matrix<T, 2, 2>& m, const vector<T, 2>& v)
 }
 
 template <class T>
-vector<T, 3> mul(const matrix<T, 3, 3>& m, const vector<T, 3>& v)
+constexpr vector<T, 3> mul(const matrix<T, 3, 3>& m, const vector<T, 3>& v)
 {
 	return
 		{
@@ -618,7 +618,7 @@ vector<T, 3> mul(const matrix<T, 3, 3>& m, const vector<T, 3>& v)
 }
 
 template <class T>
-vector<T, 4> mul(const matrix<T, 4, 4>& m, const vector<T, 4>& v)
+constexpr vector<T, 4> mul(const matrix<T, 4, 4>& m, const vector<T, 4>& v)
 {
 	return
 		{
@@ -630,7 +630,7 @@ vector<T, 4> mul(const matrix<T, 4, 4>& m, const vector<T, 4>& v)
 }
 
 template <class T>
-matrix<T, 4, 4> ortho(T left, T right, T bottom, T top, T z_near, T z_far)
+constexpr matrix<T, 4, 4> ortho(T left, T right, T bottom, T top, T z_near, T z_far)
 {
 	return
 		{{
@@ -642,7 +642,7 @@ matrix<T, 4, 4> ortho(T left, T right, T bottom, T top, T z_near, T z_far)
 }
 
 template <class T>
-matrix<T, 4, 4> ortho_half_z(T left, T right, T bottom, T top, T z_near, T z_far)
+constexpr matrix<T, 4, 4> ortho_half_z(T left, T right, T bottom, T top, T z_near, T z_far)
 {
 	return
 		{{
@@ -654,7 +654,7 @@ matrix<T, 4, 4> ortho_half_z(T left, T right, T bottom, T top, T z_near, T z_far
 }
 
 template <class T>
-matrix<T, 2, 2> outer_product(const vector<T, 2>& c, const vector<T, 2>& r)
+constexpr matrix<T, 2, 2> outer_product(const vector<T, 2>& c, const vector<T, 2>& r)
 {
 	return
 		{{
@@ -664,7 +664,7 @@ matrix<T, 2, 2> outer_product(const vector<T, 2>& c, const vector<T, 2>& r)
 }
 
 template <class T>
-matrix<T, 3, 3> outer_product(const vector<T, 3>& c, const vector<T, 3>& r)
+constexpr matrix<T, 3, 3> outer_product(const vector<T, 3>& c, const vector<T, 3>& r)
 {
 	return
 		{{
@@ -675,7 +675,7 @@ matrix<T, 3, 3> outer_product(const vector<T, 3>& c, const vector<T, 3>& r)
 }
 
 template <class T>
-matrix<T, 4, 4> outer_product(const vector<T, 4>& c, const vector<T, 4> r)
+constexpr matrix<T, 4, 4> outer_product(const vector<T, 4>& c, const vector<T, 4> r)
 {
 	return
 		{{
@@ -717,7 +717,7 @@ matrix<T, 4, 4> perspective_half_z(T vertical_fov, T aspect_ratio, T z_near, T z
 }
 
 template <std::size_t N1, std::size_t M1, class T, std::size_t N0, std::size_t M0>
-matrix<T, N1, M1> resize(const matrix<T, N0, M0>& m)
+constexpr matrix<T, N1, M1> resize(const matrix<T, N0, M0>& m)
 {
 	matrix<T, N1, M1> resized;
 
@@ -796,7 +796,7 @@ matrix3<T> rotate_z(T angle)
 }
 
 template <class T>
-matrix<T, 4, 4> scale(const matrix<T, 4, 4>& m, const vector<T, 3>& v)
+constexpr matrix<T, 4, 4> scale(const matrix<T, 4, 4>& m, const vector<T, 3>& v)
 {
 	return mul(m, matrix<T, 4, 4>
 		{{
@@ -808,7 +808,7 @@ matrix<T, 4, 4> scale(const matrix<T, 4, 4>& m, const vector<T, 3>& v)
 }
 
 template <class T>
-matrix<T, 2, 2> sub(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
+constexpr matrix<T, 2, 2> sub(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
 {
 	return
 		{{
@@ -818,7 +818,7 @@ matrix<T, 2, 2> sub(const matrix<T, 2, 2>& x, const matrix<T, 2, 2>& y)
 }
 
 template <class T>
-matrix<T, 3, 3> sub(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
+constexpr matrix<T, 3, 3> sub(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
 {
 	return
 		{{
@@ -829,7 +829,7 @@ matrix<T, 3, 3> sub(const matrix<T, 3, 3>& x, const matrix<T, 3, 3>& y)
 }
 
 template <class T>
-matrix<T, 4, 4> sub(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
+constexpr matrix<T, 4, 4> sub(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
 {
 	return
 		{{
@@ -841,7 +841,7 @@ matrix<T, 4, 4> sub(const matrix<T, 4, 4>& x, const matrix<T, 4, 4>& y)
 }
 
 template <class T>
-matrix<T, 4, 4> translate(const matrix<T, 4, 4>& m, const vector<T, 3>& v)
+constexpr matrix<T, 4, 4> translate(const matrix<T, 4, 4>& m, const vector<T, 3>& v)
 {
 	return mul(m, matrix<T, 4, 4>
 		{{
@@ -853,7 +853,7 @@ matrix<T, 4, 4> translate(const matrix<T, 4, 4>& m, const vector<T, 3>& v)
 }
 
 template <class T>
-matrix<T, 2, 2> transpose(const matrix<T, 2, 2>& m)
+constexpr matrix<T, 2, 2> transpose(const matrix<T, 2, 2>& m)
 {
 
 	return
@@ -869,7 +869,7 @@ matrix<T, 2, 2> transpose(const matrix<T, 2, 2>& m)
 }
 
 template <class T>
-matrix<T, 3, 3> transpose(const matrix<T, 3, 3>& m)
+constexpr matrix<T, 3, 3> transpose(const matrix<T, 3, 3>& m)
 {
 
 	return
@@ -889,7 +889,7 @@ matrix<T, 3, 3> transpose(const matrix<T, 3, 3>& m)
 }
 
 template <class T>
-matrix<T, 4, 4> transpose(const matrix<T, 4, 4>& m)
+constexpr matrix<T, 4, 4> transpose(const matrix<T, 4, 4>& m)
 {
 
 	return
@@ -914,13 +914,13 @@ matrix<T, 4, 4> transpose(const matrix<T, 4, 4>& m)
 
 /// @private
 template <class T2, class T1, std::size_t N, std::size_t M, std::size_t... I>
-inline matrix<T2, N, M> type_cast(const matrix<T1, N, M>& m, std::index_sequence<I...>)
+constexpr inline matrix<T2, N, M> type_cast(const matrix<T1, N, M>& m, std::index_sequence<I...>)
 {
 	return {type_cast<T2>(m[I])...};
 }
 
 template <class T2, class T1, std::size_t N, std::size_t M>
-matrix<T2, N, M> type_cast(const matrix<T1, N, M>& m)
+constexpr matrix<T2, N, M> type_cast(const matrix<T1, N, M>& m)
 {
 	return type_cast<T2>(m, std::make_index_sequence<N>{}); 
 }
diff --git a/src/math/matrix-operators.hpp b/src/math/matrix-operators.hpp
index 316b16f..f885f6c 100644
--- a/src/math/matrix-operators.hpp
+++ b/src/math/matrix-operators.hpp
@@ -25,60 +25,60 @@
 
 /// @copydoc math::add(const math::matrix<T, 2, 2>&, const math::matrix<T, 2, 2>&)
 template <class T, std::size_t N, std::size_t M>
-math::matrix<T, N, M> operator+(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y);
+constexpr math::matrix<T, N, M> operator+(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y);
 
 /// @copydoc math::mul(const math::matrix<T, 2, 2>&, const math::matrix<T, 2, 2>&)
 template <class T, std::size_t N, std::size_t M>
-math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y);
+constexpr math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y);
 
 /// @copydoc math::mul(const math::matrix<T, N, M>&, T)
 template <class T, std::size_t N, std::size_t M>
-math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& m, T s);
+constexpr math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& m, T s);
 
 /// @copydoc math::mul(const math::matrix<T, N, M>&, T)
 template <class T, std::size_t N, std::size_t M>
-math::matrix<T, N, M> operator*(T s, const math::matrix<T, N, M>& m);
+constexpr math::matrix<T, N, M> operator*(T s, const math::matrix<T, N, M>& m);
 
 /// @copydoc math::mul(const math::matrix<T, 2, 2>&, const math::vector<T, 2>&)
 template <class T, std::size_t N>
-math::vector<T, N> operator*(const math::matrix<T, N, N>& m, const math::vector<T, N>& v);
+constexpr math::vector<T, N> operator*(const math::matrix<T, N, N>& m, const math::vector<T, N>& v);
 
 /// @copydoc math::sub(const math::matrix<T, 2, 2>&, const math::matrix<T, 2, 2>&)
 template <class T, std::size_t N, std::size_t M>
-math::matrix<T, N, M> operator-(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y);
+constexpr math::matrix<T, N, M> operator-(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y);
 
 template <class T, std::size_t N, std::size_t M>
-inline math::matrix<T, N, M> operator+(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y)
+constexpr inline math::matrix<T, N, M> operator+(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y)
 {
 	return add(x, y);
 }
 
 template <class T, std::size_t N, std::size_t M>
-inline math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y)
+constexpr inline math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y)
 {
 	return mul(x, y);
 }
 
 template <class T, std::size_t N, std::size_t M>
-inline math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& m, T s)
+constexpr inline math::matrix<T, N, M> operator*(const math::matrix<T, N, M>& m, T s)
 {
 	return mul(m, s);
 }
 
 template <class T, std::size_t N, std::size_t M>
-inline math::matrix<T, N, M> operator*(T s, const math::matrix<T, N, M>& m)
+constexpr inline math::matrix<T, N, M> operator*(T s, const math::matrix<T, N, M>& m)
 {
 	return mul(m, s);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator*(const math::matrix<T, N, N>& m, const math::vector<T, N>& v)
+constexpr inline math::vector<T, N> operator*(const math::matrix<T, N, N>& m, const math::vector<T, N>& v)
 {
 	return mul(m, v);
 }
 
 template <class T, std::size_t N, std::size_t M>
-inline math::matrix<T, N, M> operator-(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y)
+constexpr inline math::matrix<T, N, M> operator-(const math::matrix<T, N, M>& x, const math::matrix<T, N, M>& y)
 {
 	return sub(x, y);
 }
diff --git a/src/math/matrix-type.hpp b/src/math/matrix-type.hpp
index ad2576e..0ecdd04 100644
--- a/src/math/matrix-type.hpp
+++ b/src/math/matrix-type.hpp
@@ -66,14 +66,26 @@ constexpr matrix<T, N, M> matrix<T, N, M>::identity = identity_matrix<T>(std::ma
 template <typename T>
 using matrix2 = matrix<T, 2, 2>;
 
+/// 2x2 matrix.
+template <typename T>
+using matrix2x2 = matrix<T, 2, 2>;
+
 /// 3x3 matrix.
 template <typename T>
 using matrix3 = matrix<T, 3, 3>;
 
+/// 3x3 matrix.
+template <typename T>
+using matrix3x3 = matrix<T, 3, 3>;
+
 /// 4x4 matrix.
 template <typename T>
 using matrix4 = matrix<T, 4, 4>;
 
+/// 4x4 matrix.
+template <typename T>
+using matrix4x4 = matrix<T, 4, 4>;
+
 } // namespace math
 
 #endif // ANTKEEPER_MATH_MATRIX_TYPE_HPP
diff --git a/src/math/vector-functions.hpp b/src/math/vector-functions.hpp
index 8281bcc..2a541b0 100644
--- a/src/math/vector-functions.hpp
+++ b/src/math/vector-functions.hpp
@@ -36,7 +36,7 @@ namespace math {
  * @return Sum of the two vectors.
  */
 template <class T, std::size_t N>
-vector<T, N> add(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<T, N> add(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Checks if all elements of a boolean vector are `true`.
@@ -45,7 +45,7 @@ vector<T, N> add(const vector<T, N>& x, const vector<T, N>& y);
  * @return `true` if all elements are `true`, `false` otherwise.
  */
 template <std::size_t N>
-bool all(const vector<bool, N>& x);
+constexpr bool all(const vector<bool, N>& x);
 
 /**
  * Checks if any elements of a boolean vector are `true`.
@@ -54,7 +54,7 @@ bool all(const vector<bool, N>& x);
  * @return `true` if any elements are `true`, `false` otherwise.
  */
 template <std::size_t N>
-bool any(const vector<bool, N>& x);
+constexpr bool any(const vector<bool, N>& x);
 
 /**
  * Reinterprets data as an `N`-dimensional vector of type `T`.
@@ -64,7 +64,7 @@ bool any(const vector<bool, N>& x);
  * @param data Data to reinterpret.
  */
 template <std::size_t N, typename T>
-vector<T, N>& as_vector(T& data);
+constexpr vector<T, N>& as_vector(T& data);
 
 /**
  * Clamps the values of a vector's elements.
@@ -75,7 +75,7 @@ vector<T, N>& as_vector(T& data);
  * @return Clamped vector.
  */
 template <class T, std::size_t N>
-vector<T, N> clamp(const vector<T, N>& x, T min_value, T max_value);
+constexpr vector<T, N> clamp(const vector<T, N>& x, T min_value, T max_value);
 
 /**
  * Clamps the length of a vector.
@@ -95,7 +95,7 @@ vector<T, N> clamp_length(const vector<T, N>& x, T max_length);
  * @return Cross product of the two vectors.
  */
 template <class T>
-vector<T, 3> cross(const vector<T, 3>& x, const vector<T, 3>& y);
+constexpr vector<T, 3> cross(const vector<T, 3>& x, const vector<T, 3>& y);
 
