blob: 44b4e29c101d6671a083726b9015cc2b4a1a1cf9 [file] [log] [blame]
Name
IMG_yuv_image
Name Strings
cl_img_yuv_image
Contributors
James Laverack
Rob Quill
Contacts
James Laverack, Imagination Technoligies (james.laverack 'at' imgtec.com)
Rob Quill, Imagination Technoligies (robert.quill 'at' imgtec.com)
Status
Approved.
Version
Revision: #2, August 8th, 2016
Number
OpenCL Extension #46
Dependencies
Requires OpenCL version 1.2 or later.
This extension is written against the wording of the OpenCL 1.2
Specification.
Overview
This extension provides new image formats to read YUV images, such
as those provided by cameras on mobile devices.
New Tokens
New image channel order enums:
CL_NV21_IMG 0x40D0
CL_YV12_IMG 0x40D1
Additions to Chapter 5 of the OpenCL 1.2 Specification (Image Objects)
In Section 5.3.1, page 93, append the following to the list of error
conditions for clCreateImage:
"CL_INVALID_IMAGE_SIZE if the image_channel_order is CL_NV21_IMG or CL_YV12_IMG
and image dimensions specified in image_desc are not supported for for
all devices in context."
"CL_INVALID_VALUE if the image_channel_order is CL_NV21_IMG or CL_YV12_IMG and
the image_desc is not CL_MEM_OBJECT_IMAGE2D."
In Section 5.3.1.1, page 94, add the following to Table 5.6:
"CL_NV21_IMG, CL_YV12_IMG. This format can only be used if channel
data type = CL_UNORM_INT8"
In Section 5.3.3, page 102, append the following to the list of error
conditions for clEnqueueReadImage and clEnququWriteImage:
"CL_INVALID_VALUE if image_channel_order is CL_NV21_IMG or CL_YV12_IMG and the region
is not the full size of the image and/or the origin is not (0, 0)."
In Section 5.3.4, page 108, append the following to the list of error
conditions for clEnqueueFillImage:
"CL_IMAGE_FORMAT_NOT_SUPPORTED if image_channel_order is CL_NV21_IMG or
CL_YV12_IMG."
In Section 5.3.6, page 116, append the following to the list of error
conditions for clEnqueueMapImage:
"CL_INVALID_VALUE if image_channel_order is CL_NV21_IMG or CL_YV12_IMG and the region
is not the full size of the image and/or the origin is not (0, 0)."
Additions to Chapter 8 of the OpenCL 1.2 Specification (Addressing and Filter Modes)
In Section 8.2, page 335, replace the first paragraph of the Note with the following
text:
"If the sampler is specified as using unnormalized coordinates (floating-point or integer
coordinates), filter mode set to CLK_FILTER_NEAREST and addressing mode set to one of the
following modes - CLK_ADDRESS_NONE, CLK_ADDRESS_CLAMP_TO_EDGE or
CLK_ADDRESS_CLAMP, and image_channel_order is not CL_NV21_IMG or CL_YV12_IMG then the location
of the image element in the image given by (i, j, k) in section 8.2 will be computed
without any loss of precision."
Revision History
#1 - Initial revision.
#2 - Replace CL_NV21 and CL_YV12 with CL_NV21_IMG and CL_YV12_IMG respectively.