Utilities and definitions used when working with VPIImageFormat and VPIPixelFormat values. More...

Macros
#define	VPI_MAKE_SWIZZLE(x, y, z, w) ((VPISwizzle)VPI_DETAIL_MAKE_SWIZZLE(x, y, z, w))
	Creates an user-defined swizzle operation. More...

#define	VPI_MAKE_IMAGE_FORMAT(colorSpace, chromaSubsamp, memLayout, dataType, swizzle, numPlanes, packing0, packing1, ...)
	Creates an user-defined image format constant using abbreviated parameters. More...

#define	VPI_MAKE_PIXEL_FORMAT(memLayout, dataType, packing)
	Creates an user-defined pixel format constant using abbreviated parameters. More...

Enumerations
enum	VPIFormatPacking
	Defines how channels are packed into an image plane element. More...

enum	VPIFormatDataType
	Defines the channel data type. More...

enum	VPIMemLayout
	Defines how the 2D plane pixels are laid out in memory. More...

enum	VPISwizzleChannel
	Defines the format channel names. More...

enum	VPISwizzle
	Defines the supported channel swizzle operations. More...

enum	VPIColorModel
	Defines color models. More...

enum	VPIColorSpace
	Defines color spaces. More...

enum	VPIChromaSubsampling
	Defines how chroma-subsampling is done. More...

enum	VPIImageFormat
	Pre-defined image formats. More...

enum	VPIPixelFormat
	Pre-defined pixel formats. More...

Functions
VPISwizzle	vpiMakeSwizzle (VPISwizzleChannel x, VPISwizzleChannel y, VPISwizzleChannel z, VPISwizzleChannel w)
	Creates an user-defined VPISwizzle operation. More...

VPIChromaSubsampling	vpiMakeChromaSubsampling (int divHoriz, int divVert, int centerSampleHoriz, int centerSampleVert)
	Creates an user-defined VPIChromaSubsampling definition. More...

VPIFormatPacking	vpiMakeFormatPacking (int fixed, int bitsX, int bitsY, int bitsZ, int bitsW)
	Returns a pre-defined VPIFormatPacking given the number of bits in each channel. More...

void	vpiFormatPackingGetParams (VPIFormatPacking packing, int fixed, int bitsX, int bitsY, int bitsZ, int *bitsW)
	Returns channels' information from a format packing. More...

int	vpiFormatPackingGetChannelCount (VPIFormatPacking packing)
	Returns the number of channels defined by the given packing. More...

int	vpiFormatPackingGetBitsPerPixel (VPIFormatPacking packing)
	Returns the number of bits per pixel of the given packing. More...

VPIImageFormat	vpiMakeImageFormat (VPIColorSpace colorSpace, VPIChromaSubsampling chromaSub, VPIMemLayout memLayout, VPIFormatDataType dataType, VPISwizzle swizzle, VPIFormatPacking packing0, VPIFormatPacking packing1, VPIFormatPacking packing2, VPIFormatPacking packing3)
	Creates an user-defined image format. More...

VPIFormatPacking	vpiImageFormatGetPlanePacking (VPIImageFormat fmt, int plane)
	Get the packing for a given plane of an image format. More...

int	vpiImageFormatGetPlaneBitsPerPixel (VPIImageFormat fmt, int plane)
	Get the image format's plane bits per pixel count. More...

VPIImageFormat	vpiImageFormatSetDataType (VPIImageFormat fmt, VPIFormatDataType dataType)
	Set the image format's data type. More...

VPIFormatDataType	vpiImageFormatGetDataType (VPIImageFormat fmt)
	Get the image format's data type. More...

VPISwizzle	vpiImageFormatGetSwizzle (VPIImageFormat fmt)
	Get the image format's channel swizzle operation. More...

VPISwizzle	vpiImageFormatGetPlaneSwizzle (VPIImageFormat fmt, int plane)
	Get the swizzle operation of the given image format's plane. More...

VPISwizzleChannel	vpiImageFormatGetSwizzleX (VPIImageFormat fmt)
	Get the color model channel that corresponds to the first component of given image format. More...

VPISwizzleChannel	vpiImageFormatGetSwizzleY (VPIImageFormat fmt)
	Get the color model channel that corresponds to the second component of given image format. More...

VPISwizzleChannel	vpiImageFormatGetSwizzleZ (VPIImageFormat fmt)
	Get the color model channel that corresponds to the third component of given image format. More...

VPISwizzleChannel	vpiImageFormatGetSwizzleW (VPIImageFormat fmt)
	Get the color model channel that corresponds to the fourth component of given image format. More...

VPIImageFormat	vpiImageFormatSetMemLayout (VPIImageFormat fmt, VPIMemLayout memLayout)
	Set the image format's memory layout. More...

VPIMemLayout	vpiImageFormatGetMemLayout (VPIImageFormat fmt)
	Get the image format's memory layout. More...

VPIImageFormat	vpiImageFormatSetColorSpace (VPIImageFormat fmt, VPIColorSpace colorSpace)
	Set the image format's color space. More...

VPIColorSpace	vpiImageFormatGetColorSpace (VPIImageFormat fmt)
	Get the image format's color space. More...

VPIColorModel	vpiImageFormatGetColorModel (VPIImageFormat fmt)
	Get the image format's color model. More...

VPIImageFormat	vpiImageFormatSetChromaSubsampling (VPIImageFormat fmt, VPIChromaSubsampling css)
	Set the image format's chroma subsampling type. More...

VPIChromaSubsampling	vpiImageFormatGetChromaSubsampling (VPIImageFormat fmt)
	Get the image format's chroma subsampling type. More...

int	vpiImageFormatGetChromaSubsamplingDivHoriz (VPIImageFormat fmt)
	Get the image format's chroma subsampling horizontal downscale factor. More...

int	vpiImageFormatGetChromaSubsamplingDivVert (VPIImageFormat fmt)
	Get the image format's chroma subsampling vertical downscale factor. More...

int	vpiImageFormatGetChromaSubsamplingCenterHoriz (VPIImageFormat fmt)
	Get the image format's chroma horizontal sample position. More...

int	vpiImageFormatGetChromaSubsamplingCenterVert (VPIImageFormat fmt)
	Get the image format's chroma vertical sample position. More...

int	vpiImageFormatGetPlaneChannelCount (VPIImageFormat fmt, int plane)
	Get the number of channels in a plane of an image format. More...

int	vpiImageFormatGetPlaneCount (VPIImageFormat fmt)
	Get the number of planes of an image format. More...

int	vpiImageFormatGetChannelCount (VPIImageFormat fmt)
	Get the total number of channels of an image format. More...

VPIPixelFormat	vpiImageFormatGetPlanePixelFormat (VPIImageFormat fmt, int plane)
	Get the pixel format of image format's plane. More...

VPIImageFormat	vpiImageFormatGetPlaneFormat (VPIImageFormat fmt, int plane)
	Get the plane format of an image format. More...