 /**
  * Calculates the distance between two points.
@@ -115,7 +115,7 @@ vector<T, N> distance(const vector<T, N>& p0, const vector<T, N>& p1);
  * @return Squared distance between the two points.
  */
 template <class T, std::size_t N>
-vector<T, N> distance_squared(const vector<T, N>& p0, const vector<T, N>& p1);
+constexpr vector<T, N> distance_squared(const vector<T, N>& p0, const vector<T, N>& p1);
 
 /**
  * Divides a vector by another vector.
@@ -125,7 +125,7 @@ vector<T, N> distance_squared(const vector<T, N>& p0, const vector<T, N>& p1);
  * @return Result of the division.
  */
 template <class T, std::size_t N>
-vector<T, N> div(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<T, N> div(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Divides a vector by a scalar.
@@ -135,7 +135,7 @@ vector<T, N> div(const vector<T, N>& x, const vector<T, N>& y);
  * @return Result of the division.
  */
 template <class T, std::size_t N>
-vector<T, N> div(const vector<T, N>& v, T s);
+constexpr vector<T, N> div(const vector<T, N>& v, T s);
 
 /**
  * Calculates the dot product of two vectors.
@@ -145,7 +145,7 @@ vector<T, N> div(const vector<T, N>& v, T s);
  * @return Dot product of the two vectors.
  */
 template <class T, std::size_t N>
-T dot(const vector<T, N>& x, const vector<T, N>& y);
+constexpr T dot(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Compares two vectors for equality
@@ -155,7 +155,7 @@ T dot(const vector<T, N>& x, const vector<T, N>& y);
  * @return Boolean vector containing the result of the element comparisons.
  */
 template <class T, std::size_t N>
-vector<bool, N> equal(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<bool, N> equal(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Performs a component-wise greater-than comparison of two vectors.
@@ -165,7 +165,7 @@ vector<bool, N> equal(const vector<T, N>& x, const vector<T, N>& y);
  * @return Boolean vector containing the result of the element comparisons.
  */
 template <class T, std::size_t N>
-vector<bool, N> greater_than(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<bool, N> greater_than(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Performs a component-wise greater-than or equal-to comparison of two vectors.
@@ -175,7 +175,7 @@ vector<bool, N> greater_than(const vector<T, N>& x, const vector<T, N>& y);
  * @return Boolean vector containing the result of the element comparisons.
  */
 template <class T, std::size_t N>
-vector<bool, N> greater_than_equal(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<bool, N> greater_than_equal(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Calculates the length of a vector.
@@ -193,7 +193,7 @@ T length(const vector<T, N>& x);
  * @return Squared length of the vector.
  */
 template <class T, std::size_t N>
-T length_squared(const vector<T, N>& x);
+constexpr T length_squared(const vector<T, N>& x);
 
 /**
  * Performs a component-wise less-than comparison of two vectors.
@@ -203,7 +203,7 @@ T length_squared(const vector<T, N>& x);
  * @return Boolean vector containing the result of the element comparisons.
  */
 template <class T, std::size_t N>
-vector<bool, N> less_than(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<bool, N> less_than(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Performs a component-wise less-than or equal-to comparison of two vectors.
@@ -213,7 +213,7 @@ vector<bool, N> less_than(const vector<T, N>& x, const vector<T, N>& y);
  * @return Boolean vector containing the result of the element comparisons.
  */
 template <class T, std::size_t N>
-vector<bool, N> less_than_equal(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<bool, N> less_than_equal(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Multiplies two vectors.
@@ -223,7 +223,7 @@ vector<bool, N> less_than_equal(const vector<T, N>& x, const vector<T, N>& y);
  * @return Product of the two vectors.
  */
 template <class T, std::size_t N>
-vector<T, N> mul(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<T, N> mul(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Multiplies a vector by a scalar.
@@ -233,7 +233,7 @@ vector<T, N> mul(const vector<T, N>& x, const vector<T, N>& y);
  * @return Product of the vector and scalar.
  */
 template <class T, std::size_t N>
-vector<T, N> mul(const vector<T, N>& v, T s);
+constexpr vector<T, N> mul(const vector<T, N>& v, T s);
 
 /**
  * Negates a vector.
@@ -242,7 +242,7 @@ vector<T, N> mul(const vector<T, N>& v, T s);
  * @return Negated vector.
  */
 template <class T, std::size_t N>
-vector<T, N> negate(const vector<T, N>& x);
+constexpr vector<T, N> negate(const vector<T, N>& x);
 
 /**
  * Calculates the unit vector in the same direction as the original vector.
@@ -260,7 +260,7 @@ vector<T, N> normalize(const vector<T, N>& x);
  * @return Logically inverted vector.
  */
 template <class T, std::size_t N>
-vector<bool, N> not(const vector<T, N>& x);
+constexpr vector<bool, N> not(const vector<T, N>& x);
 
 /**
  * Compares two vectors for inequality
@@ -270,7 +270,7 @@ vector<bool, N> not(const vector<T, N>& x);
  * @return Boolean vector containing the result of the element comparisons.
  */
 template <class T, std::size_t N>
-vector<bool, N> not_equal(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<bool, N> not_equal(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Resizes a vector. Any new elements will be set to `0`.
@@ -279,7 +279,7 @@ vector<bool, N> not_equal(const vector<T, N>& x, const vector<T, N>& y);
  * @return Resized vector.
  */
 template <std::size_t N1, class T, std::size_t N0>
-vector<T, N1> resize(const vector<T, N0>& v);
+constexpr vector<T, N1> resize(const vector<T, N0>& v);
 
 /**
  * Subtracts a vector from another vector.
@@ -289,7 +289,7 @@ vector<T, N1> resize(const vector<T, N0>& v);
  * @return Difference between the two vectors.
  */
 template <class T, std::size_t N>
-vector<T, N> sub(const vector<T, N>& x, const vector<T, N>& y);
+constexpr vector<T, N> sub(const vector<T, N>& x, const vector<T, N>& y);
 
 /**
  * Makes an m-dimensional vector by rearranging and/or duplicating elements of an n-dimensional vector.
@@ -300,7 +300,7 @@ vector<T, N> sub(const vector<T, N>& x, const vector<T, N>& y);
  * @return Vector containing elements from vector `v` in the order specified by `Indices`. The size of the returned vector is equivalent to the number of indices in `Indices`.
  */
 template <std::size_t... Indices, class T, std::size_t N>
-vector<T, sizeof...(Indices)> swizzle(const vector<T, N>& v);
+constexpr vector<T, sizeof...(Indices)> swizzle(const vector<T, N>& v);
 
 /**
  * Types casts each vector component and returns a vector of the casted type.
@@ -312,49 +312,49 @@ vector<T, sizeof...(Indices)> swizzle(const vector<T, N>& v);
  * @return Type-casted vector.
  */
 template <class T2, class T1, std::size_t N>
-vector<T2, N> type_cast(const vector<T1, N>& v);
+constexpr vector<T2, N> type_cast(const vector<T1, N>& v);
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> add(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<T, N> add(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] + y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> add(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<T, N> add(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return add(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <std::size_t N, std::size_t... I>
-inline bool all(const vector<bool, N>& x, std::index_sequence<I...>)
+constexpr inline bool all(const vector<bool, N>& x, std::index_sequence<I...>)
 {
 	return (x[I] && ...);
 }
 
 template <std::size_t N>
-inline bool all(const vector<bool, N>& x)
+constexpr inline bool all(const vector<bool, N>& x)
 {
 	return all(x, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <std::size_t N, std::size_t... I>
-inline bool any(const vector<bool, N>& x, std::index_sequence<I...>)
+constexpr inline bool any(const vector<bool, N>& x, std::index_sequence<I...>)
 {
 	return (x[I] || ...);
 }
 
 template <std::size_t N>
-inline bool any(const vector<bool, N>& x)
+constexpr inline bool any(const vector<bool, N>& x)
 {
 	return any(x, std::make_index_sequence<N>{});
 }
 
 template <std::size_t N, typename T>
-inline vector<T, N>& as_vector(T& data)
+constexpr inline vector<T, N>& as_vector(T& data)
 {
 	static_assert(std::is_pod<vector<T, N>>::value);
 	return reinterpret_cast<vector<T, N>&>(data);
@@ -362,13 +362,13 @@ inline vector<T, N>& as_vector(T& data)
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> clamp(const vector<T, N>& x, T min_value, T max_value, std::index_sequence<I...>)
+constexpr inline vector<T, N> clamp(const vector<T, N>& x, T min_value, T max_value, std::index_sequence<I...>)
 {
 	return {std::min<T>(max_value, std::max<T>(min_value, x[I]))...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> clamp(const vector<T, N>& x, T min_value, T max_value)
+constexpr inline vector<T, N> clamp(const vector<T, N>& x, T min_value, T max_value)
 {
 	return clamp(x, min_value, max_value, std::make_index_sequence<N>{});
 }
@@ -381,7 +381,7 @@ vector<T, N> clamp_length(const vector<T, N>& x, T max_length)
 }
 
 template <class T>
-inline vector<T, 3> cross(const vector<T, 3>& x, const vector<T, 3>& y)
+constexpr inline vector<T, 3> cross(const vector<T, 3>& x, const vector<T, 3>& y)
 {
 	return
 		{
@@ -399,7 +399,7 @@ inline vector<T, N> distance(const vector<T, N>& p0, const vector<T, N>& p1)
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> distance_squared(const vector<T, N>& p0, const vector<T, N>& p1)
+constexpr inline vector<T, N> distance_squared(const vector<T, N>& p0, const vector<T, N>& p1)
 {
 	static_assert(std::is_floating_point<T>::value);
 	return length_squared(sub(p0, p1));
@@ -407,78 +407,78 @@ inline vector<T, N> distance_squared(const vector<T, N>& p0, const vector<T, N>&
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> div(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<T, N> div(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] / y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> div(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<T, N> div(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return div(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> div(const vector<T, N>& v, T s, std::index_sequence<I...>)
+constexpr inline vector<T, N> div(const vector<T, N>& v, T s, std::index_sequence<I...>)
 {
 	return {(v[I] / s)...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> div(const vector<T, N>& v, T s)
+constexpr inline vector<T, N> div(const vector<T, N>& v, T s)
 {
 	return div(v, s, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline T dot(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline T dot(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return ((x[I] * y[I]) + ...);
 }
 
 template <class T, std::size_t N>
-inline T dot(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline T dot(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return dot(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<bool, N> equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<bool, N> equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] == y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<bool, N> equal(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<bool, N> equal(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return equal(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<bool, N> greater_than(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<bool, N> greater_than(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] > y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<bool, N> greater_than(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<bool, N> greater_than(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return greater_than(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<bool, N> greater_than_equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<bool, N> greater_than_equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] >= y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<bool, N> greater_than_equal(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<bool, N> greater_than_equal(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return greater_than_equal(x, y, std::make_index_sequence<N>{});
 }
@@ -491,73 +491,72 @@ inline T length(const vector<T, N>& x)
 }
 
 template <class T, std::size_t N>
-inline T length_squared(const vector<T, N>& x)
+constexpr inline T length_squared(const vector<T, N>& x)
 {
-	static_assert(std::is_floating_point<T>::value);
 	return dot(x, x);
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<bool, N> less_than(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<bool, N> less_than(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] < y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<bool, N> less_than(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<bool, N> less_than(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return less_than(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<bool, N> less_than_equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<bool, N> less_than_equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] <= y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<bool, N> less_than_equal(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<bool, N> less_than_equal(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return less_than_equal(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> mul(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<T, N> mul(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] * y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> mul(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<T, N> mul(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return mul(x, y, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> mul(const vector<T, N>& v, T s, std::index_sequence<I...>)
+constexpr inline vector<T, N> mul(const vector<T, N>& v, T s, std::index_sequence<I...>)
 {
 	return {(v[I] * s)...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> mul(const vector<T, N>& v, T s)
+constexpr inline vector<T, N> mul(const vector<T, N>& v, T s)
 {
 	return mul(v, s, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> negate(const vector<T, N>& x, std::index_sequence<I...>)
+constexpr inline vector<T, N> negate(const vector<T, N>& x, std::index_sequence<I...>)
 {
 	return {(-x[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> negate(const vector<T, N>& x)
+constexpr inline vector<T, N> negate(const vector<T, N>& x)
 {
 	return negate(x, std::make_index_sequence<N>{});
 }
@@ -571,32 +570,32 @@ inline vector<T, N> normalize(const vector<T, N>& x)
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<bool, N> not(const vector<T, N>& x, std::index_sequence<I...>)
+constexpr inline vector<bool, N> not(const vector<T, N>& x, std::index_sequence<I...>)
 {
 	return {!x[I]...};
 }
 
 template <class T, std::size_t N>
-inline vector<bool, N> not(const vector<T, N>& x)
+constexpr inline vector<bool, N> not(const vector<T, N>& x)
 {
 	return not(x, std::make_index_sequence<N>{});
 }
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<bool, N> not_equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<bool, N> not_equal(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] != y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<bool, N> not_equal(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<bool, N> not_equal(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return not_equal(x, y, std::make_index_sequence<N>{});
 }
 
 template <std::size_t N1, class T, std::size_t N0>
-vector<T, N1> resize(const vector<T, N0>& v)
+constexpr vector<T, N1> resize(const vector<T, N0>& v)
 {
 	vector<T, N1> resized;
 
@@ -611,32 +610,32 @@ vector<T, N1> resize(const vector<T, N0>& v)
 
