HDR Metadata
HDR metadata describes the signal, mastering conditions, or content characteristics used to interpret and adapt an HDR picture.
A media report can show BT.2020 primaries, PQ transfer, a 1,000-nit mastering maximum, content-light fields, and dynamic-metadata information. Those fields describe different parts of the picture's journey. Read them by role before treating them as one HDR quality score.
| Role | Typical information | What a renderer learns |
|---|---|---|
| Signal interpretation | Primaries, transfer, matrix, range | What the decoded numbers mean |
| Mastering context | Display primaries, white point, luminance limits | The reference volume used for judgment |
| Content bounds | MaxCLL, MaxFALL | Reported light-level bounds of the produced sequence |
| Adaptive instructions or statistics | Format-specific dynamic metadata | Information applicable to a shot or frame |
Signal interpretation is often called color signaling rather than HDR metadata in the narrower sense. It is essential because a valid compressed stream can reconstruct samples whose meaning still depends on those descriptions.
Color interpretation, mastering context, and content bounds
Suppose a source has BT.2020 primaries and PQ, a mastering maximum of 1,000 nits, and a reported MaxCLL of 600 nits. The first pair specifies color and light coordinates. The mastering number describes a display configuration. The content number supplies a bound for the picture under its measurement definition.
A compatible app can interpret the source, read whichever metadata its rendering path supports, and adapt the produced picture to its output. A display with a different range does not have to imitate the mastering monitor's hardware; it needs an appropriate reproduction of the authored image.
Some metadata is carried in codec supplemental messages; other signal descriptions can also appear in container structures. The container reader and video decoder make the needed information available to later processing. Correct packaging and interpretation are both parts of preserving it.
Static and dynamic describe when information applies
Static information applies to the relevant sequence or program. A mastering-display description can remain constant while the movie cuts between a dim room and a bright outdoor scene.
Dynamic metadata can vary with a shot or frame. It can describe that image's brightness distribution or carry adjustments intended for a mapping target. A converter can therefore choose different parameters for scenes with different demands, while managing transitions over time.
The format defines what those parameters mean. A dynamic block is not a hidden higher-resolution picture, and dynamic metadata from one format cannot simply be assigned to a different one. It needs the corresponding processing that interprets its values.
The individual subjects—mastering metadata, content-light bounds, and Dolby Vision RPU—explain their particular roles.
When the pixels change, their descriptions need review
During Emby video transcoding, a new picture stream is created. Several possible operations have different metadata consequences:
- Copying the encoded video into another container can preserve its picture values, while requiring correct carriage of their descriptions.
- Resizing retains or changes the relevant color representation according to the processing path; spatial size alone does not select a new transfer function.
- HDR-to-SDR conversion changes brightness, color, and transfer representation. The output descriptions must identify that converted SDR picture.
- Cutting or reordering a sequence changes which pictures correspond to timed dynamic information. The metadata must still match the delivered sequence.
Blindly copying a source field after processing can leave a plausible number attached to the wrong output. Conversely, dropping essential color signaling can make the destination assign another interpretation to otherwise valid samples.
For a playback investigation, compare the stored source record with the stream actually delivered to the app and the output mode used for presentation. That locates the change: source information, carriage, picture conversion, or final rendering.