Tone Mapping Tone Mapping (HDR-to-HDR and HDR-to-SDR)

What tone mapping is

Tone mapping is the process by which a display, projector, or video processor analyzes HDR content and renders it within the brightness, black-level, and color-volume capabilities of the actual output device. It exists because virtually no consumer display can reproduce the full BT.2100 PQ luminance range used during mastering, so the signal must be re-mapped on the way out.

The work splits into two distinct conversion problems. HDR-to-SDR is the format-level down-conversion used to derive an SDR cut from an HDR master, governed by BT.2408 and BT.2446 production guidance. HDR-to-HDR is the per-display mapping every HDR-capable TV or projector performs in real time when content is mastered above its peak luminance, addressed by BT.2390. Both share the same underlying problem — fitting a wider range into a narrower one — but the targets and viewing environments differ.

The gap that forces tone mapping to exist is large. HDR titles are commonly mastered to peak luminances of 1,000, 4,000, or theoretically 10,000 nits, while consumer TVs typically reach 500 to 2,000 nits (high-end models past 3,000) and home theater projectors only roughly 100 to 300 nits. Tone mapping bridges this gap on every playback chain, and the gap is the largest single reason different displays look different on the same HDR title.

How tone mapping works

The simplest possible response to over-range content is a hard clip: every code value above the display's peak luminance is pinned to peak, instantly losing all gradation in the brightest highlights. A soft roll-off — the actual 'tone mapping' approach — compresses the above-peak range with a curve so that brighter source values still map to slightly brighter output values, preserving relative highlight gradation at the cost of slightly dimmer overall brightness. BT.2408 explicitly notes that hard clipping is sometimes preferable to soft tone mapping in HDR-to-SDR conversion, and the choice between them is intentional, not accidental.

BT.2390 specifies a reference EETF (Electrical-to-Electrical Transfer Function) for HDR-to-HDR display mapping. The curve is piecewise: a 1:1 linear pass-through across the central midtone region, a hermite-spline 'knee' that smoothly rolls off the highlights into the display's peak luminance, and a low-end boost that lifts near-black values to compensate for the display's higher black floor. This curve serves as a vendor-neutral reference behavior — manufacturers may deviate, but BT.2390 defines what 'correct' baseline display mapping looks like.

Above the curve shape sits the question of how often it is allowed to change. Static tone mapping derives a single curve from a title's static metadata (MaxCLL, MaxFALL) and applies it to every frame, which means the curve must accommodate the brightest moment in the entire runtime — often at the cost of darker scenes. Dynamic tone mapping changes the curve scene-by-scene or frame-by-frame, either driven by per-scene metadata (HDR10+, Dolby Vision) or by real-time content analysis on the display side (LG/Samsung/Sony Dynamic Tone Mapping, JVC Frame Adapt HDR, MadVR DTM). Dynamic tone mapping yields more consistent perceived brightness across very dark and very bright scenes.

Tone mapping in home theater

Every HDR-capable TV runs a tone-mapping stage on incoming HDR content whenever the source's peak luminance exceeds the panel's peak. Reviewers describe this as the reason peak-brightness measurements matter at all in HDR — the TV must sacrifice either overall brightness or highlight detail to fit the signal into the panel's range, and a brighter panel needs to compress less. The differences between HDR TVs in side-by-side reviews are largely tone-mapping behavior, not raw panel specs.

Projectors face a much larger HDR mapping gap than TVs because their peak luminance is roughly 100–300 nits against content mastered at 1,000–4,000 nits, and most projectors do not support Dolby Vision or HDR10+. JVC's Frame Adapt HDR, introduced firmware-side circa 2019, is the canonical example of vendor-side dynamic tone mapping for projectors: it analyzes the HDR10 signal frame-by-frame or scene-by-scene with no metadata required, then re-maps the dynamic range to the projector's optical capability — replacing the older static-curve approach where users manually adjusted between titles.

Dedicated external processors like the MadVR Envy sit between the source and display, performing per-frame dynamic tone mapping in their own GPU silicon and passing an already-mapped image to the TV or projector — bypassing the display's internal tone mapper. MadVR Labs describes its DTM 2.0 as analyzing every frame in real time across roughly half a billion pixels per second at 4K60. This pattern is most common in high-end projector setups where the display's own tone mapping is the weakest link.

HGiG inverts the usual flow for games. Instead of the TV tone-mapping the console's HDR output, the game queries the display's reported peak/minimum luminance (MaxTML, MinTML, MaxFFTML) via the console, tone-maps internally, and sends a signal already fitted to the TV's range. The TV's HGIG mode then disables its own dynamic tone mapping for that input, passing the game's mapped output through verbatim — the cleanest way to keep gameplay-critical highlight detail consistent across different TVs. The full handoff is covered in the HGiG entry.

Tone-mapping pitfalls and confusions

Calling a feature 'dynamic tone mapping' is ambiguous unless you know which side is doing the work. Content-driven dynamic tone mapping (HDR10+, Dolby Vision) requires the content to carry per-scene metadata and the display to support that format. Display-driven dynamic tone mapping (LG/Samsung 'Dynamic Tone Mapping', JVC Frame Adapt, MadVR DTM) analyzes the picture in real time and works on plain HDR10 — but produces a result that may not match the colorist's intent because no metadata was carried. Static HDR10 fed to a display without DTM is the static-only baseline.

Tone-mapping behavior also sits on an aggressive-versus-conservative spectrum. An aggressive mapper pushes mid-tones and highlights closer to the display's peak — the image looks brighter and punchier, but values near the master's peak get compressed into the panel's ceiling and lose differentiation (the sun and a chrome reflection both clip white). A conservative mapper holds mid-tones lower and reserves headroom for highlights — bright detail like specular sparkles in clouds survives, but the overall scene can look dim against a brighter neighbor. Neither is 'wrong'; the right setting depends on content and viewing environment.

This is also why there is no objectively correct tone-mapping curve outside of one specific case — when the display's capability matches or exceeds the content's master peak, in which case no mapping is needed. For everything else, the 'right' tone map is a judgment call between honoring the colorist's intent (conservative, more accurate to reference) and maximizing perceptual punch on the actual viewing setup (aggressive, often more pleasing in a bright living room). This is the reason Dolby Vision and HDR10+ exist in the first place: they let the colorist describe the intended trim per scene, instead of leaving the display to guess.