Color Volume
Color volume describes which colors can be reproduced across a range of light levels.
A red light and a white reflection can ask very different things of a screen. The reflection can use all three RGB components. A saturated red uses mostly the red component, so it can run out of available output well before the screen reaches its maximum white brightness.
Color volume describes colors together with their light levels. It extends the two-dimensional idea of color gamut into the brightness dimension: which colors can be reproduced near black, at middle levels, and near the brightest output?
Why the volume is not a straight stack of gamut triangles
Imagine an ideal additive RGB display using BT.709 primaries, with linear component values limited to zero through one and 100-nit white. Its neutral white is R = G = B = 1. In this example, luminance is:
Y = 100 × (0.2126R + 0.7152G + 0.0722B) nits
The coefficient sum is one. Setting all components to one produces 100 nits, but setting just one component to one gives different results:
| Linear RGB setting | Color | Calculated luminance |
|---|---|---|
| 1, 0, 0 | Red primary | 21.26 nits |
| 0, 1, 0 | Green primary | 71.52 nits |
| 0, 0, 1 | Blue primary | 7.22 nits |
| 1, 1, 0 | Red plus green | 92.78 nits |
| 1, 1, 1 | White | 100 nits |
An idealized calculation with fixed primaries and component limits. It is not a measurement of a television.
To make that pure red brighter while retaining the same chromaticity would require increasing its red component beyond one. The display has already reached the limit. Adding green and blue can raise luminance, but moves the color toward a less saturated mixture. This is the brightness/saturation tradeoff the word “volume” helps describe.
Near the top of this ideal volume, fewer saturated colors remain available. At the very highest white level, all components are already at their maxima. Real displays add other limits: their primaries and channel behavior vary, their black floor affects dark colors, and power or thermal control can change output during a bright scene.
Brightness and saturation in an orange flame
Consider an HDR picture containing a bright orange flame. Its encoded RGB values describe both its color and light level. The display may have sufficient reach toward orange at a moderate luminance but insufficient output to reproduce that particular orange at the requested high luminance.
A conversion can reduce the flame's brightness, reduce its saturation, or combine adjustments. Keeping only the brightness target and clipping individual color components can shift hue. Keeping only the hue and saturation target can require a larger luminance reduction. A useful mapping balances the relationships that make the flame look convincing alongside the rest of the scene.
This is why tone mapping and gamut mapping interact. Their jobs can be integrated into one color-volume mapping. Treating the output as a brightness ceiling plus an independent fixed triangle misses part of the problem.
Reference volume and display measurement conditions
A color-volume percentage needs a reference volume and a specified measurement method. Different coordinate systems weight color differences differently; a percentage relative to one reference cannot be compared directly with an unrelated percentage.
Test conditions also matter. A brief small white patch, a large bright area, and a sustained colored patch can produce different output limits on the same screen. Measurements across luminance and color give more information than one peak-white number. Near black, useful separation between dark colors matters even when the high-brightness result is impressive.
Mastering-display metadata records properties of the mastering reference. During Emby playback, the receiving screen can have a different volume. The renderer adapts the produced image to the output; retaining the original source allows that adaptation to be chosen again for another display.