Skip to content

L
libyuv
Switch branch/tag
  1. 10 Jan, 2018 1 commit
  3. 09 Jan, 2018 1 commit
  5. 08 Jan, 2018 2 commits
  7. 07 Jan, 2018 1 commit
    • Frank Barchard's avatar
      H010ToAR30 optimized to 2 step conversion · 9d2cd6a3
      Frank Barchard authored 
      Previously H010ToAR30 was done in a 3 step conversion:
H010ToH420, H420ToARGB, ARGBToAR30.
This CL merges the first 2 steps into H010ToARGB, to
improve performance.
Caveat - only 10 bit YUV is supported at this time.
Previously the low level code supported different numbers
of bits - 9, 10, 12 or 16.

Was 3 step conversion:
LibYUVConvertTest.H010ToAR30_Any (1263 ms)
LibYUVConvertTest.H010ToAR30_Unaligned (951 ms)
LibYUVConvertTest.H010ToAR30_Invert (913 ms)
LibYUVConvertTest.H010ToAR30_Opt (901 ms)

Now 2 step conversion:
LibYUVConvertTest.H010ToAR30_Any (853 ms)
LibYUVConvertTest.H010ToAR30_Unaligned (811 ms)
LibYUVConvertTest.H010ToAR30_Invert (781 ms)
LibYUVConvertTest.H010ToAR30_Opt (755 ms)

Bug: libyuv:751
Test: LibYUVConvertTest.H010ToAR30_Opt
Change-Id: Ica7574040401cd57145a4827acdf3c0e58346a2a
Reviewed-on: https://chromium-review.googlesource.com/853288Reviewed-by: 's avatarFrank Barchard <fbarchard@chromium.org>
Reviewed-by: 's avatarMiguel Casas <mcasas@chromium.org>
      9d2cd6a3
  9. 05 Jan, 2018 2 commits
  11. 04 Jan, 2018 1 commit
  13. 03 Jan, 2018 1 commit
  15. 02 Jan, 2018 4 commits
  17. 28 Dec, 2017 1 commit
  19. 21 Dec, 2017 2 commits
    • Frank Barchard's avatar
      HalfFloat_SSE2 use movd from memory · c67db605
      Frank Barchard authored 
      pshufd requires 16 byte aligned memory or a register.
Use movd to a register to avoid a segfault if memory for float
is misaligned

Bug: libyuv:759
Test: 32 bit build of LibYUVPlanarTest.TestHalfFloatPlane_16bit_denormal
Change-Id: I6fdcc4317453af5acd4700f9d46425bb2f4a205b
Reviewed-on: https://chromium-review.googlesource.com/840459Reviewed-by: 's avatarMiguel Casas <mcasas@chromium.org>
Reviewed-by: 's avatarFrank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
      c67db605
    • VladimirTechMan's avatar
      Allow enabling libyuv symbols visibility for shared library targets · 56eb5a47
      VladimirTechMan authored 
      When developing WebRTC applications with custom video sources, libyuv's
API is often the required part of the project. For that, application
developers can link with a separate standalone instance of libyuv.
However, it is even better to avoid that binary code duplication and
link against libyuv as a part compiled into the WebRTC library.

When building WebRTC as a static library target, the symbols from libyuv
are normally accessible to the linker without any extra actions. When
building WebRTC as a shared library, that does not work as the exported
symbols are those marked as visible, while libyuv is built with the
hidden visibility setting by default.

This patch adds an extra flag to enable switching the symbol visibility
to default when building shared library targets depending on libyuv in
WebRTC and Chromium projects. By default the flag is not enabled,
preserving the prior behaviour.

Bug: NONE
Change-Id: I48893cb9f54b1e1b49a34e14434e67f91d3e5e79
Reviewed-on: https://chromium-review.googlesource.com/827746Reviewed-by: 's avatarFrank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
      56eb5a47
  21. 20 Dec, 2017 1 commit
  23. 19 Dec, 2017 1 commit
  25. 15 Dec, 2017 1 commit
  27. 14 Dec, 2017 1 commit
  29. 12 Dec, 2017 4 commits
    • Frank Barchard's avatar
      Add I420ToAR30 10 bit RGB · bb3180ae
      Frank Barchard authored 
      For more complete support of AR30 format, add I420ToAR30 allowing
the new RGB 10 bit format to be used from standard 8 bit I420 format.

Bug: libyuv:751
Test: I420ToAR30 unittest added
Change-Id: Ia8b0857447408bd6adab485158ce5f38d6dc2faa
Reviewed-on: https://chromium-review.googlesource.com/823243Reviewed-by: 's avatarWeiyong Yao <braveyao@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
      bb3180ae
    • Frank Barchard's avatar
      ARGBToAR30 SSSE3 use pmulhuw to replicate fields · c3677514
      Frank Barchard authored 
      AR30 is optimized with 3 techniques
1. pmulhuw is used to replicate 8 bits to 10 bits.
2. Two channels are processed at a time.  R and B, and A and G.
3. pshufb is used to shift and mask 2 channels of R and B

Bug: libyuv:751
Test: ARGBToAR30_Opt
Change-Id: I4e62d6caa4df7d0ae80395fa911d3c922b6b897b
Reviewed-on: https://chromium-review.googlesource.com/822520Reviewed-by: 's avatarrichard winterton <rrwinterton@gmail.com>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
      c3677514
    • Mirko Bonadei's avatar
      Using all_dependent_configs to pass libyuv_config around. · d94a4867
      Mirko Bonadei authored 
      Using public_configs, client projects must rely on public_deps to
propagate configurations up in the build graph. This is bad because
public_deps allows the exposition of headers that live in another
target. This can lead to a really unhealthy build.