VPIImageFormat	vpiMakeImageFormatFromPlanes (VPIImageFormat plane0, VPIImageFormat plane1, VPIImageFormat plane2, VPIImageFormat plane3)
	Constructs an image format given the format of each plane. More...

VPIPixelFormat	vpiMakePixelFormat (VPIMemLayout memLayout, VPIFormatDataType dataType, VPIFormatPacking packing)
	Creates an user-defined pixel format. More...

VPIFormatPacking	vpiPixelFormatGetPacking (VPIPixelFormat fmt)
	Get the packing of a pixel format. More...

int	vpiPixelFormatGetBitsPerPixel (VPIPixelFormat fmt)
	Get the number of bits per pixel of a pixel format. More...

VPIFormatDataType	vpiPixelFormatGetDataType (VPIPixelFormat fmt)
	Get the data type of a pixel format. More...

VPIMemLayout	vpiPixelFormatGetMemLayout (VPIPixelFormat fmt)
	Get the memory layout of a pixel format. More...

int	vpiPixelFormatGetChannelCount (VPIPixelFormat fmt)
	Get the number of channels of a pixel format. More...

Detailed Description

Utilities and definitions used when working with VPIImageFormat and VPIPixelFormat values.

Here the contituent parts of a VPIImageFormat and VPIPixelFormat are defined, along with functions to create and decompose them.

Macro Definition Documentation

◆ VPI_MAKE_IMAGE_FORMAT

#define VPI_MAKE_IMAGE_FORMAT	(	colorSpace,
		chromaSubsamp,
		memLayout,
		dataType,
		swizzle,
		numPlanes,
		packing0,
		packing1,
		...
	)

#include <vpi/ImageFormat.h>

Creates an user-defined image format constant using abbreviated parameters.

This macro allows passing abbreviated format parameters (without the parameter type prefix). Example to create a YUV422 ITU-R BT.709 block-linear format.

VPIImageFormat fmt = VPI_MAKE_IMAGE_FORMAT(YCbCr709, 422_COSITED, BL, UNSIGNED, XYZ0, 2, X8, X8_Y8);

Parameters

[in]	colorSpace	VPIColorSpace to be used, without the VPI_COLOR_SPACE_ prefix.
[in]	chromaSubsamp	VPIChromaSubsampling to be used, without the VPI_CHROMA_SUBSAMPLING_ prefix.
[in]	memLayout	VPIMemLayout to be used, without the VPI_MEM_LAYOUT_ prefix.
[in]	dataType	VPIFormatDataType to be used, without the VPI_FORMAT_DATA_TYPE_ prefix.
[in]	swizzle	VPISwizzle operation to be performed on the channels, without the VPI_SWIZZLE_ prefix.
[in]	numPlanes	Number of planes this format has.
[in]	packing0,packing1,...	Format packing of each plane, without the VPI_FORMAT_PACKING_ prefix.

Returns: The user-defined image format.

Definition at line 166 of file ImageFormat.h.

◆ VPI_MAKE_PIXEL_FORMAT

#define VPI_MAKE_PIXEL_FORMAT	(	memLayout,
		dataType,
		packing
	)

#include <vpi/PixelFormat.h>

Creates an user-defined pixel format constant using abbreviated parameters.

This macro allows passing abbreviated format parameters (without the parameter type prefix). Example to create a block-linear format two interleaved 32-bit floating point channels:

VPIPixelFormat fmt = VPI_MAKE_PIXEL_FORMAT(BL, FLOAT, X32_Y32);

Parameters

[in]	memLayout	VPIMemLayout to be used, without the VPI_MEM_LAYOUT_ prefix.
[in]	dataType	VPIFormatDataType to be used, without the VPI_FORMAT_DATA_TYPE_ prefix.
[in]	packing	Format packing used, which also defines the number of channels, without the VPI_FORMAT_PACKING_ prefix.

Returns: The user-defined pixel format.

Definition at line 118 of file PixelFormat.h.

◆ VPI_MAKE_SWIZZLE

#define VPI_MAKE_SWIZZLE	(	x,
		y,
		z,
		w
	)	((VPISwizzle)VPI_DETAIL_MAKE_SWIZZLE(x, y, z, w))

#include <vpi/Format.h>

Creates an user-defined swizzle operation.

This macro is used to create an user-defined swizzle operation if it's not predefined xxx. Parameters are specified literally by X, Y, Z, 0, 1. They correpond to VPISwizzleChannel values. Example:

VPISwizzle sw = VPI_MAKE_SWIZZLE(0,Y,Z,W);

Parameters

[in]	x	Channel that will correspond to the first component.
[in]	y	Channel that will correspond to the second component.
[in]	z	Channel that will correspond to the third component.
[in]	w	Channel that will correspond to the component component.

Returns: the user-defined VPISwizzle operation.

Definition at line 334 of file Format.h.

Enumeration Type Documentation

◆ VPIChromaSubsampling

enum VPIChromaSubsampling

#include <vpi/Format.h>

Defines how chroma-subsampling is done.

This is only applicable to image formats whose color model is YUV. Other image formats must use VPI_CHROMA_SUBSAMPLING_NONE. Chroma subsampling is defined by 4 parameters:

Horizontal resolution relative to luma resolution.
Vertical resolution relative to luma resolution.
Horizontal sample position relative to current (x) and next (x+1) luma sample.
Vertical sample position relative to current (y) and next (y+1) luma sample.

Enumerator
VPI_CHROMA_SUBSAMPLING_444	4:4:4 sub-sampling. Chroma has full horizontal and vertical resolution, meaning no chroma subsampling.
VPI_CHROMA_SUBSAMPLING_422_BT601	4:2:2 BT.601 sub-sampling. Chroma has half horizontal and full vertical resolutions with chroma sample cosited with luma.
VPI_CHROMA_SUBSAMPLING_422_COSITED	Alias to VPI_CHROMA_SUBSAMPLING_422_BT601.
VPI_CHROMA_SUBSAMPLING_420_COSITED	4:2:0 sub-sampling. Chroma has half horizontal and vertical resolutions with chroma sample cosited with luma.
VPI_CHROMA_SUBSAMPLING_420_MPEG1	4:2:0 MPEG1 sub-sampling. Chroma has half horizontal vertical resolutions with chroma sample centered horizontally and vertically.
VPI_CHROMA_SUBSAMPLING_420_MPEG2	4:2:0 MPEG2 sub-sampling. Chroma has half horizontal and vertical resolutions with chroma sample cosited horizontally and centered vertically.
VPI_CHROMA_SUBSAMPLING_NONE	Used when no chroma subsampling takes place, specially for color spaces without chroma components.

Definition at line 459 of file Format.h.

◆ VPIColorModel

enum VPIColorModel

#include <vpi/Format.h>

Defines color models.

A color model gives meaning to each channel of an image format. They are specified in a canonical XYZW ordering that can then be swizzled to the desired ordering.