 /// @private
 template <class T, std::size_t N, std::size_t... I>
-inline vector<T, N> sub(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
+constexpr inline vector<T, N> sub(const vector<T, N>& x, const vector<T, N>& y, std::index_sequence<I...>)
 {
 	return {(x[I] - y[I])...};
 }
 
 template <class T, std::size_t N>
-inline vector<T, N> sub(const vector<T, N>& x, const vector<T, N>& y)
+constexpr inline vector<T, N> sub(const vector<T, N>& x, const vector<T, N>& y)
 {
 	return sub(x, y, std::make_index_sequence<N>{});
 }
 
 template <std::size_t... Indices, class T, std::size_t N>
-inline vector<T, sizeof...(Indices)> swizzle(const vector<T, N>& v)
+constexpr inline vector<T, sizeof...(Indices)> swizzle(const vector<T, N>& v)
 {
 	return { v[Indices]... };
 }
 
 /// @private
 template <class T2, class T1, std::size_t N, std::size_t... I>
-inline vector<T2, N> type_cast(const vector<T1, N>& v, std::index_sequence<I...>)
+constexpr inline vector<T2, N> type_cast(const vector<T1, N>& v, std::index_sequence<I...>)
 {
 	return {static_cast<T2>(v[I])...};
 }
 
 template <class T2, class T1, std::size_t N>
-inline vector<T2, N> type_cast(const vector<T1, N>& v)
+constexpr inline vector<T2, N> type_cast(const vector<T1, N>& v)
 {
 	return type_cast<T2>(v, std::make_index_sequence<N>{}); 
 }
@@ -644,4 +643,3 @@ inline vector<T2, N> type_cast(const vector<T1, N>& v)
 } // namespace math
 
 #endif // ANTKEEPER_MATH_VECTOR_FUNCTIONS_HPP
-
diff --git a/src/math/vector-operators.hpp b/src/math/vector-operators.hpp
index af7f899..b4d4b9e 100644
--- a/src/math/vector-operators.hpp
+++ b/src/math/vector-operators.hpp
@@ -25,35 +25,35 @@
 
 /// @copydoc math::add(const math::vector<T, N>&, const math::vector<T, N>&)
 template <class T, std::size_t N>
-math::vector<T, N> operator+(const math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N> operator+(const math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /// @copydoc math::div(const math::vector<T, N>&, const math::vector<T, N>&)
 template <class T, std::size_t N>
-math::vector<T, N> operator/(const math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N> operator/(const math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /// @copydoc math::div(const math::vector<T, N>&, T)
 template <class T, std::size_t N>
-math::vector<T, N> operator/(const math::vector<T, N>& v, T s);
+constexpr math::vector<T, N> operator/(const math::vector<T, N>& v, T s);
 
 /// @copydoc math::mul(const math::vector<T, N>&, const math::vector<T, N>&)
 template <class T, std::size_t N>
-math::vector<T, N> operator*(const math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N> operator*(const math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /// @copydoc math::mul(const math::vector<T, N>&, T)
 template <class T, std::size_t N>
-math::vector<T, N> operator*(const math::vector<T, N>& v, T s);
+constexpr math::vector<T, N> operator*(const math::vector<T, N>& v, T s);
 
 /// @copydoc math::mul(const math::vector<T, N>&, T)
 template <class T, std::size_t N>
-math::vector<T, N> operator*(T s, const math::vector<T, N>& v);
+constexpr math::vector<T, N> operator*(T s, const math::vector<T, N>& v);
 
 /// @copydoc math::negate(const math::vector<T, N>&)
 template <class T, std::size_t N>
-math::vector<T, N> operator-(const math::vector<T, N>& x);
+constexpr math::vector<T, N> operator-(const math::vector<T, N>& x);
 
 /// @copydoc math::sub(const math::vector<T, N>&, const math::vector<T, N>&)
 template <class T, std::size_t N>
-math::vector<T, N> operator-(const math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N> operator-(const math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /**
  * Adds two vectors and stores the result in the first vector.
@@ -63,7 +63,7 @@ math::vector<T, N> operator-(const math::vector<T, N>& x, const math::vector<T,
  * @return Reference to the first vector.
  */
 template <class T, std::size_t N>
-math::vector<T, N>& operator+=(math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N>& operator+=(math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /**
  * Subtracts two vectors and stores the result in the first vector.
@@ -73,7 +73,7 @@ math::vector<T, N>& operator+=(math::vector<T, N>& x, const math::vector<T, N>&
  * @return Reference to the first vector.
  */
 template <class T, std::size_t N>
-math::vector<T, N>& operator-=(math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N>& operator-=(math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /**
  * Multiplies two vectors and stores the result in the first vector.
@@ -83,7 +83,7 @@ math::vector<T, N>& operator-=(math::vector<T, N>& x, const math::vector<T, N>&
  * @return Reference to the first vector.
  */
 template <class T, std::size_t N>
-math::vector<T, N>& operator*=(math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N>& operator*=(math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /**
  * Multiplies a vector and a scalar and stores the result in the vector.
@@ -93,7 +93,7 @@ math::vector<T, N>& operator*=(math::vector<T, N>& x, const math::vector<T, N>&
  * @return Reference to the vector.
  */
 template <class T, std::size_t N>
-math::vector<T, N>& operator*=(math::vector<T, N>& v, T s);
+constexpr math::vector<T, N>& operator*=(math::vector<T, N>& v, T s);
 
 /**
  * Divides the first vector by the second vector the result in the first vector.
@@ -103,7 +103,7 @@ math::vector<T, N>& operator*=(math::vector<T, N>& v, T s);
  * @return Reference to the first vector.
  */
 template <class T, std::size_t N>
-math::vector<T, N>& operator/=(math::vector<T, N>& x, const math::vector<T, N>& y);
+constexpr math::vector<T, N>& operator/=(math::vector<T, N>& x, const math::vector<T, N>& y);
 
 /**
  * Divides a vector by a scalar and stores the result in the vector.
@@ -113,91 +113,90 @@ math::vector<T, N>& operator/=(math::vector<T, N>& x, const math::vector<T, N>&
  * @return Reference to the vector.
  */
 template <class T, std::size_t N>
-math::vector<T, N>& operator/=(math::vector<T, N>& v, T s);
+constexpr math::vector<T, N>& operator/=(math::vector<T, N>& v, T s);
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator+(const math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N> operator+(const math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return add(x, y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator-(const math::vector<T, N>& x)
+constexpr inline math::vector<T, N> operator-(const math::vector<T, N>& x)
 {
 	return negate(x);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator-(const math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N> operator-(const math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return sub(x, y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator*(const math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N> operator*(const math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return mul(x, y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator*(const math::vector<T, N>& v, T s)
+constexpr inline math::vector<T, N> operator*(const math::vector<T, N>& v, T s)
 {
 	return mul(v, s);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator*(T s, const math::vector<T, N>& v)
+constexpr inline math::vector<T, N> operator*(T s, const math::vector<T, N>& v)
 {
 	return mul(v, s);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator/(const math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N> operator/(const math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return div(x, y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N> operator/(const math::vector<T, N>& v, T s)
+constexpr inline math::vector<T, N> operator/(const math::vector<T, N>& v, T s)
 {
 	return div(v, s);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N>& operator+=(math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N>& operator+=(math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return (x = x + y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N>& operator-=(math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N>& operator-=(math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return (x = x - y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N>& operator*=(math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N>& operator*=(math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return (x = x * y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N>& operator*=(math::vector<T, N>& v, T s)
+constexpr inline math::vector<T, N>& operator*=(math::vector<T, N>& v, T s)
 {
 	return (v = v * s);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N>& operator/=(math::vector<T, N>& x, const math::vector<T, N>& y)
+constexpr inline math::vector<T, N>& operator/=(math::vector<T, N>& x, const math::vector<T, N>& y)
 {
 	return (x = x * y);
 }
 
 template <class T, std::size_t N>
-inline math::vector<T, N>& operator/=(math::vector<T, N>& v, T s)
+constexpr inline math::vector<T, N>& operator/=(math::vector<T, N>& v, T s)
 {
 	return (v = v / s);
 }
 
 #endif // ANTKEEPER_MATH_VECTOR_OPERATORS_HPP
-
diff --git a/src/math/vector-type.hpp b/src/math/vector-type.hpp
index 84f5a84..2d7c4a4 100644
--- a/src/math/vector-type.hpp
+++ b/src/math/vector-type.hpp
@@ -20,7 +20,6 @@
 #ifndef ANTKEEPER_MATH_VECTOR_TYPE_HPP
 #define ANTKEEPER_MATH_VECTOR_TYPE_HPP
 
-#include <array>
 #include <cstddef>
 
 namespace math {
@@ -34,87 +33,107 @@ namespace math {
 template <typename T, std::size_t N>
 struct vector
 {
-	typedef T scalar_type;
-	typedef std::array<T, N> array_type;
-	scalar_type components[N];
-
-	inline constexpr scalar_type& operator[](std::size_t i) noexcept { return components[i]; }
-	inline constexpr const scalar_type& operator[](std::size_t i) const noexcept { return components[i]; }
-	inline constexpr operator array_type&() noexcept { return reinterpret_cast<array_type&>(components[0]); }
-	inline constexpr operator const array_type&() const noexcept { return reinterpret_cast<const array_type&>(components[0]); }
-	inline constexpr const scalar_type* data() const noexcept { return components; };
-	inline constexpr scalar_type* data() noexcept { return components; };
+	typedef T element_type;
+	element_type elements[N];
+
+	inline constexpr element_type& operator[](std::size_t i) noexcept { return elements[i]; }
+	inline constexpr const element_type& operator[](std::size_t i) const noexcept { return elements[i]; }
+	inline constexpr const element_type* data() const noexcept { return elements; };
+	inline constexpr element_type* data() noexcept { return elements; };
 	inline constexpr std::size_t size() const noexcept { return N; };
 };
 
 template <typename T>
 struct vector<T, 1>
 {
-	typedef T scalar_type;
-	typedef std::array<T, 1> array_type;
-	scalar_type x;
-
-	inline constexpr scalar_type& operator[](std::size_t i) noexcept { return *((&x) + i); }
-	inline constexpr const scalar_type& operator[](std::size_t i) const noexcept { return *((&x) + i); }
-	inline constexpr operator array_type&() noexcept { return reinterpret_cast<array_type&>(x); }
-	inline constexpr operator const array_type&() const noexcept { return reinterpret_cast<const array_type&>(x); }
-	inline constexpr const scalar_type* data() const noexcept { return &x; };
-	inline constexpr scalar_type* data() noexcept { return &x; };
+	typedef T element_type;
+	
+	union
+	{
+		element_type elements[1];
+		
+		struct
+		{
+			element_type x;
+		};
+	};
+
+	inline constexpr element_type& operator[](std::size_t i) noexcept { return elements[i]; }
+	inline constexpr const element_type& operator[](std::size_t i) const noexcept { return elements[i]; }
+	inline constexpr const element_type* data() const noexcept { return elements; };
+	inline constexpr element_type* data() noexcept { return elements; };
 	inline constexpr std::size_t size() const noexcept { return 1; };
 };
 
 template <typename T>
 struct vector<T, 2>
 {
-	typedef T scalar_type;
-	typedef std::array<T, 2> array_type;
-	scalar_type x;
-	scalar_type y;
-
-	inline constexpr scalar_type& operator[](std::size_t i) noexcept { return *((&x) + i); }
-	inline constexpr const scalar_type& operator[](std::size_t i) const noexcept { return *((&x) + i); }
-	inline constexpr operator array_type&() noexcept { return reinterpret_cast<array_type&>(x); }
-	inline constexpr operator const array_type&() const noexcept { return reinterpret_cast<const array_type&>(x); }
-	inline constexpr const scalar_type* data() const noexcept { return &x; };
-	inline constexpr scalar_type* data() noexcept { return &x; };
+	typedef T element_type;
+	
+	union
+	{
+		element_type elements[2];
+		
+		struct
+		{
+			element_type x;
+			element_type y;
+		};
+	};
+
+	inline constexpr element_type& operator[](std::size_t i) noexcept { return elements[i]; }
+	inline constexpr const element_type& operator[](std::size_t i) const noexcept { return elements[i]; }
+	inline constexpr const element_type* data() const noexcept { return elements; };
+	inline constexpr element_type* data() noexcept { return elements; };
 	inline constexpr std::size_t size() const noexcept { return 2; };
 };
 
 template <typename T>
 struct vector<T, 3>
 {
-	typedef T scalar_type;
-	typedef std::array<T, 3> array_type;
-	scalar_type x;
-	scalar_type y;
-	scalar_type z;
-
-	inline constexpr scalar_type& operator[](std::size_t i) noexcept { return *((&x) + i); }
-	inline constexpr const scalar_type& operator[](std::size_t i) const noexcept { return *((&x) + i); }
-	inline constexpr operator array_type&() noexcept { return reinterpret_cast<array_type&>(x); }
-	inline constexpr operator const array_type&() const noexcept { return reinterpret_cast<const array_type&>(x); }
-	inline constexpr const scalar_type* data() const noexcept { return &x; };
-	inline constexpr scalar_type* data() noexcept { return &x; };
+	typedef T element_type;
+	
+	union
+	{
+		element_type elements[3];
+		
+		struct
+		{
+			element_type x;
+			element_type y;
+			element_type z;
+		};
+	};
+
+	inline constexpr element_type& operator[](std::size_t i) noexcept { return elements[i]; }
+	inline constexpr const element_type& operator[](std::size_t i) const noexcept { return elements[i]; }
+	inline constexpr const element_type* data() const noexcept { return elements; };
+	inline constexpr element_type* data() noexcept { return elements; };
 	inline constexpr std::size_t size() const noexcept { return 3; };
-
 };
 
 template <typename T>
 struct vector<T, 4>
 {
-	typedef T scalar_type;
-	typedef std::array<T, 4> array_type;
-	scalar_type x;
-	scalar_type y;
-	scalar_type z;
-	scalar_type w;
-
-	inline constexpr scalar_type& operator[](std::size_t i) noexcept { return *((&x) + i); }
-	inline constexpr const scalar_type& operator[](std::size_t i) const noexcept { return *((&x) + i); }
-	inline constexpr operator array_type&() noexcept { return reinterpret_cast<array_type&>(x); }
-	inline constexpr operator const array_type&() const noexcept { return reinterpret_cast<const array_type&>(x); }
-	inline constexpr const scalar_type* data() const noexcept { return &x; };
-	inline constexpr scalar_type* data() noexcept { return &x; };
+	typedef T element_type;
+	
+	union
+	{
+		element_type elements[4];
+		
+		struct
+		{
+			element_type x;
+			element_type y;
+			element_type z;
+			element_type w;
+		};
+	};
+
+	inline constexpr element_type& operator[](std::size_t i) noexcept { return elements[i]; }
+	inline constexpr const element_type& operator[](std::size_t i) const noexcept { return elements[i]; }
+	inline constexpr const element_type* data() const noexcept { return elements; };
+	inline constexpr element_type* data() noexcept { return elements; };
 	inline constexpr std::size_t size() const noexcept { return 4; };
 };
 