On the other side, all_dependent_configs is automatically propagated
up in the build graph but if a target includes a libyuv header it is
forced by GN to declare the dependency (and this will propagate
libyuv_config).

Bug: webrtc:8605, webrtc:8603
Change-Id: I4d71bb5de0b5b62a4ec110349223614f0b98e655
No-Try: True
Reviewed-on: https://chromium-review.googlesource.com/822112
Commit-Queue: Mirko Bonadei <mbonadei@chromium.org>
Reviewed-by: 's avatarPatrik Höglund <phoglund@chromium.org>
      d94a4867
    • Frank Barchard's avatar
      ARGBToAR30 use vpmulhuw to replicate fields · 11dd1b95
      Frank Barchard authored 
      AR30 is optimized with 3 techniques
1. vpmulhuw is used to replicate 8 bits to 10 bits.
2. Two channels are processed at a time.  R and B, and A and G.
3. vpshufb is used to shift and mask 2 channels of R and B

Red Blue
With the 8 bit value in the upper bits, vpmulhuw by (1024+4) will produce a 10
bit value in the low 10 bits of each 16 bit value. This is whats wanted for the
blue channel. The red needs to be shifted 4 left, so multiply by (1024+4)*16 for
red.

Alpha Green
Alpha and Green are already in the high bits so vpand can zero out the other
bits, keeping just 2 upper bits of alpha and 8 bit green. The same multiplier
could be used for Green - (1024+4) putting the 10 bit green in the lsb.  Alpha
would be a simple multiplier to shift it into position.  It wants a gap of 10
above the green.  Green is 10 bits, so there are 6 bits in the low short.  4
more are needed, so a multiplier of 4 gets the 2 bits into the upper 16 bits,
and then a shift of 4 is a multiply of 16, so (4*16) = 64.  Then shift the
result left 10 to position the A and G channels.

Bug: libyuv:751
Test: ARGBToAR30_Opt
Change-Id: Ie4f20dce18203bae7b75acb1fd5232db8a8a4f11
Reviewed-on: https://chromium-review.googlesource.com/820046
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: 's avatarCheng Wang <wangcheng@google.com>
      11dd1b95
  31. 09 Dec, 2017 1 commit
  33. 08 Dec, 2017 1 commit
  35. 07 Dec, 2017 1 commit
  37. 05 Dec, 2017 1 commit
  39. 30 Nov, 2017 1 commit
  41. 28 Nov, 2017 1 commit
    • Frank Barchard's avatar
      Convert16To8Row_SSSE3 port from AVX2 · 324fa327
      Frank Barchard authored 
      H010ToAR30 uses Convert16To8Row_SSSE3 to convert 10 bit YUV to 8 bit.
Then standard YUV conversion can be used.  This improves performance
on low end CPUs.
Future CL will by pass this conversion allowing for 10 bit YUV source,
but the function will be useful as a utility for YUV conversions.

Bug: libyuv:559, libyuv:751
Test: out/Release/libyuv_unittest --gtest_filter=*H010ToAR30* --libyuv_width=1280 --libyuv_height=720 --libyuv_repeat=999 --libyuv_flags=-1 --libyuv_cpu_info=-1
Change-Id: I9b3ef22d88a5fd861de4cf1900b4c6e8fd24d0af
Reviewed-on: https://chromium-review.googlesource.com/792334
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: 's avatarFrank Barchard <fbarchard@chromium.org>
      324fa327
  43. 27 Nov, 2017 1 commit
  45. 22 Nov, 2017 1 commit
  47. 21 Nov, 2017 1 commit
  49. 20 Nov, 2017 3 commits
  51. 17 Nov, 2017 1 commit
  53. 16 Nov, 2017 1 commit
  55. 15 Nov, 2017 1 commit
  57. 10 Nov, 2017 1 commit
    • Frank Barchard's avatar
      clang-format for align_buffer_page_end and free_aligned_buffer_page_end · 3cf056f8
      Frank Barchard authored 
      clang-format does nested indents for macros that dont end with ;
example:
  align_buffer_page_end(dst_y_8, dst_y_plane_size)
      align_buffer_page_end(dst_u_8, dst_uv_plane_size)
          align_buffer_page_end(dst_v_8, dst_uv_plane_size)
              align_buffer_page_end(dst_y_16, dst_y_plane_size * 2)
                  align_buffer_page_end(dst_u_16, dst_uv_plane_size * 2)
                      align_buffer_page_end(dst_v_16, dst_uv_plane_size * 2)

use a similar allocator to the one used within libyuv in row.h which makes the caller add ;
  align_buffer_page_end(dst_y_8, dst_y_plane_size);
  align_buffer_page_end(dst_u_8, dst_uv_plane_size);
  align_buffer_page_end(dst_v_8, dst_uv_plane_size);
  align_buffer_page_end(dst_y_16, dst_y_plane_size * 2);
  align_buffer_page_end(dst_u_16, dst_uv_plane_size * 2);
  align_buffer_page_end(dst_v_16, dst_uv_plane_size * 2);

Bug: libyuv:758
Test: try bots
Change-Id: I4a0770707e7053e094a37bbfc3c5884d5663d078
Reviewed-on: https://chromium-review.googlesource.com/762757Reviewed-by: 's avatarPatrik Höglund <phoglund@chromium.org>
Reviewed-by: 's avatarFrank Barchard <fbarchard@google.com>
Commit-Queue: Frank Barchard <fbarchard@google.com>
      3cf056f8