Enumerator
VPI_COLOR_MODEL_OTHER	Other non-specified color model. Canonical channel ordering can be inferred from the color space.
VPI_COLOR_MODEL_RGBA	RGB + alpha color model.
VPI_COLOR_MODEL_YUV	YUV color model.
VPI_COLOR_MODEL_RAW	RAW color model, used for Bayer image formats.

Definition at line 340 of file Format.h.

◆ VPIColorSpace

enum VPIColorSpace

#include <vpi/Format.h>

Defines color spaces.

Color space specifies how physical colors are organized in a particular representation. Each VPIColorSpace is associated with a VPIColorModel.

Enumerator
VPI_COLOR_SPACE_NONCOLOR	Used in image formats that don't represent a color.
VPI_COLOR_SPACE_HSV	Hue, saturation, value color space.
VPI_COLOR_SPACE_HSL	Hue, saturation, lightness color space.
VPI_COLOR_SPACE_YIQ	Luma and I/Q chrominance color space.
VPI_COLOR_SPACE_LAB	CIE Lab* color space.
VPI_COLOR_SPACE_LUV	CIE Lub* color space.
VPI_COLOR_SPACE_XYZ	CIE XYZ tristimulus color space.
VPI_COLOR_SPACE_RGB	RGBA color space. Data in this color space is typically treated as linear, but in practice it contains gamma encoded data.
VPI_COLOR_SPACE_SensorRGB_LINEAR	This color space is used to represent camera sensor data after debayering. It isn't fully defined as specific characteristics of each camera sensor are needed to interpret the data. This is a linear color space, and should only be used with linear operations.
VPI_COLOR_SPACE_sRGB	sRGB color space with linear alpha.
VPI_COLOR_SPACE_RGB709	Non-linear RGB ITU-R BT.709 color space with linear alpha.
VPI_COLOR_SPACE_RGB709_LINEAR	Linear RGBA color space with ITU-R BT.709/sRGB gamut.
VPI_COLOR_SPACE_DISPLAYP3_LINEAR	Display-P3 color space (DCI-P3 primaries, linear encoding).
VPI_COLOR_SPACE_DISPLAYP3_NONLINEAR	Display-P3 color space (DCI-P3 primaries, sRGB encoding).
VPI_COLOR_SPACE_RGB2020	Non-linear RGB ITU-R BT.2020 color space with linear alpha.
VPI_COLOR_SPACE_RGB2020_LINEAR	Linear RGBA color space with ITU-R BT.2020 gamut.
VPI_COLOR_SPACE_RGB2020_PQ	Non-linear RGB color space encoded with SMPTE ST 2084 Perceptual Quantizer transfer function.
VPI_COLOR_SPACE_YCbCr601	YCbCr ITU-R BT.601 color space with standard range luma (16-235)
VPI_COLOR_SPACE_YCbCr601_ER	YCbCr ITU-R BT.601 color space with extended range luma (0-255)
VPI_COLOR_SPACE_YCbCr601_VC1	YCbCr ITU-R BT.601 color space with range reduced YCbCr for VC1 decoded planes. If picture layer of VC1 bit stream has RANGEREDFRM bit set, decoded YUV data has to be scaled up (range expanded). For this type of color space, clients should range expand the channels as follows: \begin{align} Y' &= \mathit{clip}(2(Y-128) + 128) \\ C_b' &= \mathit{clip}(2(C_b-128) + 128) \\ C_r' &= \mathit{clip}(2(C_r-128) + 128) \end{align}
VPI_COLOR_SPACE_YCbCr709	YCbCr ITU-R BT.709 color space with standard range luma (16-235).
VPI_COLOR_SPACE_YCbCr709_ER	YCbCr ITU-R BT.709 color space with extended range luma (0-255).
VPI_COLOR_SPACE_YCbCr2020	YCbCr ITU-R BT.2020 color space with standard range luma (16-235).
VPI_COLOR_SPACE_YCbCr2020_PQ_ER	YCbCr ITU-R BT.2020 color space encoded with SMPTE ST 2084 Perceptual Quantizer transfer function with extended range luma (0-255).
VPI_COLOR_SPACE_BayerRGGB	Bayer format with X channel mapped to samples as follows: span 1: R G R G R G R G span 2: G B G B G B G B (Y,Z,W are discarded)
VPI_COLOR_SPACE_BayerBGGR	Bayer format with X channel mapped to samples as follows: span 1: B G B G B G B G span 2: G R G R G R G R (Y,Z,W are discarded)
VPI_COLOR_SPACE_BayerGRBG	Bayer format with X channel mapped to samples as follows: span 1: G R G R G R G R span 2: B G B G B G B G (Y,Z,W are discarded)
VPI_COLOR_SPACE_BayerGBRG	Bayer format with X channel mapped to samples as follows: span 1: G B G B G B G B span 2: R G R G R G R G (Y,Z,W are discarded)
VPI_COLOR_SPACE_BayerCCCC	Bayer format with X channel mapped to samples as follows: span 1: C C C C C C C C span 2: C C C C C C C C (Y,Z,W are discarded)
VPI_COLOR_SPACE_BayerCRBC	Bayer format with X channel mapped to samples as follows: span 1: C R C R C R C R span 2: B C B C B C B C (Y,Z,W are discarded)

Definition at line 355 of file Format.h.

◆ VPIFormatDataType

enum VPIFormatDataType

#include <vpi/Format.h>

Defines the channel data type.

Enumerator
VPI_FORMAT_DATA_TYPE_UNSIGNED	Channels are unsigned integer values.
VPI_FORMAT_DATA_TYPE_SIGNED	Channels are signed integer values.
VPI_FORMAT_DATA_TYPE_FLOAT	Channel are floating point values.

Definition at line 210 of file Format.h.

◆ VPIFormatPacking

enum VPIFormatPacking

#include <vpi/Format.h>

Defines how channels are packed into an image plane element.

Packing encodes how many channels the plane element has, and how they are arranged in memory.

Up to 4 channels (denoted by X, Y, Z, W) can be packed into an image plane element, each one occupying a specified number of bits.

When two channels are specified one right after the other, they are ordered from most-significant bit to least-significant bit. Words are separated by underscores. For example:

X8Y8Z8W8 = a single 32-bit word containing 4 channels, 8 bits each.

In little-endian architectures:

     Address  0   ||  1   ||  2   ||  3
           WWWWWWWWZZZZZZZZYYYYYYYYXXXXXXXX

In big-endian architectures:

     Address  0   ||  1   ||  2   ||  3
           XXXXXXXXYYYYYYYYZZZZZZZZWWWWWWWW

X8_Y8_Z8_W8 = four consecutive 8-bit words, corresponding to 4 channels, 8 bits each.

In little-endian architectures:

     Address  0   ||  1   ||  2   ||  3
           XXXXXXXXYYYYYYYYZZZZZZZZWWWWWWWW

In big-endian architectures:

     Address  0   ||  1   ||  2   ||  3
           XXXXXXXXYYYYYYYYZZZZZZZZWWWWWWWW

In cases where a word is less than 8 bits (e.g., X1 1-bit channel), channels are ordered from LSB to MSB within a word.

