C Specification
When calling vkGetPhysicalDeviceVideoCapabilitiesKHR with
pVideoProfile->videoCodecOperation
specifying an encode operation, the
VkVideoEncodeCapabilitiesKHR structure must be included in the
pNext
chain of the VkVideoCapabilitiesKHR structure to retrieve
capabilities specific to video encoding.
The VkVideoEncodeCapabilitiesKHR
structure is defined as:
// Provided by VK_KHR_video_encode_queue
typedef struct VkVideoEncodeCapabilitiesKHR {
VkStructureType sType;
void* pNext;
VkVideoEncodeCapabilityFlagsKHR flags;
VkVideoEncodeRateControlModeFlagsKHR rateControlModes;
uint32_t maxRateControlLayers;
uint64_t maxBitrate;
uint32_t maxQualityLevels;
VkExtent2D encodeInputPictureGranularity;
VkVideoEncodeFeedbackFlagsKHR supportedEncodeFeedbackFlags;
} VkVideoEncodeCapabilitiesKHR;
Members
-
sType
is a VkStructureType value identifying this structure. -
pNext
isNULL
or a pointer to a structure extending this structure. -
flags
is a bitmask of VkVideoEncodeCapabilityFlagBitsKHR describing supported encoding features. -
rateControlModes
is a bitmask of VkVideoEncodeRateControlModeFlagBitsKHR indicating supported rate control modes. -
maxRateControlLayers
indicates the maximum number of rate control layers supported. -
maxBitrate
indicates the maximum supported bitrate. -
maxQualityLevels
indicates the number of discrete video encode quality levels supported. Implementations must support at least one quality level. -
encodeInputPictureGranularity
indicates the granularity at which encode input picture data is encoded and may indicate a texel granularity up to the size of the codec-specific coding block size. This capability does not impose any valid usage constraints on the application, however, depending on the contents of the encode input picture, it may have effects on the encoded bitstream, as described in more detail below. -
supportedEncodeFeedbackFlags
is a bitmask of VkVideoEncodeFeedbackFlagBitsKHR values specifying the supported flags for video encode feedback queries.
Description
Implementations must include support for at least
VK_VIDEO_ENCODE_FEEDBACK_BITSTREAM_BUFFER_OFFSET_BIT_KHR
and
VK_VIDEO_ENCODE_FEEDBACK_BITSTREAM_BYTES_WRITTEN_BIT_KHR
in
supportedEncodeFeedbackFlags
.
encodeInputPictureGranularity
provides information about the way
encode input picture data is used as input to video
encode operations.
In particular, some implementations may not be able to limit the set of
texels used to encode the output video bitstream to the image subregion
specified in the VkVideoPictureResourceInfoKHR structure corresponding
to the encode input picture (i.e. to the resolution of the image data to
encode specified in its codedExtent
member).
Note
For example, the application requests the coded extent to be 1920x1080, but
the implementation is only able to source the encode input picture data at
the granularity of the codec-specific coding block size which is 16x16
pixels (or as otherwise indicated in |
If codedExtent
rounded up to the next integer multiple of
encodeInputPictureGranularity
is greater than the extent of the image
subresource specified for the encode input picture,
then the texel values corresponding to texel coordinates outside of the
bounds of the image subresource may be undefined.
However, implementations should use well-defined default values for such
texels in order to maximize the encoding efficiency for the last coding
block row/column, and/or to ensure consistent encoding results across
repeated encoding of the same input content.
Nonetheless, the values used for such texels must not have an effect on
whether the video encode operation produces a compliant bitstream, and must
not have any other effects on the encoded picture data beyond what may
otherwise result from using these texel values as input to any compression
algorithm, as defined in the used video compression standard.
Note
While not required, it is generally a good practice for applications to make
sure that the image subresource used for the encode input picture has an
extent that is an integer multiple of the codec-specific coding block size
(or at least |
Document Notes
For more information, see the Vulkan Specification
This page is extracted from the Vulkan Specification. Fixes and changes should be made to the Specification, not directly.