Three parts in one file
The specification describes a file format that encodes a logarithmic range gain map image inside a JPEG. In practice that means three components: a normal SDR photo, a gain map that says how much brighter each region should get, and metadata telling a reader how to apply it.
What each part does
- The base image An ordinary JPEG rendition. This is what every existing viewer displays, and it is why the format did not break the world when it arrived.
- The gain map A second JPEG, much smaller, acting as a per-region multiplier. The Vaultaire converter writes it at quarter size.
- The metadata XMP values and an MPF index that tell a compatible reader the map is there and how to interpret it.
The Ultra HDR Image Format specification states that legacy readers which do not support the format read and display the conventional low dynamic range image, while readers that do support it combine the primary image with the gain map and render a high dynamic range image on compatible displays.
Why it was built this way
An HDR display can show brighter whites than an SDR one, but by how much varies. The specification is explicit that the display technique must adapt to any ratio between a display's HDR and SDR white points, which can vary significantly between devices, or even over time on a single device.
A single fixed HDR image cannot do that. A base image plus a multiplier can, because the reader decides how far to push the map for the screen in front of it. The spec also requires the technique to work without clipping highlights, crushing shadows, compressing local contrast, or changing the tonal relationships between objects in the scene. Those four constraints are why a gain map beats simply encoding a brighter picture.
The other design requirement was size. The extra data is a compressed grayscale-style map, not a second full-resolution photograph, so the file grows modestly rather than doubling.
Why your HDR photo often looks normal
Ultra HDR degrades quietly, which is its best feature and its most confusing one. Nothing errors. You just get the SDR picture.
| Cause | What happens |
|---|---|
| The viewer does not support gain maps | It shows the base image, exactly as intended |
| The display is SDR | There is no extra headroom to render into |
| Screen brightness or a power-saving mode | The available headroom shrinks, so the effect shrinks |
| An editor re-saved the file | Many editors write a plain JPEG and drop the map |
| A social platform re-encoded the upload | Support varies by app, account, device, and upload path |
Test the actual publishing path before you rely on the HDR rendition anywhere. Support changes often enough that a result from last year proves nothing about today.
Converting in both directions
The converter reads the direction from the file itself. It works on one opaque still JPEG, PNG, or WebP, or an Ultra HDR JPEG, up to 20 MB, 8,192 pixels per side, and 16,000,000 pixels total.
- Ultra HDR to plain SDR The base rendition inside the file is already the correct SDR answer, so it is read out, re-encoded as a normal JPEG, and checked to confirm the download declares no gain map. No look to choose.
- SDR to Ultra HDR A gain map is generated from the image and packaged with an SDR fallback, the required XMP values, and an MPF index. You pick Natural, Glow, or Graphic for the highlight response.
- Output is always JPEG A PNG or WebP source becomes an Ultra HDR JPEG, because the format defines the base image, gain map, XMP values, and MPF index as JPEG data.
- Verified before download The container is parsed and both embedded JPEG images are decoded first. Everything runs in a browser worker and no file is uploaded.
One honest limit on the SDR to HDR direction: generating a gain map cannot recover highlight detail the original never captured. It changes how a compatible display renders the range that is already in the file. If the photo is simply too dark, the brightening guide is the right starting point instead.
On metadata, this tool is the one exception among the Vaultaire converters. Supported source metadata is not copied, but Ultra HDR requires XMP gain-map values and an MPF index, so the output is not metadata-free by design.
Ultra HDR FAQ
What is an Ultra HDR image?
A JPEG holding a normal SDR base image, a second JPEG gain map, and metadata describing how a compatible reader combines them. The spec defines it as a logarithmic range gain map image encoded inside a JPEG file.
Why does it still open in old software?
Backward compatibility was a design requirement. Readers that do not understand the format display the conventional low dynamic range image and ignore the rest.
Why does my HDR photo look normal?
It depends on the device, display mode, brightness, and the app showing it. It also depends on the gain map surviving whatever edited or shared the file, because many services re-encode and drop it.
Can I convert Ultra HDR to a normal JPG?
Yes. The base rendition inside the file is the correct SDR result, so it is read out, re-encoded as a plain JPEG, and checked to confirm no gain map is declared.
Can I turn a normal photo into Ultra HDR?
You can package one with a generated gain map. That cannot recover highlight detail the original never captured. It shapes how a compatible display renders the range already there.
Does the file get much bigger?
Modestly. The map is a compressed reduced-size image, written here at quarter size, not a second full-resolution photograph. Keeping the extra data small was one of the format's stated requirements.