Note: Also note equivalences such as the following:; In little-endian: X8_Y8_Z8_W8 = W8Z8Y8X8.; In big-endian: X8_Y8_Z8_W8 = X8Y8Z8W8.

Some formats allow different packings when pixels' horizontal coordinate is even or odd. For instance, every pixel of YUV422 packed format contains an Y channel, while only even pixels contain the U channel, and odd pixels contain V channel. Such formats use a double-underscore to separate the even pixels from the odd pixels. The packing just described might be referred to X8_Y8__X8_Z8, where X = luma, Y = U chroma, Z = V chroma.

Enumerator
VPI_FORMAT_PACKING_0	No channels.
VPI_FORMAT_PACKING_INVALID	Denotes an invalid packing.
VPI_FORMAT_PACKING_X1	One 1-bit channel.
VPI_FORMAT_PACKING_X2	One 2-bit channel.
VPI_FORMAT_PACKING_X4	One 4-bit channel.
VPI_FORMAT_PACKING_X8	One 8-bit channel.
VPI_FORMAT_PACKING_X4Y4	Two 4-bit channels in one word.
VPI_FORMAT_PACKING_X3Y3Z2	Three 3-, 3- and 2-bit channels in one 8-bit word.
VPI_FORMAT_PACKING_X16	One 16-bit channel.
VPI_FORMAT_PACKING_X8_Y8	Two 8-bit channels in two 8-bit words.
VPI_FORMAT_PACKING_X8Y8	Two 8-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X6Y10	Two 6- and 10-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X10Y6	Two 10- and 6-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X2Y14	Two 2- and 14-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X12Y4	Two 12- and 4-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X4Y12	Two 4- and 12-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X8_Y8__X8_Z8	2 pixels of 2 8-bit channels each, totalling 4 8-bit words.
VPI_FORMAT_PACKING_Y8_X8__Z8_X8	2 pixels of 2 swapped 8-bit channels each, totalling 4 8-bit words.
VPI_FORMAT_PACKING_X5Y5Z6	Three 5-, 5- and 6-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X5Y6Z5	Three 5-, 6- and 5-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X6Y5Z5	Three 6-, 5- and 5-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X1Y5Z5W5	Four 1-, 5-, 5- and 5-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X4Y4Z4W4	Four 4-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X5Y1Z5W5	Four 5-, 1-, 5- and 5-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X5Y5Z1W5	Four 5-, 5-, 1- and 5-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X5Y5Z5W1	Four 5-, 5-, 5- and 1-bit channels in one 16-bit word.
VPI_FORMAT_PACKING_X24	One 24-bit channel.
VPI_FORMAT_PACKING_X8_Y8_Z8	Three 8-bit channels in three 8-bit words.
VPI_FORMAT_PACKING_X32	One 32-bit channel.
VPI_FORMAT_PACKING_X16_Y16	Two 16-bit channels in one 16-bit words.
VPI_FORMAT_PACKING_X16Y16	Two 16-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X12Y20	Two 12- and 20-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X10Y11Z11	Three 10-, 11- and 11-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X11Y11Z10	Three 11-, 11- and 10-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X2Y10Z10W10	Four 2-, 10-, 10- and 10-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X8Y8Z8W8	Four 8-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X8_Y8_Z8_W8	Four 8-bit channels in four 8-bit words.
VPI_FORMAT_PACKING_X10Y6Z10W6	Four 10-, 6-, 10- and 6-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X10Y10Z10W2	Four 10-, 10-, 10- and 2-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X12Y4Z12W4	Four 12-, 4-, 12- and 4-bit channels in one 32-bit word.
VPI_FORMAT_PACKING_X48	One 48-bit channel.
VPI_FORMAT_PACKING_X16_Y16_Z16	Three 16-bit channels in three 16-bit words.
VPI_FORMAT_PACKING_X64	One 64-bit channel.
VPI_FORMAT_PACKING_X32_Y32	Two 32-bit channels in two 32-bit words.
VPI_FORMAT_PACKING_X16_Y16_Z16_W16	Four 16-bit channels in four 16-bit words.
VPI_FORMAT_PACKING_X16Y16Z16W16	Four 16-bit channels in one 64-bit word.
VPI_FORMAT_PACKING_X96	One 96-bit channel.
VPI_FORMAT_PACKING_X32_Y32_Z32	Three 32-bit channels in three 32-bit words.
VPI_FORMAT_PACKING_X128	One 128-bit channel.
VPI_FORMAT_PACKING_X32_Y32_Z32_W32	Four 32-bit channels in four 32-bit words.
VPI_FORMAT_PACKING_X32Y32Z32W32	Four 32-bit channels in one 128-bit word.

Definition at line 131 of file Format.h.

◆ VPIImageFormat

enum VPIImageFormat

#include <vpi/ImageFormat.h>

Pre-defined image formats.

An image format defines how image pixels are interpreted. Each image format is defined by the following components:

VPIColorSpace
VPIChromaSubsampling method (when applicable)
VPIFormatDataType
VPISwizzle
Number of planes
Format packing of each plane.

These pre-defined formats are guaranteed to work with algoriths that explicitly support them. Image formats can also be user-defined using vpiMakeImageFormat or vpiMakeImageFormatFromPlanes functions.

Using user-defined image formats with algorithms can lead to undefined behavior (segfaults, etc), but usually it works as expected. Result of algorithms using these image formats must be checked for correctness, as it's not guaranteed that they will work.

Enumerator
VPI_IMAGE_FORMAT_INVALID	Denotes an invalid image format.
VPI_IMAGE_FORMAT_U8	Single plane with one 8-bit unsigned integer channel. Also used for grayscale images.
VPI_IMAGE_FORMAT_S8	Single plane with one 8-bit signed integer channel. Also used for grayscale images.
VPI_IMAGE_FORMAT_U16	Single plane with one 16-bit unsigned integer channel. Also used for grayscale images.
VPI_IMAGE_FORMAT_S16	Single plane with one 16-bit signed integer channel. Also used for grayscale images.
VPI_IMAGE_FORMAT_F32	Single plane with one 32-bit floating point channel. Also used for grayscale images.
VPI_IMAGE_FORMAT_F64	Single plane with one 64-bit floating point channel. Also used for grayscale images.
VPI_IMAGE_FORMAT_2F32	Single plane with two interleaved 32-bit floating point channels.
VPI_IMAGE_FORMAT_NV12	YUV420sp 8-bit pitch-linear format composed of two planes: One 8-bit channel with luma (Y). Two interleaved 8-bit channels with chroma (U/V), with half horizontal and vertical resolution of luma plane.
VPI_IMAGE_FORMAT_NV12_BL	YUV420sp 8-bit block-linear format composed of two planes: One 8-bit channel with luma (Y). Two interleaved 8-bit channels with chroma (U/V), with half horizontal and vertical resolution of luma plane.
VPI_IMAGE_FORMAT_RGB8	Single plane with interleaved RGB 8-bit channel.
VPI_IMAGE_FORMAT_BGR8	Single plane with interleaved BGR 8-bit channel.
VPI_IMAGE_FORMAT_RGBA8	Single plane with interleaved RGBA 8-bit channel.
VPI_IMAGE_FORMAT_BGRA8	Single plane with interleaved BGRA 8-bit channel.

