Gamma is the power-function curve that connects the numbers stored in a video file to the brightness a screen produces - and almost every "why does this look darker/lighter on their screen" conversation is a gamma conversation in disguise. The two numbers that anchor it: the broadcast reference display gamma is 2.404 (ITU-R BT.1886), while sRGB uses a piecewise curve that is only approximately gamma 2.2 - a small linear toe below a threshold, then a power segment above it. Neither is arbitrary, and the gap between them is exactly why a cut graded on one display can look flat on another. This is the gamma layer of the review-link color story; the primaries and white-point layer is in the color-space explainer.
The definition, with the real equation
ITU-R BT.1886 - the reference electro-optical transfer function for flat panel displays in HDTV production - specifies the display behavior as a power function: L = a(max[(V + b), 0])^γ with γ = 2.404, where V is the input signal value and L the produced luminance. The exponent is the "gamma"; the a and b terms account for real displays' black level and lag. The intuition to keep: the curve is nonlinear on purpose - human vision is more sensitive to differences in shadows than in highlights, so display curves allocate signal precision accordingly (the encoding rationale appears in how encoders spend bits too).
sRGB is not "exactly 2.2"
The web's default space, sRGB (IEC 61966-2-1), is a piecewise function: below an input of 0.0031308 it is a straight line (encoded = 12.92 × X), above it the power segment 1.055 × X^(1/2.4) − 0.055 - "gamma 2.2" is the shorthand, but the actual curve is that hybrid. This is why "sRGB vs gamma 2.4" is not one number versus another: it is a piecewise curve versus a pure 2.404 power function, and their divergence concentrates in the shadows and midtones. The W3C's CSS Color Module makes the web's default explicit - untagged colors "must be treated as being in the sRGB color space" - which is why browser video playback lives on the sRGB side of this divide.
Why a mismatch moves midtones most
A power curve's relative error against another power curve is largest where the curves bend - the mid region of the signal. In practice: display a Rec.709/BT.1886-conformed cut on a display behaving closer to a plain 2.2, and midtones (faces, walls, the whole middle of the image) shift brightness while blacks and peak whites barely move. That is the "flat" or "milky" review-link report - one of the five diagnosable sources in why video looks different on a review link, and the HDR-scale version of the same story (PQ data shown without tone mapping) is in the HDR review explainer.
2.2 vs 2.4 in practice
| Surface | Curve | What it means for you |
|---|---|---|
| Broadcast / mastering reference display | BT.1886 pure power, γ = 2.404 | The grading-suite target; creative decisions are judged here |
| sRGB web surfaces (browsers, untagged content) | Piecewise ≈2.2 (12.92 toe, then 2.4-segment) | Where review links and feeds live; slightly brighter midtones than a 2.4 suite |
| Untagged video in a web player | Whatever the player assumes - commonly sRGB | The mismatch surface; export tagged files so the player does not guess |
The practical rules fall out of the table: grade in a controlled environment (the 2.4 reference is what the industry judges against), and for delivery, export properly tagged Rec.709 files so players know which curve to assume - the export-side settings are in what format to export for client review. Expect the residual: a client's uncalibrated laptop sits wherever its factory calibration put it, which is why approval questions belong on a review platform's consistent player, not on whichever screen opens first.
FAQ
Is gamma 2.2 or 2.4?
Both, for different surfaces: BT.1886 defines the broadcast reference display at γ = 2.404, and sRGB - the web default - is a piecewise curve (a linear toe, then a 2.4-segment power function) commonly summarized as "gamma 2.2." Grading suites target 2.4; browsers live on the sRGB curve.
Is sRGB gamma exactly 2.2?
No - that is the shorthand. The IEC sRGB standard specifies a linear segment below 0.0031308 (encoded = 12.92 × X) and a power segment above it (1.055 × X^(1/2.4) − 0.055), which together only approximate a 2.2 power curve.
Why does a gamma mismatch shift midtones instead of everything?
Power curves diverge most where they bend - the middle of the signal range - so a 2.4-conformed image shown on a ≈2.2-behaving surface moves midtone brightness (faces, mid-gray walls) while shadows and highlights shift far less. That midtone-heavy shift is what reads as "flat."
What gamma should I master for web delivery?
Deliver a properly tagged Rec.709 file mastered in a controlled suite (BT.1886 reference) - the tag tells players what the data assumes. The residual variance lives in the client's display, which no export setting can fix.
Does gamma explain my review link looking different from my edit bay?
It is one of the standard sources - a 2.4-reference suite versus an sRGB browser surface shifts midtones predictably. Work the full diagnosis (tagging, color management, compression, display state) in the review-link color-shift explainer before regrading.