@@ -133,4 +152,3 @@ using vector4 = vector<T, 4>;
 } // namespace math
 
 #endif // ANTKEEPER_MATH_VECTOR_TYPE_HPP
-
diff --git a/src/physics/gas/atmosphere.hpp b/src/physics/gas/atmosphere.hpp
index 9420fc1..cd92348 100644
--- a/src/physics/gas/atmosphere.hpp
+++ b/src/physics/gas/atmosphere.hpp
@@ -32,101 +32,95 @@ namespace gas {
 namespace atmosphere {
 
 /**
- * Calculates a particle polarizability factor used in computing scattering coefficients.
+ * Calculates a particle polarizability factor.
  *
  * @param ior Atmospheric index of refraction.
- * @param density Molecular density.
- * @return Polarization factor.
+ * @param density Molecular number density, in mol/m-3.
+ * @return Polarizability factor.
  *
+ * @see Elek, O., & Kmoch, P. (2010). Real-time spectral scattering in large-scale natural participating media. Proceedings of the 26th Spring Conference on Computer Graphics - SCCG ’10. doi:10.1145/1925059.1925074
  * @see Elek, Oskar. (2009). Rendering Parametrizable Planetary Atmospheres with Multiple Scattering in Real-Time.
- * @see Real-Time Spectral Scattering in Large-Scale Natural Participating Media.
  */
 template <class T>
 T polarization(T ior, T density)
 {
-	const T ior2m1 = ior * ior - T(1.0);
-	const T num = T(2) * math::pi<T> * math::pi<T> * ior2m1 * ior2m1;
+	constexpr T k = T(2) * math::pi<T> * math::pi<T>;
+	const T ior2m1 = ior * ior - T(1);
+	const T num = k * ior2m1 * ior2m1;
 	const T den = T(3) * density * density;
 	return num / den;
 }
 
 /**
- * Calculates a Rayleigh scattering coefficient (wavelength-dependent).
+ * Calculates a wavelength-dependent scattering coefficient.
  *
+ * @param density Molecular number density of the particles, in mol/m-3.
+ * @param polarization Particle polarizability factor.
  * @param wavelength Wavelength of light, in meters.
- * @param density Molecular density of Rayleigh particles.
- * @param polarization Rayleigh particle polarization factor.
+ *
+ * @return Scattering coefficient.
  *
  * @see atmosphere::polarization
  *
+ * @see Elek, O., & Kmoch, P. (2010). Real-time spectral scattering in large-scale natural participating media. Proceedings of the 26th Spring Conference on Computer Graphics - SCCG ’10. doi:10.1145/1925059.1925074
  * @see Elek, Oskar. (2009). Rendering Parametrizable Planetary Atmospheres with Multiple Scattering in Real-Time.
- * @see Real-Time Spectral Scattering in Large-Scale Natural Participating Media.
  */
 template <class T>
-T scattering_rayleigh(T wavelength, T density, T polarization)
+T scattering(T density, T polarization, T wavelength)
 {
 	const T wavelength2 = wavelength * wavelength;
-	return math::four_pi<T> * density / (wavelength2 * wavelength2) * polarization;
+	return math::four_pi<T> * (density / (wavelength2 * wavelength2)) * polarization;
 }
 
 /**
- * Calculates a Mie scattering coefficient (wavelength-independent).
+ * Calculates a wavelength-independent scattering coefficient.
  *
- * @param density Molecular density of Mie particles.
- * @param polarization Mie particle polarization factor.
+ * @param density Molecular number density of the particles, in mol/m-3.
+ * @param polarization Particle polarizability factor.
  *
- * @return Mie scattering coefficient.
+ * @return Scattering coefficient.
  *
  * @see atmosphere::polarization
  *
+ * @see Elek, O., & Kmoch, P. (2010). Real-time spectral scattering in large-scale natural participating media. Proceedings of the 26th Spring Conference on Computer Graphics - SCCG ’10. doi:10.1145/1925059.1925074
  * @see Elek, Oskar. (2009). Rendering Parametrizable Planetary Atmospheres with Multiple Scattering in Real-Time.
- * @see Real-Time Spectral Scattering in Large-Scale Natural Participating Media.
  */
 template <class T>
-T scattering_mie(T density, T polarization)
+T scattering(T density, T polarization)
 {
 	return math::four_pi<T> * density * polarization;
 }
 
 /**
- * Calculates a Mie absorption coefficient (wavelength-independent).
+ * Calculates an absorption coefficient.
  *
- * @param scattering Mie scattering coefficient.
+ * @param scattering Scattering coefficient.
+ * @param albedo Single-scattering albedo.
  *
- * @return Mie absorption coefficient.
+ * @return Absorption coefficient.
  *
- * @see Bruneton, E. and Neyret, F. (2008), Precomputed Atmospheric Scattering. Computer Graphics Forum, 27: 1079-1086. https://doi.org/10.1111/j.1467-8659.2008.01245.x
+ * @see https://en.wikipedia.org/wiki/Single-scattering_albedo
  */
 template <class T>
-T absorption_mie(T scattering)
+T absorption(T scattering, T albedo)
 {
-	return scattering / T(9);
+	return scattering * (T(1) / albedo - T(1));
 }
 
 /**
- * Calculates attenuation due to extinction (absorption + out-scattering).
+ * Calculates an extinction coefficient.
  *
- * @param ec Extinction coefficient (absorption coefficient + scattering coefficient).
- * @param s Scale factor.
- * @return Attenuation factor.
- */
-template <class T>
-T extinction(T ec, T s)
-{
-	return std::exp(-(ec * s));
-}
-
-/**
- * Calculates the single-scattering albedo (SSA) given a scattering coefficient and an extinction coefficient.
+ * @param scattering Scattering coefficient.
+ * @param albedo Single-scattering albedo.
+ *
+ * @return Extinction coefficient.
  *
- * @param s Scattering coefficient.
- * @param e Extinction coefficient.
- * @return Single-scattering albedo.
+ * @see https://en.wikipedia.org/wiki/Single-scattering_albedo
  */
 template <class T>
-T albedo(T s, T e)
+T extinction(T scattering, T albedo)
 {
-	return s / e;
+	return scattering / albedo;
 }
 
 /**
diff --git a/src/physics/gas/ozone.hpp b/src/physics/gas/ozone.hpp
index c8d17f4..f3a2459 100644
--- a/src/physics/gas/ozone.hpp
+++ b/src/physics/gas/ozone.hpp
@@ -20,51 +20,559 @@
 #ifndef ANTKEEPER_PHYSICS_GAS_OZONE_HPP
 #define ANTKEEPER_PHYSICS_GAS_OZONE_HPP
 
+#include "math/interpolation.hpp"
+
 namespace physics {
 namespace gas {
-	
+
 /// Ozone-related constants and functions.
 namespace ozone {
 