Definition at line 93 of file ImageFormat.h.

◆ VPIMemLayout

enum VPIMemLayout

#include <vpi/Format.h>

Defines how the 2D plane pixels are laid out in memory.

This defines how a pixel are addressed, i.e., given its \((x,y)\) coordinate, what's its memory address. Block-linear formats have a proprietary memory representation and aren't supposed to be addressed by the user directly.

Enumerator
VPI_MEM_LAYOUT_PITCH_LINEAR	Pixels are laid out in row-major order, i.e., \((x,y) = y \times \mathit{pitch} + x \times \mathit{pixel stride}\).
VPI_MEM_LAYOUT_BLOCK1_LINEAR	Pixels are laid out in block-linear format with height = 1.
VPI_MEM_LAYOUT_BLOCK2_LINEAR	Pixels are laid out in block-linear format with height = 2.
VPI_MEM_LAYOUT_BLOCK4_LINEAR	Pixels are laid out in block-linear format with height = 4.
VPI_MEM_LAYOUT_BLOCK8_LINEAR	Pixels are laid out in block-linear format with height = 8.
VPI_MEM_LAYOUT_BLOCK16_LINEAR	Pixels are laid out in block-linear format with height = 16.
VPI_MEM_LAYOUT_BLOCK32_LINEAR	Pixels are laid out in block-linear format with height = 32.
VPI_MEM_LAYOUT_BLOCK_LINEAR	Default block-linear format. It's guaranteed to be valid in all algorithms that support block-linear format.
VPI_MEM_LAYOUT_PL	Useful alias.
VPI_MEM_LAYOUT_BL	Useful alias.

Definition at line 223 of file Format.h.

◆ VPIPixelFormat

enum VPIPixelFormat

#include <vpi/PixelFormat.h>

Pre-defined pixel formats.

Pixel formats defines the geometry of pixels in a image plane without taking into account what the value represents. For example, a VPI_IMAGE_FORMAT_NV12 is composed of 2 planes, each one with the following pixel formats:

VPI_PIXEL_FORMAT_U8 representing pixels as 8-bit unsigned values.
VPI_PIXEL_FORMAT_2U8 representing pixels as two interleaved 32-bit floating-point values.

Enumerator
VPI_PIXEL_FORMAT_DEFAULT	Used to signal that the pixel format must be inferred from image format.
VPI_PIXEL_FORMAT_INVALID	Signal format conversion errors.
VPI_PIXEL_FORMAT_U8	One channel of unsigned 8-bit values.
VPI_PIXEL_FORMAT_S8	One channel of signed 8-bit values.
VPI_PIXEL_FORMAT_U16	One channel of unsigned 16-bit values.
VPI_PIXEL_FORMAT_S16	One channel of signed 16-bit values.
VPI_PIXEL_FORMAT_2U8	Two interleaved channels of unsigned 8-bit values.
VPI_PIXEL_FORMAT_3U8	Three interleaved channels of unsigned 8-bit values.
VPI_PIXEL_FORMAT_4U8	Four interleaved channels of unsigned 8-bit values.
VPI_PIXEL_FORMAT_F32	One channel of 32-bit float values.
VPI_PIXEL_FORMAT_2F32	Two interleaved channels of 32-bit float values.
VPI_PIXEL_FORMAT_F64	One channel of 64-bit float values.

Definition at line 79 of file PixelFormat.h.

◆ VPISwizzle

enum VPISwizzle

#include <vpi/Format.h>

Defines the supported channel swizzle operations.

The operations map an input vector \((x,y,z,w)\) into an output vector \((x',y',z',w')\). Any output channel can select any of the input channels, or the constants zero or one. For example, the swizzle "X000" selects the first channel, whereas swizzle "ZYXW" swaps the X and Z channels, needed for conversion between RGBA and BGRA image formats.

Enumerator
VPI_SWIZZLE_0000	Swizzle operation.
VPI_SWIZZLE_1000	Swizzle operation.
VPI_SWIZZLE_0001	Swizzle operation.
VPI_SWIZZLE_XYZW	Swizzle operation.
VPI_SWIZZLE_ZYXW	Swizzle operation.
VPI_SWIZZLE_WXYZ	Swizzle operation.
VPI_SWIZZLE_WZYX	Swizzle operation.
VPI_SWIZZLE_YZWX	Swizzle operation.
VPI_SWIZZLE_XYZ1	Swizzle operation.
VPI_SWIZZLE_XYZ0	Swizzle operation.
VPI_SWIZZLE_YZW1	Swizzle operation.
VPI_SWIZZLE_XXX1	Swizzle operation.
VPI_SWIZZLE_XZY1	Swizzle operation.
VPI_SWIZZLE_ZYX1	Swizzle operation.
VPI_SWIZZLE_ZYX0	Swizzle operation.
VPI_SWIZZLE_WZY1	Swizzle operation.
VPI_SWIZZLE_X000	Swizzle operation.
VPI_SWIZZLE_0X00	Swizzle operation.
VPI_SWIZZLE_00X0	Swizzle operation.
VPI_SWIZZLE_000X	Swizzle operation.
VPI_SWIZZLE_Y000	Swizzle operation.
VPI_SWIZZLE_0Y00	Swizzle operation.
VPI_SWIZZLE_00Y0	Swizzle operation.
VPI_SWIZZLE_000Y	Swizzle operation.
VPI_SWIZZLE_0XY0	Swizzle operation.
VPI_SWIZZLE_XXXY	Swizzle operation.
VPI_SWIZZLE_YYYX	Swizzle operation.
VPI_SWIZZLE_0YX0	Swizzle operation.
VPI_SWIZZLE_X00Y	Swizzle operation.
VPI_SWIZZLE_Y00X	Swizzle operation.
VPI_SWIZZLE_X001	Swizzle operation.
VPI_SWIZZLE_XY01	Swizzle operation.
VPI_SWIZZLE_XY00	Swizzle operation.
VPI_SWIZZLE_0XZ0	Swizzle operation.
VPI_SWIZZLE_0ZX0	Swizzle operation.
VPI_SWIZZLE_XZY0	Swizzle operation.
VPI_SWIZZLE_YZX1	Swizzle operation.
VPI_SWIZZLE_ZYW1	Swizzle operation.

Definition at line 275 of file Format.h.

◆ VPISwizzleChannel

enum VPISwizzleChannel

#include <vpi/Format.h>

Defines the format channel names.

The channels are color model-agnostic.

