KNOWLEDGE MAP · SPACES & STANDARDS
Same code values? Decode before mixing.
The web and graphics space without the “gamma 2.2” shortcut. Compare linear-light and encoded mixing, then examine profiles, ranges and video handoff.
Read the numbers. Notice the assumptions. Then apply the idea to your footage.
Mix light, or mix codes?
Y 0.50000Y 0.214040.73536 / 0.73536 / 0.73536Y_linear − Y_codes = 0.285959Both strips are calculated from the exact piecewise sRGB curve. Linear Y is relative, not monitor nits. Code-space mixing is a different operation, not automatically a broken file. This is not a visual calibration test.
TOPIC ILLUSTRATIONTHE ESSENTIAL IDEA
sRGB
sRGB combines Rec.709 primaries, D65 white and its own piecewise encoding. Numbers may be stored correctly yet mixed or interpreted incorrectly.
- Understand why midpoint code is not midpoint light.
- Distinguish image profiles from video tags.
- Prepare graphics for a known video pipeline.
Why the Internet needed a default RGB space
sRGB emerged as a shared interpretation for images crossing devices and applications. The historical Hewlett-Packard/Microsoft proposal sought a practical colour baseline for the internet and desktop systems, particularly when detailed device profiles were absent.
It does not make all monitors identical. It defines the image’s source space; applications and colour management must relate those coordinates to the actual display. “An sRGB file” and “a calibrated screen” are different statements.
What this means in practice
An embedded profile helps explain the numbers; it does not correct a misadjusted monitor white.
What sRGB actually defines
sRGB uses R (0.640, 0.330), G (0.300, 0.600), B (0.150, 0.060), D65 white and a piecewise transfer function. Coordinates belong to a defined colour representation, not merely three slider positions.
Code [0.5, 0.5, 0.5] is neutral relative to this white, but it is not half of linear light. Decoding yields approximately 0.214 per channel. Do not confuse this relative value with physical screen luminance in cd/m².
What this means in practice
Name the quantity first: code 0.5, linear RGB 0.5 and 50 cd/m² are three different statements.
Why sRGB and Rec.709 have the same gamut
Shared Rec.709 primaries imply the same linear RGB matrix when white normalization also matches. They do not remove differences in encoding and display assumptions.
A graphic and video may need no gamut-basis conversion, yet still need tone-interpretation handling. Establish whether you are preserving finished artwork’s appearance or performing another explicitly defined operation. Simply replacing one tag with another does not specify that intent.
What this means in practice
Check a neutral gradient first: no white-point change is needed here, but midtones may be interpreted differently.
The exact sRGB transfer function: where 12.92 and 2.4 come from
Decode C using C/12.92 when C ≤ 0.04045; otherwise use ((C + 0.055)/1.055)^2.4. The inverse uses a 0.0031308 breakpoint. These branches explain why sRGB is neither exactly gamma 2.2 nor a pure 2.4 power.
Light calculations operate after decoding. A half-and-half linear-light black/white mixture encodes to roughly 0.735, whereas averaging codes gives 0.5. Both are defined operations with different purposes; linear mixing does not promise perceptual uniformity.
What this means in practice
Switch the mixing domain at the same 50% position. The label and calculated values should change together.
Reference display: why sRGB is more than chromaticities
The ICC registry relates sRGB to a reference display context, including D65 and reference white at 80 cd/m². This is a descriptive model, not a statement that your browser window physically emits that luminance.
Device brightness, display profile, application colour management and surround remain separate factors. Correct coordinate conversion alone cannot guarantee the same subjective appearance in a bright room and a darkened studio.
What this means in practice
Do not treat 80 as the user’s measured luminance in a browser demonstration.
ICC profiles: what the file tells an application
A profile relates file values to a defined colour meaning. Assigning a profile describes existing numbers; converting to a profile recalculates coordinates under a chosen transform. These are different actions.
Do not assign Display P3 to a correctly tagged sRGB image just to intensify colours: that reinterprets the numbers. Missing profiles require investigating the source, not choosing the prettiest result. An application’s sRGB default is an assumption, not source evidence.
What this means in practice
Use two copies: assign a different profile to one and convert the other. Compare both appearance and numbers.
Why sRGB is still the default in CSS and browsers
Ordinary CSS HEX, rgb() and hsl() specify sRGB colours. #ff0000 is not the most saturated red the current display can produce. Another space needs an explicit notation such as color(display-p3 …).
CSS Color 4 separately describes spaces, interpolation and out-of-gamut handling. A checked draft specification is not a guarantee that a particular browser implements every feature. Test the required functions and the readability of fallbacks.
What this means in practice
A green button must retain text contrast when switching between sRGB and wider-gamut presentation.
Wide-gamut web does not replace sRGB
Wide-gamut web presentation requires a coherent chain: a correct colour definition, application support, colour management and a capable display. A P3 tag alone cannot extend hardware capabilities.
The swatches in this chapter’s labs deliberately use sRGB. That makes the preview path explicit. Evaluating actual wide-gamut output needs an appropriate display path; numerical conversions should still be checked separately from visual impressions.
What this means in practice
Identical clipped previews do not prove that the original wide-gamut coordinates are identical.