 /**
- * Cross section of ozone at wavelength 680nm and temperature 293K, in m-2/molecule.
- *
- * @see https://www.iup.uni-bremen.de/gruppen/molspec/databases/referencespectra/o3spectra2011/index.html
- */
-template <class T>
-constexpr T cross_section_680nm_293k = T(1.36820899679147e-25);
-
-/**
- * Cross section of ozone at wavelength 550nm and temperature 293K, in m-2/molecule.
+ * Cross sections of ozone from wavelength 280nm to 780nm, in increments of 1nm, at a temperature of 293K, in m-2/molecule.
  *
  * @see https://www.iup.uni-bremen.de/gruppen/molspec/databases/referencespectra/o3spectra2011/index.html
  */
-template <class T>
-constexpr T cross_section_550nm_293k = T(3.31405330400124e-25);
+constexpr double cross_sections_280nm_780nm_293k[781 - 280] =
+{
+	4.08020162329358e-22,
+	3.71300005805609e-22,
+	3.33329613050011e-22,
+	3.08830771169272e-22,
+	2.7952452827933e-22,
+	2.54507006594824e-22,
+	2.31059906830132e-22,
+	2.06098663896454e-22,
+	1.77641837487505e-22,
+	1.60649162987982e-22,
+	1.43573256268989e-22,
+	1.28028940869356e-22,
+	1.11222715354512e-22,
+	1.00812840644175e-22,
+	8.62757275283743e-23,
+	7.74492225274189e-23,
+	6.74998572150431e-23,
+	5.92584261096169e-23,
+	5.09573191001311e-23,
+	4.58263122360992e-23,
+	3.90608822487099e-23,
+	3.4712884682176e-23,
+	3.05004947986011e-23,
+	2.61948471583207e-23,
+	2.35950204633063e-23,
+	2.01696279467763e-23,
+	1.77716671613151e-23,
+	1.57884103804348e-23,
+	1.35332466460157e-23,
+	1.23148992566335e-23,
+	1.00674476685553e-23,
+	9.14966132477946e-24,
+	8.0373280016411e-24,
+	6.82346538467894e-24,
+	6.42747947320284e-24,
+	5.12300259416845e-24,
+	4.77060350699273e-24,
+	4.0011431423959e-24,
+	3.72184265884461e-24,
+	2.71828160861414e-24,
+	3.14980819763314e-24,
+	2.01180905161078e-24,
+	2.23741500460792e-24,
+	2.04586872483057e-24,
+	1.14774957440809e-24,
+	1.7873766001915e-24,
+	1.14326155695545e-24,
+	7.61393733215954e-25,
+	1.30841854222986e-24,
+	6.50013011248201e-25,
+	4.47253301018733e-25,
+	7.22254319278391e-25,
+	4.42923420258652e-25,
+	3.22168537281097e-25,
+	5.60700006481047e-25,
+	2.51991724359347e-25,
+	1.64975530971913e-25,
+	2.69863618345909e-25,
+	2.2962216595934e-25,
+	1.2768604186372e-25,
+	1.6428080690911e-25,
+	1.15252741016694e-25,
+	6.88123014503947e-26,
+	9.01478526179051e-26,
+	1.38671804168295e-25,
+	7.28648586727799e-26,
+	6.3468766437083e-26,
+	4.1705109317344e-26,
+	3.22298756116943e-26,
+	2.52476541455397e-26,
+	2.96595291121762e-26,
+	2.10981495904366e-26,
+	3.0696184107227e-26,
+	2.38627682184272e-26,
+	1.35645160962203e-26,
+	1.14737436955784e-26,
+	1.00293019429341e-26,
+	1.42699666872085e-26,
+	1.03892014298915e-26,
+	7.46029500400895e-27,
+	7.86729405487508e-27,
+	6.49493671023213e-27,
+	3.90586420068501e-27,
+	3.3969327080211e-27,
+	2.69156849765275e-27,
+	4.89998022220222e-27,
+	4.18363151355665e-27,
+	2.41505691668684e-27,
+	1.52926807176976e-27,
+	1.74010836686791e-27,
+	1.43997701486263e-27,
+	1.61611242687813e-27,
+	1.09444991940235e-27,
+	8.68337506495441e-28,
+	1.28660044051837e-27,
+	1.07534571825705e-27,
+	7.59223090396554e-28,
+	6.75850905941831e-28,
+	6.05594086115429e-28,
+	7.40387066310015e-28,
+	6.04999720618854e-28,
+	5.17923583652293e-28,
+	4.0858846895433e-28,
+	6.47448369216067e-28,
+	5.29992493931534e-28,
+	5.63128808710364e-28,
+	4.17695119955099e-28,
+	5.15330384762416e-28,
+	5.25185850011986e-28,
+	6.17618171380047e-28,
+	6.37119303086457e-28,
+	6.8496876006444e-28,
+	5.9229625229341e-28,
+	5.99998795876176e-28,
+	7.06136396740756e-28,
+	9.93040926727483e-28,
+	1.03668944015898e-27,
+	9.60158757124091e-28,
+	9.00247643555494e-28,
+	1.01801330597124e-27,
+	1.0893103182854e-27,
+	1.30096252080506e-27,
+	1.42018105948242e-27,
+	1.53477820553937e-27,
+	1.60452920542724e-27,
+	1.45613806382105e-27,
+	1.63671034304492e-27,
+	1.88680147053984e-27,
+	2.26659932677371e-27,
+	2.57113808884621e-27,
+	2.73584069155576e-27,
+	2.84524835939992e-27,
+	2.65294565386989e-27,
+	2.52007318042733e-27,
+	2.56347774832346e-27,
+	2.73272493525925e-27,
+	3.21972108535573e-27,
+	3.64879308272645e-27,
+	3.86875166703077e-27,
+	3.77657059174972e-27,
+	3.67917967079418e-27,
+	3.84603321414093e-27,
+	4.47110813921024e-27,
+	5.19879478264214e-27,
+	6.13707395634462e-27,
+	6.88229484256763e-27,
+	7.11935416561536e-27,
+	7.15812015493665e-27,
+	6.67142397990209e-27,
+	6.35218112458219e-27,
+	6.34510826220203e-27,
+	6.90321809802859e-27,
+	7.82871587803871e-27,
+	8.52938477234511e-27,
+	8.74335964163491e-27,
+	8.6718822639457e-27,
+	8.59495406258221e-27,
+	8.90719500501358e-27,
+	9.90176156668504e-27,
+	1.14857395824068e-26,
+	1.36017282525383e-26,
+	1.53232356175871e-26,
+	1.75024000506424e-26,
+	1.8179765858316e-26,
+	1.80631911188605e-26,
+	1.70102948254892e-26,
+	1.56536231218255e-26,
+	1.51141962474665e-26,
+	1.57847025841346e-26,
+	1.72161452840856e-26,
+	1.90909798194127e-26,
+	1.96774621921165e-26,
+	1.99812678813396e-26,
+	1.96900102296014e-26,
+	1.95126617395258e-26,
+	2.06408915381352e-26,
+	2.28866836725858e-26,
+	2.57662977048169e-26,
+	2.96637551360212e-26,
+	3.34197426701549e-26,
+	3.73735792971477e-26,
+	4.03453827718193e-26,
+	4.1291323324037e-26,
+	4.07643587970195e-26,
+	3.84067732691303e-26,
+	3.6271000065179e-26,
+	3.50502931147362e-26,
+	3.48851626868318e-26,
+	3.73778737646723e-26,
+	3.97091132121154e-26,
+	4.21773978585713e-26,
+	4.21620738550362e-26,
+	4.22087900183437e-26,
+	4.27973060694892e-26,
+	4.36010682588909e-26,
+	4.60584575680881e-26,
+	4.91428506793045e-26,
+	5.4918846417406e-26,
+	6.10573296817762e-26,
+	6.83025566941932e-26,
+	7.51469261186479e-26,
+	8.08197962688924e-26,
+	8.44082645474868e-26,
+	8.4843465766735e-26,
+	8.4126775729461e-26,
+	8.14411830190209e-26,
+	7.83636723569755e-26,
+	7.60974711104334e-26,
+	7.57877917471603e-26,
+	7.77887132347866e-26,
+	8.07522339055262e-26,
+	8.32310316258138e-26,
+	8.59015818152486e-26,
+	8.67834106505007e-26,
+	8.72244226406716e-26,
+	8.84734167611993e-26,
+	9.0711580597939e-26,
+	9.51778147590566e-26,
+	1.01490385328969e-25,
+	1.09448341447976e-25,
+	1.18257044868557e-25,
+	1.28105938778111e-25,
+	1.37732704252934e-25,
+	1.47491161151436e-25,
+	1.55090701035304e-25,
+	1.60575752538508e-25,
+	1.62886093543744e-25,
+	1.62802718439262e-25,
+	1.60288510229631e-25,
+	1.57216917046401e-25,
+	1.54475957021763e-25,
+	1.534341089264e-25,
+	1.5409967492982e-25,
+	1.56495784865383e-25,
+	1.60012763627009e-25,
+	1.63952588489707e-25,
+	1.67232066912218e-25,
+	1.70167894480834e-25,
+	1.7335194060265e-25,
+	1.7602731663686e-25,
+	1.79953556347172e-25,
+	1.84084607422109e-25,
+	1.89999243847493e-25,
+	1.97490644310087e-25,
+	2.05279021301286e-25,
+	2.14427839223598e-25,
+	2.24098369182092e-25,
+	2.34362490982003e-25,
+	2.44138366650567e-25,
+	2.53826212075759e-25,
+	2.62577562731292e-25,
+	2.70621837640467e-25,
+	2.76622780914432e-25,
+	2.80661519633223e-25,
+	2.82737499370866e-25,
+	2.82968166962191e-25,
+	2.82659484597549e-25,
+	2.81717325255129e-25,
+	2.82197485088881e-25,
+	2.84600833360439e-25,
+	2.88048754046166e-25,
+	2.92686210579676e-25,
+	2.98551267943544e-25,
+	3.04464034622896e-25,
+	3.09724266051229e-25,
+	3.14551726028915e-25,
+	3.18670379817661e-25,
+	3.22330249380314e-25,
+	3.25463784914917e-25,
+	3.2854797958699e-25,
+	3.31405330400124e-25,
+	3.34398013095565e-25,
+	3.38005272562664e-25,
+	3.41218557614901e-25,
+	3.45555599852459e-25,
+	3.5043277615423e-25,
+	3.55984911371566e-25,
+	3.6264721972979e-25,
+	3.70647322984569e-25,
+	3.79188179306458e-25,
+	3.88981479760571e-25,
+	3.98973596432023e-25,
+	4.08527421216539e-25,
+	4.17464990288797e-25,
+	4.25462181228461e-25,
+	4.32317712812093e-25,
+	4.38401384845366e-25,
+	4.44216978945654e-25,
+	4.49508440820886e-25,
+	4.5516564927479e-25,
+	4.60931329475278e-25,
+	4.66173637960526e-25,
+	4.70665064920011e-25,
+	4.74051362440113e-25,
+	4.75812011867871e-25,
+	4.7570031038151e-25,
+	4.73747927545327e-25,
+	4.71027119364443e-25,
+	4.66860282624977e-25,
+	4.6288533265784e-25,
+	4.57997871538082e-25,
+	4.53548773884213e-25,
+	4.49607272653461e-25,
+	4.4568818824566e-25,
+	4.42881930827398e-25,
+	4.40157474149992e-25,
+	4.38677564524713e-25,
+	4.37182142194489e-25,
+	4.37104565551326e-25,
+	4.37919711899623e-25,
+	4.39683352146027e-25,
+	4.42484514000691e-25,
+	4.47212673326828e-25,
+	4.53157968174371e-25,
+	4.60815452736499e-25,
+	4.69376508835705e-25,
+	4.78054316070552e-25,
+	4.87030272266768e-25,
+	4.95220907227656e-25,
+	5.02898332230558e-25,
+	5.09027352924287e-25,
+	5.13210890748271e-25,
+	5.15374454317407e-25,
+	5.15075653533686e-25,
+	5.13171378996911e-25,
+	5.09303097809138e-25,
+	5.03697164998998e-25,
+	4.97218676726656e-25,
+	4.89461758141549e-25,
+	4.80937594526597e-25,
+	4.72371968798565e-25,
+	4.63859449875443e-25,
+	4.55132403268817e-25,
+	4.46787830232533e-25,
+	4.38816245012175e-25,
+	4.31406743009557e-25,
+	4.24483372714822e-25,
+	4.17474755876221e-25,
+	4.11305082072427e-25,
+	4.05554427636835e-25,
+	4.0030863998631e-25,
+	3.95393149188544e-25,
+	3.90576318741963e-25,
+	3.86085282288514e-25,
+	3.81425489414328e-25,
+	3.76584427585746e-25,
+	3.71837025816073e-25,
+	3.66911165810168e-25,
+	3.61941817240908e-25,
+	3.56533744970388e-25,
+	3.51159798289664e-25,
+	3.46140984744989e-25,
+	3.41173597486151e-25,
+	3.36073006491665e-25,
+	3.30753716054224e-25,
+	3.25252799457143e-25,
+	3.19530424634748e-25,
+	3.13651908668849e-25,
+	3.07841534489121e-25,
+	3.01880275032991e-25,
+	2.95965464815758e-25,
+	2.90194743931008e-25,
+	2.84696394478385e-25,
+	2.79488978476361e-25,
+	2.74800171431914e-25,
+	2.70053589638645e-25,
+	2.65588371839819e-25,
+	2.6091795190684e-25,
+	2.56708069403319e-25,
+	2.52375403058723e-25,
+	2.48382556862202e-25,
+	2.44458617524673e-25,
+	2.40594587498642e-25,
+	2.36627105488787e-25,
+	2.32528309566254e-25,
+	2.28436716651676e-25,
+	2.24424313328781e-25,
+	2.20515093141858e-25,
+	2.16647251017334e-25,
+	2.12914125517962e-25,
+	2.09073684368918e-25,
+	2.05335637747553e-25,
+	2.01912040845123e-25,
+	1.98301051757051e-25,
+	1.94465669006103e-25,
+	1.9057000954812e-25,
+	1.86388414183128e-25,
+	1.82241187234978e-25,
+	1.78160123951627e-25,
+	1.74179167768809e-25,
+	1.70080577609513e-25,
+	1.6615348247565e-25,
+	1.62305192083274e-25,
+	1.58738563014106e-25,
+	1.55171430112041e-25,
+	1.51949383874537e-25,
+	1.48607225456117e-25,
+	1.45360366466218e-25,
+	1.42252183610792e-25,
+	1.39462753606039e-25,
+	1.36820899679147e-25,
+	1.34377836247288e-25,
+	1.3233906891166e-25,
+	1.30492894377563e-25,
+	1.28681831161393e-25,
+	1.2663174075999e-25,
+	1.2420229295933e-25,
+	1.21412305909061e-25,
+	1.18502869999655e-25,
+	1.15382448376731e-25,
+	1.12540171803974e-25,
+	1.09558453899584e-25,
+	1.06672010609186e-25,
+	1.03896476876362e-25,
+	1.01172047316729e-25,
+	9.84602582159865e-26,
+	9.58059205575365e-26,
+	9.33099533469407e-26,
+	9.08936155224937e-26,
+	8.8491955813636e-26,
+	8.64349280941688e-26,
+	8.44550335501804e-26,
+	8.25653238461191e-26,
+	8.08238949115334e-26,
+	7.94009406468206e-26,
+	7.80797137891831e-26,
+	7.67364989571709e-26,
+	7.54965718671858e-26,
+	7.43779079827743e-26,
+	7.31964666857758e-26,
+	7.2232944312979e-26,
+	7.12771524608971e-26,
+	7.0510155861502e-26,
+	6.99911412369698e-26,
+	6.96383131319201e-26,
+	6.92914966812787e-26,
+	6.85928437919081e-26,
+	6.74428511228458e-26,
+	6.59224480420957e-26,
+	6.38840433633278e-26,
+	6.15855599572905e-26,
+	5.95772659493798e-26,
+	5.76494205198861e-26,
+	5.60397496087846e-26,
+	5.46017309852595e-26,
+	5.32285644696103e-26,
+	5.22365816180628e-26,
+	5.07899578121548e-26,
+	4.93592723237266e-26,
+	4.79667132541204e-26,
+	4.68132550170927e-26,
+	4.56404612656666e-26,
+	4.48276475068268e-26,
+	4.40173864033052e-26,
+	4.35417448301629e-26,
+	4.30681941574933e-26,
+	4.28871372503407e-26,
+	4.24230311436515e-26,
+	4.22301315090103e-26,
+	4.19685390394596e-26,
+	4.18917236558952e-26,
+	4.17687017488352e-26,
+	4.22019512128238e-26,
+	4.2557462015274e-26,
+	4.31172890769302e-26,
+	4.36741466155527e-26,
+	4.41740590419353e-26,
+	4.44874945269751e-26,
+	4.42029497925774e-26,
+	4.34317624813875e-26,
+	4.21853858585038e-26,
+	4.01907542789597e-26,
+	3.80281792816081e-26,
+	3.60902775684479e-26,
+	3.41921953398109e-26,
+	3.25291539681547e-26,
+	3.10743399295997e-26,
+	2.99157340432027e-26,
+	2.89660087626517e-26,
+	2.82185391364051e-26,
+	2.76872520775773e-26,
+	2.72983807771732e-26,
+	2.65963957090044e-26,
+	2.62737083974039e-26,
+	2.56625477748644e-26,
+	2.53980768456023e-26,
+	2.49759827680602e-26,
+	2.47017705195044e-26,
+	2.47697548770545e-26,
+	2.46973660414457e-26,
+	2.47714280793403e-26,
+	2.50190654239581e-26,
+	2.56196471101148e-26,
+	2.64525680175745e-26,
+	2.72275203692747e-26,
+	2.826728922873e-26,
+	2.94507453632992e-26,
+	3.05247618840877e-26,
+	3.1005889405393e-26,
+	3.1508637563266e-26,
+	3.13148927506362e-26
+};
 
 /**
- * Cross section of ozone at wavelength 550nm and temperature 293K, in m-2/molecule.
+ * Returns the cross section of ozone at temperature 293k and a given wavelength in the visible spectrum.
  *
- * @see https://www.iup.uni-bremen.de/gruppen/molspec/databases/referencespectra/o3spectra2011/index.html
+ * @param wavelength Wavelength, in nanometers, on `[280, 780)`.
+ * @return Ozone cross section at temperature 293k and the given wavelength, in m-2/molecule.
  */
 template <class T>
-constexpr T cross_section_440nm_293k = T(1.36017282525383e-26);
+constexpr T cross_section_293k(T wavelength)
+{
+	int i = static_cast<int>(wavelength);
+	int j = static_cast<int>(wavelength + T(1));
+	
+	if (i < 280 || j > 780)
+		return T(0);
+	
+	const T x = static_cast<T>(cross_sections_280nm_780nm_293k[i - 280]);
+	const T y = static_cast<T>(cross_sections_280nm_780nm_293k[j - 280]);
+	
+	return math::lerp<T>(x, y, wavelength - static_cast<T>(i));
+}
 
 /**
  * Calculates an ozone absorption coefficient (wavelength-dependent).
  *
  * @param cross_section Ozone cross section of a wavelength, in m-2/molecule.
- * @param n Molecular number density of standard air, in molecules/m-3.
- * @param c Concentration of ozone in the atmosphere, unitless.
+ * @param density Molecular number density of ozone, in mol/m-3.
  *
  * @return Ozone absorption coefficient.
  *
  * @see https://skyrenderer.blogspot.com/2012/10/ozone-absorption.html
  */
 template <class T>
-T absorption(T cross_section, T n, T c)
+T absorption(T cross_section, T density)
 {
-	return cross_section * n * c;
+	return cross_section * density;
 }
 
