Learn / HDR Formats

What is Dolby Vision?

The proprietary, dynamic-metadata HDR format built on top of PQ — and why it's a different animal from the gain-map photo HDR this site actually produces.

Dolby Vision is Dolby's proprietary HDR video format — the one you'll see called out by name on streaming service badges, TV boxes, and Blu-ray cases, usually right next to (or instead of) HDR10. It's built on the same underlying transfer function as HDR10 — PQ, SMPTE ST 2084 — but adds something HDR10 doesn't have: metadata that changes over the course of the video instead of staying fixed.

Static metadata vs. dynamic metadata

A baseline HDR10 stream carries one set of brightness metadata (MaxCLL and MaxFALL — the maximum content light level and maximum frame-average light level) for the entire piece of content, start to finish. That's a real limitation: a movie with one blindingly bright sunrise shot and one dim, moody interior scene gets tone-mapped using the same numbers for both, because the display has no way to know the two scenes are that different.

Dolby Vision carries metadata that can change scene by scene, or even frame by frame, describing how that specific content should be tone-mapped for the specific display showing it. In principle, that means each shot gets treated on its own terms instead of being squeezed through one compromise curve for the whole runtime.

The cost of that precision

Dolby Vision isn't an open standard. Studios pay to master content in it, device makers pay to license decoder support, and the whole pipeline — camera, post-production, delivery, display — has to be Dolby-certified to carry the name. HDR10 is royalty-free and universally supported by comparison, which is exactly why it's the mandatory baseline for UHD Blu-ray and most streaming HDR, with Dolby Vision layered on top as an optional, premium addition where a studio and platform have both paid for it.

Dolby Vision is a video-and-broadcast-pipeline format, not a still-photo one. It has no real relationship to the gain-map approach this site's Encode tool uses — don't expect a Dolby Vision decoder to have any idea what to do with a gain-map JPEG, or vice versa.

HDR10+ exists for the same reason, without the license fee

Samsung and Amazon co-developed HDR10+ specifically as a royalty-free answer to Dolby Vision's dynamic metadata — same core idea (per-scene tone-mapping instructions on top of a PQ signal), open format instead of licensed. Support for it is narrower than either HDR10 or Dolby Vision, but it exists precisely because dynamic metadata was worth having without the certification overhead.

Where this fits with what Headroom actually does

Everything on this site is about still images, not video, and about gain-map HDR specifically — the same family of ideas as Ultra HDR and Adobe's Gain Map spec, not Dolby Vision or HDR10. If you're trying to understand what's actually inside an HDR photo you already have, start with what a gain map is, not with video HDR terminology — the two live in genuinely different parts of the HDR landscape.

Questions people actually ask

Can Headroom create Dolby Vision content?
No, and that's by design, not a limitation to fix. Dolby Vision is a licensed, proprietary video format with its own certification process. Headroom builds real gain-map HDR JPEGs for still photos, which is a completely different mechanism serving a different medium.
Is Dolby Vision better than HDR10?
It offers more precision — dynamic, per-scene or per-frame metadata instead of one static value for an entire video — but that comes with licensing costs and certification requirements that HDR10 doesn't have. "Better" depends on whether you're the one paying for that precision or just watching the result; well-mastered HDR10 content can look very close in practice.
Does a gain-map JPEG have anything to do with Dolby Vision?
No — different mechanism entirely. A gain map is an auxiliary grayscale image bundled into a photo file. Dolby Vision is metadata (static or dynamic tone-mapping instructions) that rides alongside a PQ-encoded video signal. Both solve a version of "tell the display how to render HDR content correctly," but neither format understands the other.