Enumerator
VPI_SWIZZLE_CHANNEL_0	Don't select a channel.
VPI_SWIZZLE_CHANNEL_X	Selects the first channel of the color model.
VPI_SWIZZLE_CHANNEL_Y	Selects the second channel of the color model.
VPI_SWIZZLE_CHANNEL_Z	Selects the third channel of the color model.
VPI_SWIZZLE_CHANNEL_W	Selects the fourth channel of the color model.
VPI_SWIZZLE_CHANNEL_1	Sets the corresponding channel to have its maximum value.

Definition at line 258 of file Format.h.

Function Documentation

◆ vpiFormatPackingGetBitsPerPixel()

int vpiFormatPackingGetBitsPerPixel ( VPIFormatPacking packing )

#include <vpi/Format.h>

Returns the number of bits per pixel of the given packing.

Parameters

[in] packing The format packing to be queried.

Returns: Total number of bits per pixel of the given packing. It's the sum of number of bits occupied by all packing channels.

◆ vpiFormatPackingGetChannelCount()

int vpiFormatPackingGetChannelCount ( VPIFormatPacking packing )

#include <vpi/Format.h>

Returns the number of channels defined by the given packing.

Parameters

[in] packing The format packing to be queried.

Returns: Number of channels from the given format packing. It's value between 0 and 4.

◆ vpiFormatPackingGetParams()

void vpiFormatPackingGetParams	(	VPIFormatPacking	packing,
		int *	fixed,
		int *	bitsX,
		int *	bitsY,
		int *	bitsZ,
		int *	bitsW
	)

#include <vpi/Format.h>

Returns channels' information from a format packing.

Parameters

[in]	packing	The format packing to be queried.
[out]	fixed	1 if each component is assigned to its own word, 0 if all components share the same word.
[out]	bitsX	Number of bits in the first channel. Value can't be 0.
[out]	bitsY	Number of bits in the second channel. Can be 0 if channel isn't used.
[out]	bitsZ	Number of bits in the second channel. Can be 0 if bitsY is 0 and this channel isn't used.
[out]	bitsW	Number of bits in the second channel. Can be 0 if bitsZ is 0 and this channel isn't used.

◆ vpiImageFormatGetChannelCount()

int vpiImageFormatGetChannelCount ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the total number of channels of an image format.

Parameters

[in] fmt Image format to be queried.

Returns: The sum of all channel counts in all planes.

◆ vpiImageFormatGetChromaSubsampling()

VPIChromaSubsampling vpiImageFormatGetChromaSubsampling ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's chroma subsampling type.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's chroma subsampling type.

◆ vpiImageFormatGetChromaSubsamplingCenterHoriz()

int vpiImageFormatGetChromaSubsamplingCenterHoriz ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's chroma horizontal sample position.

Parameters

[in] fmt Image format to be queried.

Return values

0	Chroma sample is horizontally aligned with the corresponding luma sample.
1	Chroma sample is between the current (x) and the next (x+1) luma sample.

◆ vpiImageFormatGetChromaSubsamplingCenterVert()

int vpiImageFormatGetChromaSubsamplingCenterVert ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's chroma vertical sample position.

Parameters

[in] fmt Image format to be queried.

Return values

0	Chroma sample is vertically aligned with the corresponding luma sample.
1	Chroma sample is between the current (y) and the next (y+1) luma sample.

◆ vpiImageFormatGetChromaSubsamplingDivHoriz()

int vpiImageFormatGetChromaSubsamplingDivHoriz ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's chroma subsampling horizontal downscale factor.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's chroma subsampling horizontal downscale factor.

◆ vpiImageFormatGetChromaSubsamplingDivVert()

int vpiImageFormatGetChromaSubsamplingDivVert ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's chroma subsampling vertical downscale factor.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's chroma subsampling vertical downscale factor.

◆ vpiImageFormatGetColorModel()

VPIColorModel vpiImageFormatGetColorModel ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's color model.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's color model.

◆ vpiImageFormatGetColorSpace()

VPIColorSpace vpiImageFormatGetColorSpace ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's color space.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's color space.

◆ vpiImageFormatGetDataType()

VPIFormatDataType vpiImageFormatGetDataType ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's data type.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's data type.

◆ vpiImageFormatGetMemLayout()

VPIMemLayout vpiImageFormatGetMemLayout ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's memory layout.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's memory layout.

◆ vpiImageFormatGetPlaneBitsPerPixel()

int vpiImageFormatGetPlaneBitsPerPixel	(	VPIImageFormat	fmt,
		int	plane
	)

#include <vpi/ImageFormat.h>

Get the image format's plane bits per pixel count.

Parameters

[in]	fmt	Image format to be queried.
[in]	plane	Which plane is to be queried. Valid values range from 0 (first plane) to 3 (fourth plane).

Returns: The number of bits per pixel the given format plane has.

◆ vpiImageFormatGetPlaneChannelCount()

int vpiImageFormatGetPlaneChannelCount	(	VPIImageFormat	fmt,
		int	plane
	)

#include <vpi/ImageFormat.h>

Get the number of channels in a plane of an image format.

Parameters

[in]	fmt	Image format to be queried.
[in]	plane	Plane to be queried. Valid values range from 0 (first) to 3 (fourth and last) plane.

Returns: Number of channels in the given plane.

◆ vpiImageFormatGetPlaneCount()

int vpiImageFormatGetPlaneCount ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the number of planes of an image format.

Parameters

[in] fmt Image format to be queried.

Returns: Number of planes defined by the given image format.

◆ vpiImageFormatGetPlaneFormat()

VPIImageFormat vpiImageFormatGetPlaneFormat	(	VPIImageFormat	fmt,
		int	plane
	)

#include <vpi/ImageFormat.h>

Get the plane format of an image format.

Parameters

[in]	fmt	Image format to be queried.
[in]	plane	Plane to be queried. Valid values range from 0 (first) to 3 (fourth and last) plane.

Returns: The image format of the given plane.

◆ vpiImageFormatGetPlanePacking()

VPIFormatPacking vpiImageFormatGetPlanePacking	(	VPIImageFormat	fmt,
		int	plane
	)

#include <vpi/ImageFormat.h>

Get the packing for a given plane of an image format.

Parameters

[in]	fmt	Image format to be queried.
[in]	plane	Which plane whose packing must be returned. Valid values range from 0 (first plane) to 3 (fourth plane).

Returns: The plane's format packing, or VPI_FORMAT_PACKING_INVALID if the format doesn't have the given plane.

◆ vpiImageFormatGetPlanePixelFormat()

VPIPixelFormat vpiImageFormatGetPlanePixelFormat	(	VPIImageFormat	fmt,
		int	plane
	)

#include <vpi/ImageFormat.h>

Get the pixel format of image format's plane.

Parameters

[in]	fmt	Image format to be queried.
[in]	plane	Plane to be queried. Valid values range from 0 (first) to 3 (fourth and last) plane.

Returns: The pixel format of the given plane.

