ffmpeg | branch: master | Martin Storsjö | Mon Aug 8
11:33:26 2022 +0300| [1eaa575cf11c054b0b724480f3070fc908faf8ef] | committer:
Martin Storsjö
tools: Make sure to create the tools directory before building decode_simple.o
This directory dependency is normally added implicitly by rules
in
ffmpeg | branch: master | Martin Storsjö | Tue Jul 12
23:58:01 2022 +0300| [d69d12a5b9236b9d2f1fd247ea452f84cdd1aaf9] | committer:
Martin Storsjö
checkasm: motion: Test different h parameters
Previously, the checkasm test always passed h=8, so no other cases
were tested.
Out of the me_cmp
ffmpeg | branch: master | Martin Storsjö | Wed Jul 13
23:39:40 2022 +0300| [dc55e6357818e21e26afbcdfdf2dd7368ec0e8fd] | committer:
Martin Storsjö
x86: Don't hardcode the height to 8 in sad8_xy2_mmx
The height is hardcoded in some of the me_cmp functions, but not
in all of them. But in the
ffmpeg | branch: master | Martin Storsjö | Tue Aug 16
23:46:35 2022 +0300| [21c2c57ba5a8b426ad9c07902ec957dbbb20cf36] | committer:
Martin Storsjö
checkasm: Provide enough alignment in the new yuv2plane1 test
This fixes the checkasm test in some setups on x86.
Signed-off-by: Martin Storsjö
ffmpeg | branch: master | J. Dekker | Thu Jun 23 20:04:05
2022 +0200| [ea6ecb12aa9ebfbc985f71938a6cccf5046ca826] | committer: J. Dekker
checkasm/hevc_add_res: add 12bit test
Also fix the bug where in every other byte only the lower 2 bits were
used in the 8bit test.
Signed-off-by: J. Dekker
ffmpeg | branch: master | J. Dekker | Thu Jun 23 20:04:06
2022 +0200| [aa9eabb7a5283fd90b3274ac4b6ba0d16e42] | committer: J. Dekker
lavc/aarch64: reformat add_res funcs
Signed-off-by: J. Dekker
>
ffmpeg | branch: master | Swinney, Jonathan | Sat Aug 13
20:56:06 2022 +| [0d7caa5b09b680aa2c8ae677a8d0fdc134b7e658] | committer:
Martin Storsjö
swscale/aarch64: add vscale specializations
This commit adds new code paths for vscale when filterSize is 2, 4, or
8. By using specialized code
ffmpeg | branch: master | Swinney, Jonathan | Sat Aug 13
20:55:55 2022 +| [4dcd191a507b76003bc8a1d0d845d99e3ba9fabf] | committer:
Martin Storsjö
checkasm: updated tests for sw_scale
Change the reference to exactly match the C reference in swscale,
instead of exactly matching the x86 SIMD
ffmpeg | branch: master | Timo Rothenpieler | Tue Aug
16 12:31:03 2022 +0200| [317f5252c09d6deee7025907eea2dfe44935c891] | committer:
Timo Rothenpieler
doc/APIchanges: add missing rgbaf16 pixfmt entry
>
ffmpeg | branch: master | Anton Khirnov | Sat Aug 13
17:03:39 2022 +0200| [ab3147383006f1a31978efce23c6bad38a754e05] | committer:
Anton Khirnov
fftools/ffmpeg: store a separate copy of input codec parameters
Use it instead of AVStream.codecpar in the main thread. While
AVStream.codecpar is
ffmpeg | branch: master | Swinney, Jonathan | Sat Aug 13
20:48:30 2022 +| [75ffca7eef557bcc714d924048a6e184b39fa470] | committer:
Martin Storsjö
libswscale/aarch64: add another hscale specialization
This specialization handles the case where filtersize is 4 mod 8, e.g.
12, 20, etc.
11 matches
Mail list logo