1*4882a593Smuzhiyun.. SPDX-License-Identifier: GFDL-1.1-no-invariants-or-later 2*4882a593Smuzhiyun 3*4882a593Smuzhiyun.. _V4L2-PIX-FMT-NV12MT: 4*4882a593Smuzhiyun 5*4882a593Smuzhiyun**************************** 6*4882a593SmuzhiyunV4L2_PIX_FMT_NV12MT ('TM12') 7*4882a593Smuzhiyun**************************** 8*4882a593Smuzhiyun 9*4882a593SmuzhiyunFormats with ½ horizontal and vertical chroma resolution. This format 10*4882a593Smuzhiyunhas two planes - one for luminance and one for chrominance. Chroma 11*4882a593Smuzhiyunsamples are interleaved. The difference to ``V4L2_PIX_FMT_NV12`` is the 12*4882a593Smuzhiyunmemory layout. Pixels are grouped in macroblocks of 64x32 size. The 13*4882a593Smuzhiyunorder of macroblocks in memory is also not standard. 14*4882a593Smuzhiyun 15*4882a593Smuzhiyun 16*4882a593SmuzhiyunDescription 17*4882a593Smuzhiyun=========== 18*4882a593Smuzhiyun 19*4882a593SmuzhiyunThis is the two-plane versions of the YUV 4:2:0 format where data is 20*4882a593Smuzhiyungrouped into 64x32 macroblocks. The three components are separated into 21*4882a593Smuzhiyuntwo sub-images or planes. The Y plane has one byte per pixel and pixels 22*4882a593Smuzhiyunare grouped into 64x32 macroblocks. The CbCr plane has the same width, 23*4882a593Smuzhiyunin bytes, as the Y plane (and the image), but is half as tall in pixels. 24*4882a593SmuzhiyunThe chroma plane is also grouped into 64x32 macroblocks. 25*4882a593Smuzhiyun 26*4882a593SmuzhiyunWidth of the buffer has to be aligned to the multiple of 128, and height 27*4882a593Smuzhiyunalignment is 32. Every four adjacent buffers - two horizontally and two 28*4882a593Smuzhiyunvertically are grouped together and are located in memory in Z or 29*4882a593Smuzhiyunflipped Z order. 30*4882a593Smuzhiyun 31*4882a593SmuzhiyunLayout of macroblocks in memory is presented in the following figure. 32*4882a593Smuzhiyun 33*4882a593Smuzhiyun 34*4882a593Smuzhiyun.. _nv12mt: 35*4882a593Smuzhiyun 36*4882a593Smuzhiyun.. kernel-figure:: nv12mt.svg 37*4882a593Smuzhiyun :alt: nv12mt.svg 38*4882a593Smuzhiyun :align: center 39*4882a593Smuzhiyun 40*4882a593Smuzhiyun V4L2_PIX_FMT_NV12MT macroblock Z shape memory layout 41*4882a593Smuzhiyun 42*4882a593SmuzhiyunThe requirement that width is multiple of 128 is implemented because, 43*4882a593Smuzhiyunthe Z shape cannot be cut in half horizontally. In case the vertical 44*4882a593Smuzhiyunresolution of macroblocks is odd then the last row of macroblocks is 45*4882a593Smuzhiyunarranged in a linear order. 46*4882a593Smuzhiyun 47*4882a593SmuzhiyunIn case of chroma the layout is identical. Cb and Cr samples are 48*4882a593Smuzhiyuninterleaved. Height of the buffer is aligned to 32. 49*4882a593Smuzhiyun 50*4882a593Smuzhiyun 51*4882a593Smuzhiyun.. _nv12mt_ex: 52*4882a593Smuzhiyun 53*4882a593Smuzhiyun.. kernel-figure:: nv12mt_example.svg 54*4882a593Smuzhiyun :alt: nv12mt_example.svg 55*4882a593Smuzhiyun :align: center 56*4882a593Smuzhiyun 57*4882a593Smuzhiyun Example V4L2_PIX_FMT_NV12MT memory layout of macroblocks 58*4882a593Smuzhiyun 59*4882a593SmuzhiyunMemory layout of macroblocks of ``V4L2_PIX_FMT_NV12MT`` format in most 60*4882a593Smuzhiyunextreme case. 61