Why a Figma PNG and Rec.709 video can disagree in Resolve
Artwork, photographs and clips may arrive in different source representations. A finished sRGB PNG does not become camera log merely because it enters a DWG/Intermediate timeline. Transparent files also require separate alpha and edge checks.
Agree the source space, desired appearance preservation and working domain. If darkening appears only after compositing, inspect blending and alpha interpretation as well: the profile is not necessarily the source of the error.
What this means in practice
Compare a logo alone and over a neutral background. An error confined to translucent edges is a useful separate clue.
sRGB → Rec.709: when no matrix is needed but a transform still may be
With matching primaries and D65, no linear RGB basis change is necessary. Yet a finished display image and a pre-rendering signal have different meanings even when both are called RGB.
Converting sRGB artwork for video begins by decoding it, then performs the necessary reconciliation and output encoding. The exact process depends on intent: preserve a finished appearance, integrate an image into a scene, or prepare another master. Do not stack every available transform indiscriminately.
What this means in practice
Test neutral and saturated samples, not only white: white may match even when the tone function is wrong.
Practical workflow: web graphic → Resolve → video master
Before delivery, verify source provenance, profile assignment, working space and destination. For web imagery, check the profile and managed browser view; for video, also check levels, codec and colour tags.
Preserve embedded profiles where the workflow expects them. File size, resolution and profile presence alone do not prove colour correctness. Keep a control frame and a short test to detect changes during re-export or transfer to another application.
What this means in practice
One source image can have coordinated but numerically different web and video deliveries.
A MODEL YOU CAN CHECK
Code, light and profile are different.
Keep the meaning, the calculation and the limitation together.
C = 12.92L (L ≤ 0.0031308); otherwise 1.055L^(1/2.4) − 0.055First define what a blend is averaging. Linear-light mixing and sRGB-code averaging are different operations; creative interpolation may deliberately choose the latter.
| Concept | What it describes | What not to conclude |
|---|---|---|
| Linear blend | Average of decoded light contributions. | Does not guarantee perceptual uniformity. |
| Encoded blend | Average of sRGB numbers. | Not equivalent to average light. |
| ICC profile | Description of image interpretation. | Assignment alone does not recalculate pixels. |
| sRGB / Rec.709 | Shared primaries and D65. | Not an identical complete viewing pipeline. |
DO. OBSERVE. EXPLAIN.
Compare two explicit blends.
Choose black/white, 50% and no quantization in the lab. Repeat with coloured endpoints.
- 01
Average codes
Select encoded mixing at 50%. Record midpoint code C and linear Y beneath the corresponding swatch.
CHECKA code near 0.5 gives a linear level near 0.214, not 0.5.
- 02
Average light
Keep the black/white pair and 50% weight; select linear mixing and compare both numeric readouts.
CHECKLinear Y is near 0.5 and the sRGB code near 0.735: the operation changes, not the endpoints.
- 03
Add quantization
Choose coloured endpoints and compare unquantized, 8-bit and 10-bit results at one fixed blend weight.
CHECKQuantization changes precision, not the sRGB definition.
- 04
Follow the graphics
In a project copy, identify the image’s input profile and intended video output.
CHECKYou do not treat finished sRGB graphics as camera Log.
Check your understanding
TEST THE EXPLANATION
A situation, not a trick question.
Answer first. Open the reasoning afterwards.
01Two gradients differ despite identical endpoints.+
They may use different interpolation spaces. Compare the operation before blaming a damaged profile.
02Can a tag alone change sRGB into Rec.709?+
First define the intended signal meaning. Shared primaries do not remove transfer and viewing differences; arbitrary reassignment can alter appearance.
03Does a 10-bit file cease to be sRGB?+
No. Storage precision and space definition are separate properties; verify format support separately.
Avoid these shortcuts
- Reducing exact sRGB to a 2.2 power.
- Calling code 128/255 half of linear light.
- Confusing profile assignment with conversion.
From the model to your workflow.
In production, identify the source, transform placement and required output separately. The lab verifies a bounded mathematical model; it does not measure your display, replace project settings or certify an exported file.
CHECK THE FOUNDATION
Sources and model limits.
Editorial review: 1 October 2026. Links identify the definitions and formulas; exercises are independent teaching examples.
- [1]ITU-R · BT.709-6 (2015) ↗
HDTV parameters: primaries, source OETF, component signals and digital levels. Tables 1 and 3 checked in the PDF.
- [2]ITU-R · BT.1886 (2011) ↗
Annex 1: reference EOTF including black and white luminance. Lab luminances are model calculations, not display measurements.
- [4]ICC · sRGB ↗
Primary coordinates and D65; piecewise sRGB transfer function.
- [5]ICC · Display P3 ↗
P3, D65 and the sRGB transfer: a distinct definition, not DCI-P3.
- [10]W3C · CSS Color 4 ↗
Candidate Recommendation Draft, 30 September 2026: explicit CSS spaces, interpolation and gamut mapping. A work in progress.
- [21]Blackmagic Design · Color ↗
Color and colour-management context. Setting names and locations depend on Resolve version and configuration; live projects were not tested here.
- [22]ITU-T · H.273 (2024) ↗
Identification of video primaries, transfer characteristics and matrix coefficients.
- [23]W3C · A Standard Default Color Space for the Internet ↗
Historical sRGB proposal. Exact transfer coefficients are taken from the ICC registry.