 } // namespace ozone
diff --git a/src/physics/light/exposure.hpp b/src/physics/light/exposure.hpp
index f74db89..fa93e87 100644
--- a/src/physics/light/exposure.hpp
+++ b/src/physics/light/exposure.hpp
@@ -93,16 +93,16 @@ T to_luminance(T ev, T s, T k)
 }
 
 /**
- * Exposure value to luminance.
+ * Exposure value to illuminance.
  *
  * @param ev Exposure value.
  * @param s ISO arithmetic speed. A value of `100` corresponds to ISO 100.
- * @param k Incident-light meter calibration constant. A common value is `250`.
+ * @param c Incident-light meter calibration constant. A common value is `250`.
  *
  * @return Illuminance, in lux.
  */
 template <class T>
-T to_luminance(T ev, T s, T c)
+T to_illuminance(T ev, T s, T c)
 {
 	return  (c * std::exp2(ev)) / s;
 }
diff --git a/src/physics/time/jd.hpp b/src/physics/time/jd.hpp
index 3b1a3b8..258877b 100644
--- a/src/physics/time/jd.hpp
+++ b/src/physics/time/jd.hpp
@@ -26,6 +26,14 @@ namespace time {
 /// Julian day (JD).
 namespace jd {
 
+/// Number of Julian days per Julian century.
+template <class T>
+constexpr T days_per_century = T{36525};
+
+/// Number of Julian centuries per Julian day.
+template <class T>
+constexpr T centuries_per_day = T{1} / T{36525};
+
 /**
  * Converts JD to UT1.
  *
diff --git a/src/render/context.hpp b/src/render/context.hpp
index dcc6c7e..15311bc 100644
--- a/src/render/context.hpp
+++ b/src/render/context.hpp
@@ -53,9 +53,18 @@ struct context
 	/// Camera culling volume.
 	const geom::bounding_volume<float>* camera_culling_volume;
 	
-	/// Near clipping plane of the camera
+	/// Near clipping plane of the camera.
 	geom::plane<float> clip_near;
 	
+	/// Camera view matrix.
+	float4x4 view;
+	
+	/// Camera projection matrix.
+	float4x4 projection;
+	
+	/// Camera view projection matrix.
+	float4x4 view_projection;
+	
 	/// Camera exposure normalization factor.
 	float exposure;
 	
diff --git a/src/render/passes/ground-pass.cpp b/src/render/passes/ground-pass.cpp
index 1a8b5c8..05c6f57 100644
--- a/src/render/passes/ground-pass.cpp
+++ b/src/render/passes/ground-pass.cpp
@@ -86,10 +86,10 @@ void ground_pass::render(const render::context& ctx, render::queue& queue) const
 	float clip_far = camera.get_clip_far_tween().interpolate(ctx.alpha);
 	float3 model_scale = float3{1.0f, 1.0f, 1.0f} * (clip_near + clip_far) * 0.5f;
 	float4x4 model = math::scale(math::matrix4<float>::identity, model_scale);
-	float4x4 view = math::resize<4, 4>(math::resize<3, 3>(camera.get_view_tween().interpolate(ctx.alpha)));
+	float4x4 view = math::resize<4, 4>(math::resize<3, 3>(ctx.view));
 	float4x4 model_view = view * model;
-	float4x4 projection = camera.get_projection_tween().interpolate(ctx.alpha);
-	float4x4 view_projection = camera.get_view_projection_tween().interpolate(ctx.alpha);
+	const float4x4& projection = ctx.projection;
+	const float4x4& view_projection = ctx.view_projection;
 	float4x4 model_view_projection = projection * model_view;
 	
 	
diff --git a/src/render/passes/material-pass.cpp b/src/render/passes/material-pass.cpp
index 4448a55..1b5f18b 100644
--- a/src/render/passes/material-pass.cpp
+++ b/src/render/passes/material-pass.cpp
@@ -120,9 +120,9 @@ void material_pass::render(const render::context& ctx, render::queue& queue) con
 	float2 resolution = {static_cast<float>(std::get<0>(viewport)), static_cast<float>(std::get<1>(viewport))};
 	
 	const float3& camera_position = ctx.camera_transform.translation;
-	float4x4 view = ctx.camera->get_view_tween().interpolate(ctx.alpha);
-	float4x4 projection = ctx.camera->get_projection_tween().interpolate(ctx.alpha);
-	float4x4 view_projection = projection * view;
+	const float4x4& view = ctx.view;
+	const float4x4& projection = ctx.projection;
+	const float4x4& view_projection = ctx.view_projection;
 	float4x4 model_view_projection;
 	float4x4 model;
 	float4x4 model_view;
diff --git a/src/render/passes/sky-pass.cpp b/src/render/passes/sky-pass.cpp
index f040914..adfca21 100644
--- a/src/render/passes/sky-pass.cpp
+++ b/src/render/passes/sky-pass.cpp
@@ -46,6 +46,7 @@
 #include <cmath>
 #include <stdexcept>
 #include <glad/glad.h>
+#include <iostream>
 
 namespace render {
 
@@ -64,10 +65,6 @@ sky_pass::sky_pass(gl::rasterizer* rasterizer, const gl::framebuffer* framebuffe
 	stars_model_vao(nullptr),
 	star_material(nullptr),
 	star_shader_program(nullptr),
-	clouds_model(nullptr),
-	clouds_model_vao(nullptr),
-	cloud_material(nullptr),
-	cloud_shader_program(nullptr),
 	observer_position_tween({0, 0, 0}, math::lerp<float3, float>),
 	sun_position_tween(float3{1.0f, 0.0f, 0.0f}, math::lerp<float3, float>),
 	sun_luminance_tween(float3{0.0f, 0.0f, 0.0f}, math::lerp<float3, float>),
@@ -81,6 +78,8 @@ sky_pass::sky_pass(gl::rasterizer* rasterizer, const gl::framebuffer* framebuffe
 	moon_sunlight_illuminance_tween(float3{0, 0, 0}, math::lerp<float3, float>),
 	moon_planetlight_direction_tween(float3{0, 0, 0}, math::lerp<float3, float>),
 	moon_planetlight_illuminance_tween(float3{0, 0, 0}, math::lerp<float3, float>),
+	moon_illuminance_tween(float3{0.0f, 0.0f, 0.0f}, math::lerp<float3, float>),
+	render_transmittance_lut(false),
 	magnification(1.0f)
 {
 	// Build quad VBO and VAO
@@ -107,15 +106,13 @@ sky_pass::sky_pass(gl::rasterizer* rasterizer, const gl::framebuffer* framebuffe
 	quad_vao->bind(render::vertex_attribute::position, quad_position_attribute);
 	
 	// Create transmittance LUT texture and framebuffer (32F color, no depth)
-	int transmittance_width = 256;
-	int transmittance_height = 64;
-	transmittance_inverse_lut_resolution = {1.0f / static_cast<float>(transmittance_width), 1.0f / static_cast<float>(transmittance_height)};
-	transmittance_texture = new gl::texture_2d(transmittance_width, transmittance_height, gl::pixel_type::float_32, gl::pixel_format::rgb);
-	transmittance_texture->set_wrapping(gl::texture_wrapping::extend, gl::texture_wrapping::extend);
-	transmittance_texture->set_filters(gl::texture_min_filter::linear, gl::texture_mag_filter::linear);
-	transmittance_texture->set_max_anisotropy(0.0f);
-	transmittance_framebuffer = new gl::framebuffer(transmittance_width, transmittance_height);
-	transmittance_framebuffer->attach(gl::framebuffer_attachment_type::color, transmittance_texture);
+	transmittance_lut_texture = new gl::texture_2d(256, 64, gl::pixel_type::float_32, gl::pixel_format::rgb);
+	transmittance_lut_texture->set_wrapping(gl::texture_wrapping::extend, gl::texture_wrapping::extend);
+	transmittance_lut_texture->set_filters(gl::texture_min_filter::linear, gl::texture_mag_filter::linear);
+	transmittance_lut_texture->set_max_anisotropy(0.0f);
+	transmittance_lut_framebuffer = new gl::framebuffer({transmittance_lut_texture->get_width(), transmittance_lut_texture->get_height()});
+	transmittance_lut_framebuffer->attach(gl::framebuffer_attachment_type::color, transmittance_lut_texture);
+	transmittance_lut_resolution = {static_cast<float>(transmittance_lut_texture->get_width()), static_cast<float>(transmittance_lut_texture->get_height())};
 	
 	// Load transmittance LUT shader
 	transmittance_shader_program = resource_manager->load<gl::shader_program>("transmittance-lut.glsl");
@@ -124,13 +121,41 @@ sky_pass::sky_pass(gl::rasterizer* rasterizer, const gl::framebuffer* framebuffe
 	transmittance_mie_parameters_input = transmittance_shader_program->get_input("mie_parameters");
 	transmittance_ozone_distribution_input = transmittance_shader_program->get_input("ozone_distribution");
 	transmittance_ozone_absorption_input = transmittance_shader_program->get_input("ozone_absorption");
-	transmittance_inverse_lut_resolution_input = transmittance_shader_program->get_input("inverse_lut_resolution");
+	transmittance_resolution_input = transmittance_shader_program->get_input("resolution");
+	
+	// Create sky LUT texture and framebuffer (32F color, no depth)
+	int sky_lut_width = 200;
+	int sky_lut_height = 100;
+	sky_lut_resolution = {static_cast<float>(sky_lut_width), static_cast<float>(sky_lut_height)};
+	sky_lut_texture = new gl::texture_2d(sky_lut_width, sky_lut_height, gl::pixel_type::float_32, gl::pixel_format::rgb);
+	sky_lut_texture->set_wrapping(gl::texture_wrapping::extend, gl::texture_wrapping::extend);
+	sky_lut_texture->set_filters(gl::texture_min_filter::linear, gl::texture_mag_filter::linear);
+	sky_lut_texture->set_max_anisotropy(0.0f);
+	sky_lut_framebuffer = new gl::framebuffer(sky_lut_width, sky_lut_height);
+	sky_lut_framebuffer->attach(gl::framebuffer_attachment_type::color, sky_lut_texture);
+	
+	// Load sky LUT shader
+	sky_lut_shader_program = resource_manager->load<gl::shader_program>("sky-illuminance-lut.glsl");
+	sky_lut_light_direction_input = sky_lut_shader_program->get_input("light_direction");
+	sky_lut_light_illuminance_input = sky_lut_shader_program->get_input("light_illuminance");
+	sky_lut_atmosphere_radii_input = sky_lut_shader_program->get_input("atmosphere_radii");
+	sky_lut_observer_position_input = sky_lut_shader_program->get_input("observer_position");
+	sky_lut_rayleigh_parameters_input = sky_lut_shader_program->get_input("rayleigh_parameters");
+	sky_lut_mie_parameters_input = sky_lut_shader_program->get_input("mie_parameters");
+	sky_lut_ozone_distribution_input = sky_lut_shader_program->get_input("ozone_distribution");
+	sky_lut_ozone_absorption_input = sky_lut_shader_program->get_input("ozone_absorption");
+	sky_lut_airglow_illuminance_input = sky_lut_shader_program->get_input("airglow_illuminance");
+	sky_lut_resolution_input = sky_lut_shader_program->get_input("resolution");
+	sky_lut_transmittance_lut_input = sky_lut_shader_program->get_input("transmittance_lut");
+	sky_lut_transmittance_lut_resolution_input = sky_lut_shader_program->get_input("transmittance_lut_resolution");
 }
 
 sky_pass::~sky_pass()
 {
-	delete transmittance_framebuffer;
-	delete transmittance_texture;
+	delete sky_lut_framebuffer;
+	delete sky_lut_texture;
+	delete transmittance_lut_framebuffer;
+	delete transmittance_lut_texture;
 	delete quad_vao;
 	delete quad_vbo;
 }
@@ -143,33 +168,35 @@ void sky_pass::render(const render::context& ctx, render::queue& queue) const
 	glEnable(GL_CULL_FACE);
 	glCullFace(GL_BACK);
 	
-	// Render transmittance LUT
-	auto transmittance_viewport = transmittance_framebuffer->get_dimensions();
-	rasterizer->set_viewport(0, 0, std::get<0>(transmittance_viewport), std::get<1>(transmittance_viewport));
-	rasterizer->use_framebuffer(*transmittance_framebuffer);
-	rasterizer->use_program(*transmittance_shader_program);
-	transmittance_atmosphere_radii_input->upload(atmosphere_radii);
-	transmittance_rayleigh_parameters_input->upload(rayleigh_parameters);
-	transmittance_mie_parameters_input->upload(mie_parameters);
-	transmittance_ozone_distribution_input->upload(ozone_distribution);
-	transmittance_ozone_absorption_input->upload(ozone_absorption);
-	if (transmittance_inverse_lut_resolution_input)
-		transmittance_inverse_lut_resolution_input->upload(transmittance_inverse_lut_resolution);
-	rasterizer->draw_arrays(*quad_vao, gl::drawing_mode::triangles, 0, 6);
-	
-	rasterizer->use_framebuffer(*framebuffer);
-	auto viewport = framebuffer->get_dimensions();
-	rasterizer->set_viewport(0, 0, std::get<0>(viewport), std::get<1>(viewport));
-	float2 resolution = {static_cast<float>(std::get<0>(viewport)), static_cast<float>(std::get<1>(viewport))};
+	// Render transmittance LUT if transmittance parameters have been altered.
+	if (render_transmittance_lut)
+	{
+		// Render transmittance LUT
+		rasterizer->set_viewport(0, 0, transmittance_lut_texture->get_width(), transmittance_lut_texture->get_height());
+		rasterizer->use_framebuffer(*transmittance_lut_framebuffer);
+		rasterizer->use_program(*transmittance_shader_program);
+		transmittance_atmosphere_radii_input->upload(atmosphere_radii);
+		transmittance_rayleigh_parameters_input->upload(rayleigh_parameters);
+		transmittance_mie_parameters_input->upload(mie_parameters);
+		transmittance_ozone_distribution_input->upload(ozone_distribution);
+		transmittance_ozone_absorption_input->upload(ozone_absorption);
+		if (transmittance_resolution_input)
+			transmittance_resolution_input->upload(transmittance_lut_resolution);
+		rasterizer->draw_arrays(*quad_vao, gl::drawing_mode::triangles, 0, 6);
+		
+		// Don't render transmittance LUT next frame unless parameters have changed.
+		render_transmittance_lut = false;
+	}
 	