◆ vpiImageFormatGetPlaneSwizzle()

VPISwizzle vpiImageFormatGetPlaneSwizzle	(	VPIImageFormat	fmt,
		int	plane
	)

#include <vpi/ImageFormat.h>

Get the swizzle operation of the given image format's plane.

Parameters

[in]	fmt	Image format to be queried.
[in]	plane	Plane to be queried. Valid values range from 0 (first) to 3 (fourth and last) plane.

Returns: The channel swizzle operation performed in the given plane.

◆ vpiImageFormatGetSwizzle()

VPISwizzle vpiImageFormatGetSwizzle ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the image format's channel swizzle operation.

Parameters

[in] fmt Image format to be queried.

Returns: The image format's swizzle operation.

◆ vpiImageFormatGetSwizzleW()

VPISwizzleChannel vpiImageFormatGetSwizzleW ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the color model channel that corresponds to the fourth component of given image format.

For example, given format VPI_IMAGE_FORMAT_BGRA8, its fourth component is alpha, which corresponds to the VPI_SWIZZLE_CHANNEL_W (fourth) channel of the color model VPI_COLOR_MODEL_RGBA.

Parameters

[in] fmt Image format to be queried.

Returns: The color model channel assigned to the fourth component.

◆ vpiImageFormatGetSwizzleX()

VPISwizzleChannel vpiImageFormatGetSwizzleX ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the color model channel that corresponds to the first component of given image format.

For example, given format VPI_IMAGE_FORMAT_BGRA8, its first component is blue, which corresponds to the VPI_SWIZZLE_CHANNEL_Z (third) channel of the color model VPI_COLOR_MODEL_RGBA.

Parameters

[in] fmt Image format to be queried.

Returns: The color model channel assigned to the first component.

◆ vpiImageFormatGetSwizzleY()

VPISwizzleChannel vpiImageFormatGetSwizzleY ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the color model channel that corresponds to the second component of given image format.

For example, given format VPI_IMAGE_FORMAT_BGRA8, its second component is green, which corresponds to the VPI_SWIZZLE_CHANNEL_Y (second) channel of the color model VPI_COLOR_MODEL_RGBA.

Parameters

[in] fmt Image format to be queried.

Returns: The color model channel assigned to the second component.

◆ vpiImageFormatGetSwizzleZ()

VPISwizzleChannel vpiImageFormatGetSwizzleZ ( VPIImageFormat fmt )

#include <vpi/ImageFormat.h>

Get the color model channel that corresponds to the third component of given image format.

For example, given format VPI_IMAGE_FORMAT_BGRA8, its third component is red, which corresponds to the VPI_SWIZZLE_CHANNEL_X (first) channel of the color model VPI_COLOR_MODEL_RGBA.

Parameters

[in] fmt Image format to be queried.

Returns: The color model channel assigned to the third component.

◆ vpiImageFormatSetChromaSubsampling()

VPIImageFormat vpiImageFormatSetChromaSubsampling	(	VPIImageFormat	fmt,
		VPIChromaSubsampling	css
	)

#include <vpi/ImageFormat.h>

Set the image format's chroma subsampling type.

Parameters

[in]	fmt	Image format have its chroma subsampling type replaced.
[in]	css	The new chroma subsampling type.

Returns: The new image format based on input's, but with the user-provided chroma subsampling type.

◆ vpiImageFormatSetColorSpace()

VPIImageFormat vpiImageFormatSetColorSpace	(	VPIImageFormat	fmt,
		VPIColorSpace	colorSpace
	)

#include <vpi/ImageFormat.h>

Set the image format's color space.

Parameters

[in]	fmt	Image format have its color space replaced.
[in]	colorSpace	The new color space.

Returns: The new image format based on input's, but with the user-provided color space.

◆ vpiImageFormatSetDataType()

VPIImageFormat vpiImageFormatSetDataType	(	VPIImageFormat	fmt,
		VPIFormatDataType	dataType
	)

#include <vpi/ImageFormat.h>

Set the image format's data type.

Parameters

[in]	fmt	Image format have its data type replaced.
[in]	dataType	The new data type.

Returns: The new image format based on input's, but with the user-provided data type.

◆ vpiImageFormatSetMemLayout()

VPIImageFormat vpiImageFormatSetMemLayout	(	VPIImageFormat	fmt,
		VPIMemLayout	memLayout
	)

#include <vpi/ImageFormat.h>

Set the image format's memory layout.

Parameters

[in]	fmt	Image format have its memory layout replaced.
[in]	memLayout	The new memory layout.

Returns: The new image format based on input's, but with the user-provided memory layout.

◆ vpiMakeChromaSubsampling()

VPIChromaSubsampling vpiMakeChromaSubsampling	(	int	divHoriz,
		int	divVert,
		int	centerSampleHoriz,
		int	centerSampleVert
	)

#include <vpi/Format.h>

Creates an user-defined VPIChromaSubsampling definition.

Parameters

[in]	divHoriz	How many horizontal luma samples correspond to each chroma component. Must be 1 or 2.
[in]	divVert	How many vertical luma samples correspond to each chroma component. Must be 1 or 2.
[in]	centerSampleHoriz	If 0, chroma sample is aligned horizontally with corresponding luma sample. If 1, it's in the middle horizontally between its corresponding luma sample (x) and the next (x+1).
[in]	centerSampleVert	If 0, chroma sample is aligned vertically with corresponding luma sample. If 1, it's in the middle vertically between its corresponding luma sample (y) and the next (y+1).

Returns: Chroma subsampling value corresponding to the parameters given.

◆ vpiMakeFormatPacking()

VPIFormatPacking vpiMakeFormatPacking	(	int	fixed,
		int	bitsX,
		int	bitsY,
		int	bitsZ,
		int	bitsW
	)

#include <vpi/Format.h>

Returns a pre-defined VPIFormatPacking given the number of bits in each channel.

This function calculates the VPIFormatPacking based on the channel characteristics at run time.

Parameters

[in]	fixed	1 if each component is assigned to its own word, 0 if all components share the same word.
[in]	bitsX	Number of bits in the first channel. Value can't be 0.
[in]	bitsY	Number of bits in the second channel. Can be 0 if channel isn't used.
[in]	bitsZ	Number of bits in the second channel. Can be 0 if bitsY is 0 and this channel isn't used.
[in]	bitsW	Number of bits in the second channel. Can be 0 if bitsZ is 0 and this channel isn't used.

Returns: the corresponding VPIFormatPacking, or VPI_FORMAT_PACKING_INVALID if packing isn't defined.

◆ vpiMakeImageFormat()

VPIImageFormat vpiMakeImageFormat	(	VPIColorSpace	colorSpace,
		VPIChromaSubsampling	chromaSub,
		VPIMemLayout	memLayout,
		VPIFormatDataType	dataType,
		VPISwizzle	swizzle,
		VPIFormatPacking	packing0,
		VPIFormatPacking	packing1,
		VPIFormatPacking	packing2,
		VPIFormatPacking	packing3
	)

#include <vpi/ImageFormat.h>

Creates an user-defined image format.

