VkVideoDecodeAV1PictureInfoKHR

The VkVideoDecodeAV1PictureInfoKHR structure is defined as:

typedef struct VkVideoDecodeAV1PictureInfoKHR {
    VkStructureType sType;
    const void* pNext;
    const StdVideoDecodeAV1PictureInfo* pStdPictureInfo;
    int32_t referenceNameSlotIndices[VK_MAX_VIDEO_AV1_REFERENCES_PER_FRAME_KHR];
    uint32_t frameHeaderOffset;
    uint32_t tileCount;
    const uint32_t* pTileOffsets;
    const uint32_t* pTileSizes;
} VkVideoDecodeAV1PictureInfoKHR;

• sType is a VkStructureType value identifying this structure.
• pNext is NULL or a pointer to a structure extending this structure.
• pStdPictureInfo is a pointer to a StdVideoDecodeAV1PictureInfo structure specifying AV1 picture information.
• referenceNameSlotIndices is an array of seven (VK_MAX_VIDEO_AV1_REFERENCES_PER_FRAME_KHR, which is equal to the Video Std definition STD_VIDEO_AV1_REFS_PER_FRAME) signed integer values specifying the index of the DPB slot or a negative integer value for each AV1 reference name used for inter coding. In particular, the DPB slot index for the AV1 reference name frame is specified in referenceNameSlotIndices[frame - STD_VIDEO_AV1_REFERENCE_NAME_LAST_FRAME].
• frameHeaderOffset is the byte offset of the AV1 frame header OBU, as defined in section 5.9 of the AV1 Specification, within the video bitstream buffer range specified in VkVideoDecodeInfoKHR.
• tileCount is the number of elements in pTileOffsets and pTileSizes.
• pTileOffsets is a pointer to an array of tileCount integers specifying the byte offset of the tiles of the picture within the video bitstream buffer range specified in VkVideoDecodeInfoKHR.
• pTileSizes is a pointer to an array of tileCount integers specifying the byte size of the tiles of the picture within the video bitstream buffer range specified in VkVideoDecodeInfoKHR.

This structure is specified in the pNext chain of the VkVideoDecodeInfoKHR structure passed to vkCmdDecodeVideoKHR to specify the codec-specific picture information for an AV1 decode operation.

Decode Output Picture Information

When this structure is specified in the pNext chain of the VkVideoDecodeInfoKHR structure passed to vkCmdDecodeVideoKHR, the information related to the decode output picture is defined as follows:

• The image subregion used is determined according to the AV1 Decode Picture Data Access section.
• The decode output picture is associated with the AV1 picture information provided in pStdPictureInfo.

Std Picture Information

The members of the StdVideoDecodeAV1PictureInfo structure pointed to by pStdPictureInfo are interpreted as follows:

• flags.reserved, reserved1, and reserved2 are used only for padding purposes and are otherwise ignored;
• flags.apply_grain indicates that film grain is enabled for the decoded picture, as defined in section 6.8.20 of the AV1 Specification;
• tg_start and tg_end are interpreted as defined in section 6.10.1 of the AV1 Specification;
• OrderHint, OrderHints, and expectedFrameId are interpreted as defined in section 6.8.2 of the AV1 Specification;
• the StdVideoAV1TileInfo structure pointed to by pTileInfo is interpreted as follows:
  • flags.reserved and reserved1 are used only for padding purposes and are otherwise ignored;
  • pMiColStarts is a pointer to an array of TileCols number of unsigned integers that corresponds to MiColStarts defined in section 6.8.14 of the AV1 Specification;
  • pMiRowStarts is a pointer to an array of TileRows number of unsigned integers that corresponds to MiRowStarts defined in section 6.8.14 of the AV1 Specification;
  • pWidthInSbsMinus1 is a pointer to an array of TileCols number of unsigned integers that corresponds to width_in_sbs_minus_1 defined in section 6.8.14 of the AV1 Specification;
  • pHeightInSbsMinus1 is a pointer to an array of TileRows number of unsigned integers that corresponds to height_in_sbs_minus_1 defined in section 6.8.14 of the AV1 Specification;
  • all other members of StdVideoAV1TileInfo are interpreted as defined in section 6.8.14 of the AV1 Specification;
• the StdVideoAV1Quantization structure pointed to by pQuantization is interpreted as follows:
  • flags.reserved is used only for padding purposes and is otherwise ignored;
  • all other members of StdVideoAV1Quantization are interpreted as defined in section 6.8.11 of the AV1 Specification;
• if flags.segmentation_enabled is set, then the StdVideoAV1Segmentation structure pointed to by pSegmentation is interpreted as follows:
  • the elements of FeatureEnabled are bitmasks where bit index j of element i corresponds to FeatureEnabled[i][j] as defined in section 6.8.13 of the AV1 Specification;
  • FeatureData is interpreted as defined in section 6.8.13 of the AV1 Specification;
• the StdVideoAV1LoopFilter structure pointed to by pLoopFilter is interpreted as follows:
  • flags.reserved is used only for padding purposes and is otherwise ignored;
  • update_ref_delta is a bitmask where bit index i is interpreted as the value of update_ref_delta corresponding to element i of loop_filter_ref_deltas as defined in section 6.8.10 of the AV1 Specification;
  • update_mode_delta is a bitmask where bit index i is interpreted as the value of update_mode_delta corresponding to element i of loop_filter_mode_deltas as defined in section 6.8.10 of the AV1 Specification;
  • all other members of StdVideoAV1LoopFilter are interpreted as defined in section 6.8.10 of the AV1 Specification;
• if flags.enable_cdef is set in the active sequence header, then the members of the StdVideoAV1CDEF structure pointed to by pCDEF are interpreted as follows:
  • cdef_y_sec_strength and cdef_uv_sec_strength are the bitstream values of the corresponding syntax elements defined in section 5.9.19 of the AV1 Specification;
  • all other members of StdVideoAV1CDEF are interpreted as defined in section 6.10.14 of the AV1 Specification;
• the StdVideoAV1LoopRestoration structure pointed to by pLoopRestoration is interpreted as follows:
  • LoopRestorationSize[plane] is interpreted as log2(size) - 5, where size is the value of LoopRestorationSize[plane] as defined in section 6.10.15 of the AV1 Specification.
  • all other members of StdVideoAV1LoopRestoration are defined as in section 6.10.15 of the AV1 Specification;
• the members of the StdVideoAV1GlobalMotion structure provided in global_motion are interpreted as defined in section 7.10 of the AV1 Specification;
• if flags.film_grain_params_present is set in the active sequence header, then the StdVideoAV1FilmGrain structure pointed to by pFilmGrain is interpreted as follows:
  • flags.reserved is used only for padding purposes and is otherwise ignored;
  • all other members of StdVideoAV1FilmGrain are interpreted as defined in section 6.8.20 of the AV1 Specification;
• all other members are interpreted as defined in section 6.8 of the AV1 Specification.

When film grain is enabled for the decoded frame, the flags.update_grain and film_grain_params_ref_idx values specified in StdVideoAV1FilmGrain are ignored by AV1 decode operations and the load_grain_params function, as defined in section 6.8.20 of the AV1 Specification, is not executed. Instead, the application is responsible for specifying the effective film grain parameters for the frame in StdVideoAV1FilmGrain.

When film grain is enabled for the decoded frame, the application is required to specify a different decode output picture resource in VkVideoDecodeInfoKHR::dstPictureResource compared to the reconstructed picture specified in VkVideoDecodeInfoKHR::pSetupReferenceSlot→pPictureResource even if the implementation does not report support for VK_VIDEO_DECODE_CAPABILITY_DPB_AND_OUTPUT_DISTINCT_BIT_KHR in VkVideoDecodeCapabilitiesKHR::flags for the video decode profile.

Reference picture setup is controlled by the value of StdVideoDecodeAV1PictureInfo::refresh_frame_flags. If it is not zero and a reconstructed picture is specified, then the latter is used as the target of picture reconstruction to activate the DPB slot specified in pDecodeInfo→pSetupReferenceSlot→slotIndex. If StdVideoDecodeAV1PictureInfo::refresh_frame_flags is zero, but a reconstructed picture is specified, then the corresponding picture reference associated with the DPB slot is invalidated, as described in the DPB Slot States section.

Active Parameter Sets

The active sequence header is the AV1 sequence header stored in the bound video session parameters object.