+	// Construct matrices
 	const scene::camera& camera = *ctx.camera;
 	float clip_near = camera.get_clip_near_tween().interpolate(ctx.alpha);
 	float clip_far = camera.get_clip_far_tween().interpolate(ctx.alpha);
 	float3 model_scale = float3{1.0f, 1.0f, 1.0f} * (clip_near + clip_far) * 0.5f;
 	float4x4 model = math::scale(math::matrix4<float>::identity, model_scale);
-	float4x4 view = math::resize<4, 4>(math::resize<3, 3>(camera.get_view_tween().interpolate(ctx.alpha)));
+	float4x4 view = math::resize<4, 4>(math::resize<3, 3>(ctx.view));
 	float4x4 model_view = view * model;
-	float4x4 projection = camera.get_projection_tween().interpolate(ctx.alpha);
+	const float4x4& projection = ctx.projection;
 	float4x4 view_projection = projection * view;
 	float4x4 model_view_projection = projection * model_view;
 	
@@ -191,6 +218,47 @@ void sky_pass::render(const render::context& ctx, render::queue& queue) const
 	float3 sun_luminance = sun_luminance_tween.interpolate(ctx.alpha) * ctx.exposure;
 	float3 sun_illuminance = sun_illuminance_tween.interpolate(ctx.alpha) * ctx.exposure;
 	
+	float3 moon_position = moon_position_tween.interpolate(ctx.alpha);
+	float3 moon_direction = math::normalize(moon_position);
+	float3 moon_illuminance = moon_illuminance_tween.interpolate(ctx.alpha) * ctx.exposure;
+	float moon_angular_radius = moon_angular_radius_tween.interpolate(ctx.alpha) * magnification;
+	
+	float sun_y = color::aces::ap1<float>.luminance(sun_transmitted_illuminance);
+	float moon_y = color::aces::ap1<float>.luminance(moon_transmitted_illuminance);
+	float3 dominant_light_direction = (sun_y > moon_y) ? sun_direction : moon_direction;
+	float3 dominant_light_illuminance = (sun_y > moon_y) ? sun_illuminance : moon_illuminance;
+	
+	if (moon_y > sun_y)
+		sun_luminance *= 0.0f;
+	
+	// Render sky illuminance LUT
+	auto sky_lut_viewport = sky_lut_framebuffer->get_dimensions();
+	rasterizer->set_viewport(0, 0, std::get<0>(sky_lut_viewport), std::get<1>(sky_lut_viewport));
+	rasterizer->use_framebuffer(*sky_lut_framebuffer);
+	rasterizer->use_program(*sky_lut_shader_program);
+	sky_lut_light_direction_input->upload(dominant_light_direction);
+	sky_lut_light_illuminance_input->upload(dominant_light_illuminance);
+	sky_lut_atmosphere_radii_input->upload(atmosphere_radii);
+	sky_lut_observer_position_input->upload(observer_position);
+	sky_lut_rayleigh_parameters_input->upload(rayleigh_parameters);
+	sky_lut_mie_parameters_input->upload(mie_parameters);
+	sky_lut_ozone_distribution_input->upload(ozone_distribution);
+	sky_lut_ozone_absorption_input->upload(ozone_absorption);
+	sky_lut_airglow_illuminance_input->upload(airglow_illuminance * ctx.exposure);
+	if (sky_lut_resolution_input)
+		sky_lut_resolution_input->upload(sky_lut_resolution);
+	sky_lut_transmittance_lut_input->upload(transmittance_lut_texture);
+	if (sky_lut_transmittance_lut_resolution_input)
+		sky_lut_transmittance_lut_resolution_input->upload(transmittance_lut_resolution);
+	rasterizer->draw_arrays(*quad_vao, gl::drawing_mode::triangles, 0, 6);
+	
+	
+	
+	rasterizer->use_framebuffer(*framebuffer);
+	auto viewport = framebuffer->get_dimensions();
+	rasterizer->set_viewport(0, 0, std::get<0>(viewport), std::get<1>(viewport));
+	float2 resolution = {static_cast<float>(std::get<0>(viewport)), static_cast<float>(std::get<1>(viewport))};
+	
 	// Draw atmosphere
 	if (sky_model)
 	{
@@ -203,61 +271,26 @@ void sky_pass::render(const render::context& ctx, render::queue& queue) const
 			mouse_input->upload(mouse_position);
 		if (resolution_input)
 			resolution_input->upload(resolution);
-		if (time_input)
-			time_input->upload(ctx.t);
-		if (exposure_input)
-			exposure_input->upload(ctx.exposure);
-		if (sun_direction_input)
-			sun_direction_input->upload(sun_direction);
+		if (light_direction_input)
+			light_direction_input->upload(dominant_light_direction);
 		if (sun_luminance_input)
 			sun_luminance_input->upload(sun_luminance);
-		if (sun_illuminance_input)
-			sun_illuminance_input->upload(sun_illuminance);
 		if (sun_angular_radius_input)
 			sun_angular_radius_input->upload(sun_angular_radius * magnification);
 		if (atmosphere_radii_input)
 			atmosphere_radii_input->upload(atmosphere_radii);
 		if (observer_position_input)
 			observer_position_input->upload(observer_position);
-		if (rayleigh_parameters_input)
-			rayleigh_parameters_input->upload(rayleigh_parameters);
-		if (mie_parameters_input)
-			mie_parameters_input->upload(mie_parameters);
-		if (ozone_distribution_input)
-			ozone_distribution_input->upload(ozone_distribution);
-		if (ozone_absorption_input)
-			ozone_absorption_input->upload(ozone_absorption);
-		if (transmittance_lut_input)
-			transmittance_lut_input->upload(transmittance_texture);
-		if (inverse_transmittance_lut_resolution_input)
-			inverse_transmittance_lut_resolution_input->upload(transmittance_inverse_lut_resolution);
+		if (sky_illuminance_lut_input)
+			sky_illuminance_lut_input->upload(sky_lut_texture);
+		if (sky_illuminance_lut_resolution_input)
+			sky_illuminance_lut_resolution_input->upload(sky_lut_resolution);
 		
 		sky_material->upload(ctx.alpha);
 
 		rasterizer->draw_arrays(*sky_model_vao, sky_model_drawing_mode, sky_model_start_index, sky_model_index_count);
 	}
 	
-	// Draw clouds
-	if (clouds_model)
-	{
-		rasterizer->use_program(*cloud_shader_program);
-		
-		if (cloud_model_view_projection_input)
-			cloud_model_view_projection_input->upload(model_view_projection);
-		if (cloud_sun_direction_input)
-			cloud_sun_direction_input->upload(sun_direction);
-		if (cloud_sun_illuminance_input)
-			cloud_sun_illuminance_input->upload(sun_illuminance);
-		if (cloud_camera_position_input)
-			cloud_camera_position_input->upload(ctx.camera_transform.translation);
-		if (cloud_camera_exposure_input)
-			cloud_camera_exposure_input->upload(ctx.exposure);
-		
-		cloud_material->upload(ctx.alpha);
-
-		rasterizer->draw_arrays(*clouds_model_vao, clouds_model_drawing_mode, clouds_model_start_index, clouds_model_index_count);
-	}
-	
 	glEnable(GL_BLEND);
 	glBlendFunc(GL_SRC_ALPHA, GL_ONE);
 	//glBlendFunc(GL_ONE, GL_ONE);
@@ -288,8 +321,6 @@ void sky_pass::render(const render::context& ctx, render::queue& queue) const
 	}
 	
 	// Draw moon model
-	float3 moon_position = moon_position_tween.interpolate(ctx.alpha);
-	float moon_angular_radius = moon_angular_radius_tween.interpolate(ctx.alpha) * magnification;
 	//if (moon_position.y >= -moon_angular_radius)
 	{
 		float moon_distance = (clip_near + clip_far) * 0.5f;		
@@ -352,20 +383,13 @@ void sky_pass::set_sky_model(const model* model)
 				model_view_projection_input = sky_shader_program->get_input("model_view_projection");
 				mouse_input = sky_shader_program->get_input("mouse");
 				resolution_input = sky_shader_program->get_input("resolution");
-				time_input = sky_shader_program->get_input("time");
-				exposure_input = sky_shader_program->get_input("camera.exposure");				
-				sun_direction_input = sky_shader_program->get_input("sun_direction");
+				light_direction_input = sky_shader_program->get_input("light_direction");
 				sun_luminance_input = sky_shader_program->get_input("sun_luminance");
-				sun_illuminance_input = sky_shader_program->get_input("sun_illuminance");
 				sun_angular_radius_input = sky_shader_program->get_input("sun_angular_radius");
 				atmosphere_radii_input = sky_shader_program->get_input("atmosphere_radii");
 				observer_position_input = sky_shader_program->get_input("observer_position");
-				rayleigh_parameters_input = sky_shader_program->get_input("rayleigh_parameters");
-				mie_parameters_input = sky_shader_program->get_input("mie_parameters");
-				ozone_distribution_input = sky_shader_program->get_input("ozone_distribution");
-				ozone_absorption_input = sky_shader_program->get_input("ozone_absorption");
-				transmittance_lut_input = sky_shader_program->get_input("transmittance_lut");
-				inverse_transmittance_lut_resolution_input = sky_shader_program->get_input("inverse_transmittance_lut_resolution");
+				sky_illuminance_lut_input = sky_shader_program->get_input("sky_illuminance_lut");
+				sky_illuminance_lut_resolution_input = sky_shader_program->get_input("sky_illuminance_lut_resolution");
 			}
 		}
 	}
@@ -451,43 +475,6 @@ void sky_pass::set_stars_model(const model* model)
 	}
 }
 
-void sky_pass::set_clouds_model(const model* model)
-{
-	clouds_model = model;
-	
-	if (clouds_model)
-	{
-		clouds_model_vao = model->get_vertex_array();
-
-		const std::vector<model_group*>& groups = *model->get_groups();
-		for (model_group* group: groups)
-		{
-			cloud_material = group->get_material();
-			clouds_model_drawing_mode = group->get_drawing_mode();
-			clouds_model_start_index = group->get_start_index();
-			clouds_model_index_count = group->get_index_count();
-		}
-		
-		if (cloud_material)
-		{
-			cloud_shader_program = cloud_material->get_shader_program();
-			
-			if (cloud_shader_program)
-			{
-				cloud_model_view_projection_input = cloud_shader_program->get_input("model_view_projection");
-				cloud_sun_direction_input = cloud_shader_program->get_input("sun_direction");
-				cloud_sun_illuminance_input = cloud_shader_program->get_input("sun_illuminance");
-				cloud_camera_position_input = cloud_shader_program->get_input("camera.position");
-				cloud_camera_exposure_input = cloud_shader_program->get_input("camera.exposure");
-			}
-		}
-	}
-	else
-	{
-		clouds_model = nullptr;
-	}
-}
-
 void sky_pass::update_tweens()
 {
 	observer_position_tween.update();
@@ -504,6 +491,7 @@ void sky_pass::update_tweens()
 	moon_sunlight_illuminance_tween.update();
 	moon_planetlight_direction_tween.update();
 	moon_planetlight_illuminance_tween.update();
+	moon_illuminance_tween.update();
 }
 
 void sky_pass::set_magnification(float magnification)
@@ -522,9 +510,10 @@ void sky_pass::set_sun_position(const float3& position)
 	sun_position_tween[1] = position;
 }
 
-void sky_pass::set_sun_illuminance(const float3& illuminance)
+void sky_pass::set_sun_illuminance(const float3& illuminance, const float3& transmitted_illuminance)
 {
 	sun_illuminance_tween[1] = illuminance;
+	sun_transmitted_illuminance = transmitted_illuminance;
 }
 
 void sky_pass::set_sun_luminance(const float3& luminance)
@@ -543,6 +532,9 @@ void sky_pass::set_planet_radius(float radius)
 	atmosphere_radii.y = atmosphere_radii.x + atmosphere_upper_limit;
 	atmosphere_radii.z = atmosphere_radii.y * atmosphere_radii.y;
 	observer_position_tween[1] = {0.0f, atmosphere_radii.x + observer_elevation, 0.0f};
+	
+	// Trigger transmittance LUT render
+	render_transmittance_lut = true;
 }
 
 void sky_pass::set_atmosphere_upper_limit(float limit)
@@ -550,6 +542,9 @@ void sky_pass::set_atmosphere_upper_limit(float limit)
 	atmosphere_upper_limit = limit;
 	atmosphere_radii.y = atmosphere_radii.x + atmosphere_upper_limit;
 	atmosphere_radii.z = atmosphere_radii.y * atmosphere_radii.y;
+	
+	// Trigger transmittance LUT render
+	render_transmittance_lut = true;
 }
 
 void sky_pass::set_observer_elevation(float elevation)
@@ -567,17 +562,23 @@ void sky_pass::set_rayleigh_parameters(float scale_height, const float3& scatter
 		scattering.y,
 		scattering.z
 	};
+	
+	// Trigger transmittance LUT render
+	render_transmittance_lut = true;
 }
 