When the pre-defined image formats aren't enough, user-defined image formats can be created.

Warning: It's not guaranteed that algorithms will work correctly with use-defined image formats. It's recommended to check if the results are correct prior deploying the solution in a production environment.

Parameters

[in]	colorSpace	VPIColorSpace to be used.
[in]	chromaSub	VPIChromaSubsampling to be used.
[in]	memLayout	VPIMemLayout to be used.
[in]	dataType	VPIFormatDataType to be used.
[in]	swizzle	VPISwizzle operation to be performed on the channels.
[in]	packing0,packing1,packing2,packing3	Format packing of each plane. When remaining planes aren't needed, pass VPI_FORMAT_PACKING_INVALID for them.

Returns: The user-defined image format.

◆ vpiMakeImageFormatFromPlanes()

VPIImageFormat vpiMakeImageFormatFromPlanes	(	VPIImageFormat	plane0,
		VPIImageFormat	plane1,
		VPIImageFormat	plane2,
		VPIImageFormat	plane3
	)

#include <vpi/ImageFormat.h>

Constructs an image format given the format of each plane.

Parameters

[in] plane0,plane1,plane2,plane3 Image format of each plane. When plane doesn't exist, pass VPI_IMAGE_FORMAT_INVALID.

Returns: The image format whose planes has the given formats.

◆ vpiMakePixelFormat()

VPIPixelFormat vpiMakePixelFormat	(	VPIMemLayout	memLayout,
		VPIFormatDataType	dataType,
		VPIFormatPacking	packing
	)

#include <vpi/PixelFormat.h>

Creates an user-defined pixel format.

When the pre-defined pixel formats aren't enough, user-defined formats can be created.

Parameters

[in]	memLayout	VPIMemLayout to be used.
[in]	dataType	VPIFormatDataType to be used.
[in]	packing	Format packing used, which also defines the number of channels.

Returns: The user-defined pixel format.

◆ vpiMakeSwizzle()

VPISwizzle vpiMakeSwizzle	(	VPISwizzleChannel	x,
		VPISwizzleChannel	y,
		VPISwizzleChannel	z,
		VPISwizzleChannel	w
	)

#include <vpi/Format.h>

Creates an user-defined VPISwizzle operation.

This is similar to VPI_MAKE_SWIZZLE, but accepts the swizzle channels as runtime variables.

Parameters

[in]	x	Channel that will correspond to the first component.
[in]	y	Channel that will correspond to the second component.
[in]	z	Channel that will correspond to the third component.
[in]	w	Channel that will correspond to the component component.

Returns: the user-defined VPISwizzle operation.

◆ vpiPixelFormatGetBitsPerPixel()

int vpiPixelFormatGetBitsPerPixel ( VPIPixelFormat fmt )

#include <vpi/PixelFormat.h>

Get the number of bits per pixel of a pixel format.

Parameters

[in] fmt Pixel format to be queried.

Returns: The number of bits per pixel.

◆ vpiPixelFormatGetChannelCount()

int vpiPixelFormatGetChannelCount ( VPIPixelFormat fmt )

#include <vpi/PixelFormat.h>

Get the number of channels of a pixel format.

Parameters

[in] fmt Pixel format to be queried.

Returns: The number of channels of the pixel format.

◆ vpiPixelFormatGetDataType()

VPIFormatDataType vpiPixelFormatGetDataType ( VPIPixelFormat fmt )

#include <vpi/PixelFormat.h>

Get the data type of a pixel format.

Parameters

[in] fmt Pixel format to be queried.

Returns: The data type of the pixel format.

◆ vpiPixelFormatGetMemLayout()

VPIMemLayout vpiPixelFormatGetMemLayout ( VPIPixelFormat fmt )

#include <vpi/PixelFormat.h>

Get the memory layout of a pixel format.

Parameters

[in] fmt Pixel format to be queried.

Returns: The memory layout of the pixel format.

◆ vpiPixelFormatGetPacking()

VPIFormatPacking vpiPixelFormatGetPacking ( VPIPixelFormat fmt )

#include <vpi/PixelFormat.h>

Get the packing of a pixel format.

Parameters

[in] fmt Pixeel format to be queried.

Returns: The format's packing.

VPI - Vision Programming Interface

0.4.4 Release

Macros

Enumerations

Functions

Detailed Description

Macro Definition Documentation

◆ VPI_MAKE_IMAGE_FORMAT

◆ VPI_MAKE_PIXEL_FORMAT

◆ VPI_MAKE_SWIZZLE

Enumeration Type Documentation

◆ VPIChromaSubsampling

◆ VPIColorModel

◆ VPIColorSpace

◆ VPIFormatDataType

◆ VPIFormatPacking

◆ VPIImageFormat

◆ VPIMemLayout

◆ VPIPixelFormat

◆ VPISwizzle

◆ VPISwizzleChannel

Function Documentation

◆ vpiFormatPackingGetBitsPerPixel()

◆ vpiFormatPackingGetChannelCount()

◆ vpiFormatPackingGetParams()

◆ vpiImageFormatGetChannelCount()

◆ vpiImageFormatGetChromaSubsampling()

◆ vpiImageFormatGetChromaSubsamplingCenterHoriz()

◆ vpiImageFormatGetChromaSubsamplingCenterVert()

◆ vpiImageFormatGetChromaSubsamplingDivHoriz()

◆ vpiImageFormatGetChromaSubsamplingDivVert()

◆ vpiImageFormatGetColorModel()

◆ vpiImageFormatGetColorSpace()

◆ vpiImageFormatGetDataType()

◆ vpiImageFormatGetMemLayout()

◆ vpiImageFormatGetPlaneBitsPerPixel()

◆ vpiImageFormatGetPlaneChannelCount()

◆ vpiImageFormatGetPlaneCount()

◆ vpiImageFormatGetPlaneFormat()

◆ vpiImageFormatGetPlanePacking()

◆ vpiImageFormatGetPlanePixelFormat()

◆ vpiImageFormatGetPlaneSwizzle()

◆ vpiImageFormatGetSwizzle()

◆ vpiImageFormatGetSwizzleW()

◆ vpiImageFormatGetSwizzleX()

◆ vpiImageFormatGetSwizzleY()

◆ vpiImageFormatGetSwizzleZ()

◆ vpiImageFormatSetChromaSubsampling()

◆ vpiImageFormatSetColorSpace()

◆ vpiImageFormatSetDataType()

◆ vpiImageFormatSetMemLayout()

◆ vpiMakeChromaSubsampling()

◆ vpiMakeFormatPacking()

◆ vpiMakeImageFormat()

◆ vpiMakeImageFormatFromPlanes()

◆ vpiMakePixelFormat()

◆ vpiMakeSwizzle()

◆ vpiPixelFormatGetBitsPerPixel()

◆ vpiPixelFormatGetChannelCount()

◆ vpiPixelFormatGetDataType()

◆ vpiPixelFormatGetMemLayout()

◆ vpiPixelFormatGetPacking()