ST.2086 Demystified: From Codec Constraints to Metadata Mastery with HDRmaster
High Dynamic Range (HDR) has revolutionized visual storytelling. But how is this visual splendor maintained across different encoding formats and display interfaces? The unsung hero is often ST.2086, a SMPTE standard that defines static metadata for HDR content.
The History of ST.2086 Implementations in Consumer Electronics
The advent of ST.2086 marked a significant milestone in the HDR ecosystem. Initially, its implementation was limited to professional-grade monitors and broadcasting equipment. However, as HDR gained traction, consumer electronics manufacturers started incorporating ST.2086 support into their devices.
Early adopters included brands like Sony, Samsung, and LG, who integrated ST.2086 into their high-end television models. Over the years, the standard has trickled down to even budget-friendly models, making HDR a mainstream feature.
ST.2086: The Universal Language of HDR
ST.2086 is more than just a standard; it’s a universal language that HDR-enabled systems understand. It provides metadata like MaxFALL and MaxCLL to guide the rendering of HDR content, ensuring a uniform HDR experience across different platforms and formats.
ST.2086 Across Codecs: H.264, H.265, H.266
- H.264: In H.264, ST.2086 metadata is embedded in Supplemental Enhancement Information (SEI) messages, which are part of the bitstream.
- H.265: Similar to H.264, H.265 also uses SEI messages for conveying ST.2086 metadata. However, H.265 allows for more efficient encoding, making it a more attractive option for HDR content.
- H.266: The newest in the lineup, H.266 offers even greater efficiency but also comes with its own set of SEI messages and structures that can house ST.2086 metadata.
SEI vs VUI: The Carriers of ST.2086 Metadata in HDR Ecosystems
Supplemental Enhancement Information (SEI)
SEI messages are flexible, insertable at various points in the video stream, and particularly useful for carrying metadata that may change during the video or for inserting additional data that wasn’t originally included. In the context of ST.2086:
- H.264: SEI messages carry the ST.2086 metadata and are part of the bitstream.
- H.265 (HEVC): Like H.264, H.265 also uses SEI messages for ST.2086 metadata. These messages are more efficiently encoded in H.265.
- H.266 (VVC): H.266 introduces a new and more flexible set of SEI messages that can also carry ST.2086 metadata, allowing for easier future-proofing.
Video Usability Information (VUI)
VUI is integral for HDR as it sets other essential parameters that influence HDR:
- Color Space Information: Specifies color primaries, transfer characteristics, and matrix coefficients.
- Bit Depth: Indicates the bit depth of the luma and chroma samples, which is critical for HDR’s increased dynamic range.
- EOTF: Defines the Electro-Optical Transfer Function, which is pivotal in HDR rendering. It sets the relationship between encoded pixel values and displayed brightness. In the context of HDR, EOTF like ST.2084 (often represented numerically as “16”) can be specified.
In different codecs:
- H.264: Defines essential groundwork for HDR rendering, including the ability to specify EOTF like ST.2084.
- H.265 (HEVC): Similar to H.264, can define EOTF and sets other parameters that influence HDR rendering, like ST.2084.
- H.266 (VVC): Continues to offer the ability to define EOTF and sets essential parameters for HDR interpretation, including the option for ST.2084.
SEI vs. VUI conclusion
- SEI is More Flexible: SEI messages are more flexible but potentially less stable.
- VUI is More Stable: VUI is generally more stable but less flexible.
- ST.2086 is Typically in SEI: ST.2086 metadata is usually carried in SEI messages.
- VUI Defines EOTF for HDR: VUI can define the EOTF, often specifying standards like ST.2084, setting the stage for proper HDR rendering.
From Codec to Display: HDMI AVI InfoFrames
When the HDR content moves from a source device to a display via HDMI, the ST.2086 metadata is often transliterated into HDMI AVI (Auxiliary Video Information) InfoFrames. This ensures that the HDR rendering instructions are faithfully communicated to the display.
Metadata Handling with HDRmaster
The InnoPQ HDRmaster software, available at InnoPQ HDRmaster, solves the challenges of injecting and editing ST.2086 metadata for both H.264 and H.265. It also offers:
- Quality Control: Enables in-depth Q/C of video content.
- HDR10+ Support: Allows for the creation and injection of HDR10+ metadata (H.265 only).
Note: HDR10+ metadata can only be injected into H.265, not H.264 or H.266.
Terms, References and Standards
- ST 2086:2018 – SMPTE Standard – Mastering Display Color Volume Metadata Supporting High Luminance and Wide Color Gamut Images,” in ST 2086:2018 , vol., no., pp.1-8, 27 April 2018
- Supplemental Enhancement Information (SEI): SEI messages are flexible, insertable at various points in the video stream, and particularly useful for carrying metadata that may change during the video or for inserting additional data that wasn’t originally included. In the context of ST.2086:
- H.264: SEI messages carry the ST.2086 metadata and are part of the bitstream. Link to ITU-T H.264
- H.265 (HEVC): Like H.264, H.265 also uses SEI messages for ST.2086 metadata. These messages are more efficiently encoded in H.265. Link to ITU-T H.265
- H.266 (VVC): H.266 introduces a new and more flexible set of SEI messages that can also carry ST.2086 metadata, allowing for easier future-proofing. Link to ITU-T H.266
- Video Usability Information (VUI): VUI is integral for HDR as it sets other essential parameters that influence HDR:
- Color Space Information: Specifies color primaries, transfer characteristics, and matrix coefficients.
- Bit Depth: Indicates the bit depth of the luma and chroma samples, which is critical for HDR’s increased dynamic range.
- EOTF: Defines the Electro-Optical Transfer Function, which is pivotal in HDR rendering. It sets the relationship between encoded pixel values and displayed brightness. In the context of HDR, EOTF like ST.2084 (often represented numerically as “16”) can be specified.