-void sky_pass::set_mie_parameters(float scale_height, float scattering, float absorption, float anisotropy)
+void sky_pass::set_mie_parameters(float scale_height, float scattering, float extinction, float anisotropy)
 {
 	mie_parameters =
 	{
 		-1.0f / scale_height,
 		scattering,
-		absorption,
+		extinction,
 		anisotropy
 	};
+	
+	// Trigger transmittance LUT render
+	render_transmittance_lut = true;
 }
 
 void sky_pass::set_ozone_parameters(float lower_limit, float upper_limit, float mode, const float3& absorption)
@@ -589,6 +590,14 @@ void sky_pass::set_ozone_parameters(float lower_limit, float upper_limit, float
 		mode
 	};
 	ozone_absorption = absorption;
+	
+	// Trigger transmittance LUT render
+	render_transmittance_lut = true;
+}
+
+void sky_pass::set_airglow_illuminance(const float3& illuminance)
+{
+	airglow_illuminance = illuminance;
 }
 
 void sky_pass::set_moon_position(const float3& position)
@@ -626,6 +635,22 @@ void sky_pass::set_moon_planetlight_illuminance(const float3& illuminance)
 	moon_planetlight_illuminance_tween[1] = illuminance;
 }
 
+void sky_pass::set_moon_illuminance(const float3& illuminance, const float3& transmitted_illuminance)
+{
+	moon_illuminance_tween[1] = illuminance;
+	moon_transmitted_illuminance = transmitted_illuminance;
+}
+
+void sky_pass::set_transmittance_lut_resolution(std::uint16_t width, std::uint16_t height)
+{
+	transmittance_lut_texture->resize(width, height, nullptr);
+	transmittance_lut_framebuffer->resize({transmittance_lut_texture->get_width(), transmittance_lut_texture->get_height()});
+	transmittance_lut_resolution = {static_cast<float>(width), static_cast<float>(height)};
+	
+	// Trigger transmittance LUT render
+	render_transmittance_lut = true;
+}
+
 void sky_pass::handle_event(const mouse_moved_event& event)
 {
 	mouse_position = {static_cast<float>(event.x), static_cast<float>(event.y)};
diff --git a/src/render/passes/sky-pass.hpp b/src/render/passes/sky-pass.hpp
index 9216fb7..d561ac0 100644
--- a/src/render/passes/sky-pass.hpp
+++ b/src/render/passes/sky-pass.hpp
@@ -60,20 +60,20 @@ public:
 	void set_sky_model(const model* model);
 	void set_moon_model(const model* model);
 	void set_stars_model(const model* model);
-	void set_clouds_model(const model* model);
 	
 	void set_icrf_to_eus(const math::transformation::se3<float>& transformation);
 	
 	void set_sun_position(const float3& position);
 	void set_sun_luminance(const float3& luminance);
-	void set_sun_illuminance(const float3& illuminance);
+	void set_sun_illuminance(const float3& illuminance, const float3& transmitted_illuminance);
 	void set_sun_angular_radius(float radius);
 	void set_planet_radius(float radius);
 	void set_atmosphere_upper_limit(float limit);
 	void set_observer_elevation(float elevation);
 	void set_rayleigh_parameters(float scale_height, const float3& scattering);
-	void set_mie_parameters(float scale_height, float scattering, float absorption, float anisotropy);
+	void set_mie_parameters(float scale_height, float scattering, float extinction, float anisotropy);
 	void set_ozone_parameters(float lower_limit, float upper_limit, float mode, const float3& absorption);
+	void set_airglow_illuminance(const float3& illuminance);
 	
 	void set_moon_position(const float3& position);
 	void set_moon_rotation(const math::quaternion<float>& rotation);
@@ -82,42 +82,62 @@ public:
 	void set_moon_sunlight_illuminance(const float3& illuminance);
 	void set_moon_planetlight_direction(const float3& direction);
 	void set_moon_planetlight_illuminance(const float3& illuminance);
+	void set_moon_illuminance(const float3& illuminance, const float3& transmitted_illuminance);
+	
+	/**
+	 * Sets the resolution of transmittance LUT.
+	 *
+	 * @param width Transmittance LUT width, in pixels.
+	 * @param height Transmittance LUT height, in pixels.
+	 */
+	void set_transmittance_lut_resolution(std::uint16_t width, std::uint16_t height);
 
 private:
 	virtual void handle_event(const mouse_moved_event& event);
 	
 	gl::vertex_buffer* quad_vbo;
 	gl::vertex_array* quad_vao;
-	gl::texture_2d* transmittance_texture;
-	gl::framebuffer* transmittance_framebuffer;
-	float2 transmittance_inverse_lut_resolution;
 	
+	gl::texture_2d* transmittance_lut_texture;
+	gl::framebuffer* transmittance_lut_framebuffer;
+	float2 transmittance_lut_resolution;
 	gl::shader_program* transmittance_shader_program;
 	const gl::shader_input* transmittance_atmosphere_radii_input;
 	const gl::shader_input* transmittance_rayleigh_parameters_input;
 	const gl::shader_input* transmittance_mie_parameters_input;
 	const gl::shader_input* transmittance_ozone_distribution_input;
 	const gl::shader_input* transmittance_ozone_absorption_input;
-	const gl::shader_input* transmittance_inverse_lut_resolution_input;
+	const gl::shader_input* transmittance_resolution_input;
+	mutable bool render_transmittance_lut;
+	
+	gl::texture_2d* sky_lut_texture;
+	gl::framebuffer* sky_lut_framebuffer;
+	gl::shader_program* sky_lut_shader_program;
+	float2 sky_lut_resolution;
+	const gl::shader_input* sky_lut_light_direction_input;
+	const gl::shader_input* sky_lut_light_illuminance_input;
+	const gl::shader_input* sky_lut_atmosphere_radii_input;
+	const gl::shader_input* sky_lut_observer_position_input;
+	const gl::shader_input* sky_lut_rayleigh_parameters_input;
+	const gl::shader_input* sky_lut_mie_parameters_input;
+	const gl::shader_input* sky_lut_ozone_distribution_input;
+	const gl::shader_input* sky_lut_ozone_absorption_input;
+	const gl::shader_input* sky_lut_airglow_illuminance_input;
+	const gl::shader_input* sky_lut_resolution_input;
+	const gl::shader_input* sky_lut_transmittance_lut_input;
+	const gl::shader_input* sky_lut_transmittance_lut_resolution_input;
 
 	gl::shader_program* sky_shader_program;
 	const gl::shader_input* model_view_projection_input;
 	const gl::shader_input* mouse_input;
 	const gl::shader_input* resolution_input;
-	const gl::shader_input* time_input;
-	const gl::shader_input* exposure_input;
-	const gl::shader_input* sun_direction_input;
+	const gl::shader_input* light_direction_input;
 	const gl::shader_input* sun_luminance_input;
-	const gl::shader_input* sun_illuminance_input;
 	const gl::shader_input* sun_angular_radius_input;
 	const gl::shader_input* atmosphere_radii_input;
 	const gl::shader_input* observer_position_input;
-	const gl::shader_input* rayleigh_parameters_input;
-	const gl::shader_input* mie_parameters_input;
-	const gl::shader_input* ozone_distribution_input;
-	const gl::shader_input* ozone_absorption_input;
-	const gl::shader_input* transmittance_lut_input;
-	const gl::shader_input* inverse_transmittance_lut_resolution_input;
+	const gl::shader_input* sky_illuminance_lut_input;
+	const gl::shader_input* sky_illuminance_lut_resolution_input;
 	
 	gl::shader_program* moon_shader_program;
 	const gl::shader_input* moon_model_input;
@@ -155,13 +175,6 @@ private:
 	const gl::shader_input* star_exposure_input;
 	const gl::shader_input* star_distance_input;
 	
-	const model* clouds_model;
-	const material* cloud_material;
-	const gl::vertex_array* clouds_model_vao;
-	gl::drawing_mode clouds_model_drawing_mode;
-	std::size_t clouds_model_start_index;
-	std::size_t clouds_model_index_count;
-	
 	gl::shader_program* cloud_shader_program;
 	const gl::shader_input* cloud_model_view_projection_input;
 	const gl::shader_input* cloud_sun_direction_input;
@@ -169,14 +182,12 @@ private:
 	const gl::shader_input* cloud_camera_position_input;
 	const gl::shader_input* cloud_camera_exposure_input;
 
-	const gl::texture_2d* blue_noise_map;
-	const gl::texture_2d* sky_gradient;
-	const gl::texture_2d* sky_gradient2;
 	float2 mouse_position;
 	
 	tween<float3> sun_position_tween;
 	tween<float3> sun_luminance_tween;
 	tween<float3> sun_illuminance_tween;
+	float3 sun_transmitted_illuminance;
 	tween<float3> icrf_to_eus_translation;
 	tween<math::quaternion<float>> icrf_to_eus_rotation;
 	
@@ -187,6 +198,9 @@ private:
 	tween<float3> moon_sunlight_illuminance_tween;
 	tween<float3> moon_planetlight_direction_tween;
 	tween<float3> moon_planetlight_illuminance_tween;
+	tween<float3> moon_illuminance_tween;
+	float3 moon_transmitted_illuminance;
+
 	
 	float sun_angular_radius;
 	float atmosphere_upper_limit;
@@ -197,6 +211,7 @@ private:
 	float4 mie_parameters;
 	float3 ozone_distribution;
 	float3 ozone_absorption;
+	float3 airglow_illuminance;
 	
 	float magnification;
 };
diff --git a/src/render/renderer.cpp b/src/render/renderer.cpp
index 68016c7..f56fa8f 100644
--- a/src/render/renderer.cpp
+++ b/src/render/renderer.cpp
@@ -107,6 +107,9 @@ void renderer::render(float t, float dt, float alpha, const scene::collection& c
 		ctx.camera_forward = ctx.camera_transform.rotation * config::global_forward;
 		ctx.camera_up = ctx.camera_transform.rotation * config::global_up;
 		ctx.clip_near = camera->get_view_frustum().get_near(); ///< @TODO: tween this
+		ctx.view = camera->get_view_tween().interpolate(alpha);
+		ctx.projection = camera->get_projection_tween().interpolate(alpha);
+		ctx.view_projection = ctx.projection * ctx.view;
 		ctx.exposure = std::exp2(-camera->get_exposure_tween().interpolate(alpha));
 		
 		// Create render queue
diff --git a/src/resources/entity-archetype-loader.cpp b/src/resources/entity-archetype-loader.cpp
index e9aa406..d42e598 100644
--- a/src/resources/entity-archetype-loader.cpp
+++ b/src/resources/entity-archetype-loader.cpp
@@ -44,32 +44,37 @@ static bool load_component_atmosphere(entity::archetype& archetype, const json&
 		component.upper_limit = element["upper_limit"].get<double>();
 	if (element.contains("index_of_refraction"))
 		component.index_of_refraction = element["index_of_refraction"].get<double>();
-	if (element.contains("air_concentration"))
-		component.air_concentration = element["air_concentration"].get<double>();
 	
+	if (element.contains("rayleigh_concentration"))
+		component.rayleigh_concentration = element["rayleigh_concentration"].get<double>();
 	if (element.contains("rayleigh_scale_height"))
 		component.rayleigh_scale_height = element["rayleigh_scale_height"].get<double>();
-	if (element.contains("rayleigh_density"))
-		component.rayleigh_density = element["rayleigh_density"].get<double>();
 	
+	if (element.contains("mie_concentration"))
+		component.mie_concentration = element["mie_concentration"].get<double>();
 	if (element.contains("mie_scale_height"))
 		component.mie_scale_height = element["mie_scale_height"].get<double>();
-	if (element.contains("mie_density"))
-		component.mie_density = element["mie_density"].get<double>();
 	if (element.contains("mie_anisotropy"))
 		component.mie_anisotropy = element["mie_anisotropy"].get<double>();
+	if (element.contains("mie_albedo"))
+		component.mie_albedo = element["mie_albedo"].get<double>();
 	
+	if (element.contains("ozone_concentration"))
+		component.ozone_concentration = element["ozone_concentration"].get<double>();
 	if (element.contains("ozone_lower_limit"))
 		component.ozone_lower_limit = element["ozone_lower_limit"].get<double>();
 	if (element.contains("ozone_upper_limit"))
 		component.ozone_upper_limit = element["ozone_upper_limit"].get<double>();
 	if (element.contains("ozone_mode"))
 		component.ozone_mode = element["ozone_mode"].get<double>();
-	if (element.contains("ozone_concentration"))
-		component.ozone_concentration = element["ozone_concentration"].get<double>();
 	
-	if (element.contains("airglow"))
-		component.airglow = element["airglow"].get<double>();
+	if (element.contains("airglow_illuminance"))
+	{
+		const auto& airglow_illuminance = element["airglow_illuminance"];
+		component.airglow_illuminance.x = airglow_illuminance[0].get<double>();
+		component.airglow_illuminance.y = airglow_illuminance[1].get<double>();
+		component.airglow_illuminance.z = airglow_illuminance[2].get<double>();
+	}
 	
 	archetype.set<entity::component::atmosphere>(component);
 
diff --git a/src/resources/material-loader.cpp b/src/resources/material-loader.cpp
index a469037..82ff427 100644
--- a/src/resources/material-loader.cpp
+++ b/src/resources/material-loader.cpp
@@ -161,7 +161,7 @@ static bool load_vector_property(render::material* material, const std::string&
 			T value;
 			std::size_t j = 0;
 			for (const auto& value_element: vector_element)
-				value[j++] = value_element.get<typename T::scalar_type>();
+				value[j++] = value_element.get<typename T::element_type>();
 			
 			// Set property values
 			property->set_value(i++, value);
@@ -176,7 +176,7 @@ static bool load_vector_property(render::material* material, const std::string&
 		T value;
 		std::size_t i = 0;
 		for (const auto& value_element: json)
-			value[i++] = value_element.get<typename T::scalar_type>();
+			value[i++] = value_element.get<typename T::element_type>();
 		
 		// Set property values
 		property->set_value(value);