Commits · d9dccc03890a976dba59d66ed3b5aceeaa33d14c · Linshizhi / ffmpeg.wasm-core

06 Oct, 2016 1 commit
- hevc: x86: Refactor IDCT macro declarations · d9dccc03
  Diego Biurrun authored Oct 05, 2016
  
  d9dccc03
05 Oct, 2016 1 commit

d3d11va: Use the proper decoding slice index · be630b1e

Steve Lhomme authored Oct 05, 2016

The decoding buffer index expected by D3D11VA is the one from the
ID3D11Texture2D not the one from the ID3D11VideoDecoderOutputView array
in AVD3D11VAContext.

Otherwise, when providing decoder slices that do not start from 0,
pictures appear in bogus order. For an invalid index crashes and
image corruption can occur.
Signed-off-by: Diego Biurrun <diego@biurrun.de>

be630b1e

04 Oct, 2016 23 commits
- vp9lpf/x86: make filter_16_h work on 32-bit. · 715f139c
  Ronald S. Bultje authored Dec 26, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  715f139c
- vp9lpf/x86: make filter_48/84/88_h work on 32-bit. · 8915320d
  Ronald S. Bultje authored Dec 26, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  8915320d
- vp9lpf/x86: make filter_44_h work on 32-bit. · 725a2164
  Ronald S. Bultje authored Dec 26, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  725a2164
- vp9lpf/x86: make filter_16_v work on 32-bit. · 5bfa96c4
  Ronald S. Bultje authored Dec 26, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  5bfa96c4
- vp9lpf/x86: make filter_48/84_v work on 32-bit. · b905e8d2
  Ronald S. Bultje authored Dec 26, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  b905e8d2
- vp9lpf/x86: make filter_88_v work on 32-bit. · 37637e65
  Ronald S. Bultje authored Dec 27, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  37637e65
- vp9lpf/x86: make filter_44_v work on 32-bit. · be10834b
  Ronald S. Bultje authored Dec 27, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  be10834b
- vp9lpf/x86: save one register in SIGN_ADD/SUB. · 7c62891e
  Ronald S. Bultje authored Dec 27, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  7c62891e
- vp9lpf/x86: store unpacked intermediates for filter6/14 on stack. · c6375a83
  Ronald S. Bultje authored Dec 24, 2014
```
filter16 goes from 508 to 482 (h) or 346 to 314 (v) cycles; filter88
goes from 240 to 238 (h) or 174 to 165 (v) cycles, measured on TOS.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  c6375a83
- vp9lpf/x86: move variable assigned inside macro branch. · 4ce8ba72
  Ronald S. Bultje authored Dec 24, 2014
```
The value is not used outside the branch.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  4ce8ba72
- vp9lpf/x86: simplify ABSSUM_CMP by inverting the comparison meaning. · e4961035
  Ronald S. Bultje authored Dec 22, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  e4961035
- vp9lpf/x86: remove unused register from ABSSUB_CMP macro. · 683da278
  Ronald S. Bultje authored Dec 20, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  683da278
- vp9lpf/x86: slightly simplify 44/48/84/88 h stores. · 6e74e963
  Ronald S. Bultje authored Dec 20, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  6e74e963
- vp9lpf/x86: make cglobal statement more conservative in register allocation. · 6411c328
  Ronald S. Bultje authored Dec 20, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  6411c328
- vp9lpf/x86: save one register in loopfilter surface coverage. · a6e288d6
  Ronald S. Bultje authored Dec 20, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  a6e288d6
- vp9lpf/x86: add ff_vp9_loop_filter_[vh]_44_16_{sse2,ssse3,avx}. · 0ed21bdc
  Clément Bœsch authored Feb 05, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  0ed21bdc
- vp9lpf/x86: add ff_vp9_loop_filter_h_{48,84}_16_{sse2,ssse3,avx}(). · f2e3d706
  Clément Bœsch authored Jan 30, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  f2e3d706
- vp9lpf/x86: add an SSE2 version of vp9_loop_filter_[vh]_88_16 · 92d47550
  James Almer authored Jan 28, 2014
```
Similar gains as the ssse3 version once again

Additional improvements by Clément Bœsch <u@pkh.me>.
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  92d47550
- vp9lpf/x86: add ff_vp9_loop_filter_[vh]_88_16_{ssse3,avx}. · 6bea4781
  Clément Bœsch authored Jan 25, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  6bea4781
- vp9lpf/x86: add ff_vp9_loop_filter_[vh]_16_16_sse2(). · 1f451eed
  James Almer authored Jan 17, 2014
```
Similar gains in performance as the SSSE3 version
Signed-off-by: James Almer <jamrial@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  1f451eed
- vp9lpf/x86: add x86 SSSE3/AVX SIMD for vp9_loop_filter_[vh]_16_16. · a692724c
  Clément Bœsch authored Jan 05, 2014
```
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  a692724c
- checkasm: add VP9 loopfilter tests. · c935b54b
  Ronald S. Bultje authored Sep 17, 2015
```
The randomize_buffer() implementation assures that "most of the time",
we'll do a good mix of wide16/wide8/hev/regular/no filters for complete
code coverage. However, this is not mathematically assured because that
would make the code either much more complex, or much less random.

Some fixes and improvements by Rodger Combs <rodger.combs@gmail.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  c935b54b
- vp9: ignore reference segmentation map if error_resilience flag is set. · a451324d
  Ronald S. Bultje authored Aug 12, 2014
```
Fixes ffvp9_fails_where_libvpx.succeeds.webm.

Bug-Id: ffmpeg/3849.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
```
  a451324d
02 Oct, 2016 15 commits
- fate: Add test for rscc palette · dc3fe45f
  Vittorio Giovara authored Sep 19, 2016
  
  dc3fe45f
- rscc: Support palette format · c19830aa
  Carl Eugen Hoyos authored Sep 19, 2016
```
Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
```
  c19830aa
- avcodec: Document AV_PKT_DATA_PALETTE side data type · b8d5070d
  Vittorio Giovara authored Sep 20, 2016
  
  b8d5070d
- avidec: Set palette alpha as fully opaque · 497c0879
  Vittorio Giovara authored Oct 02, 2016
```
Palette format is always in RGBA.
```
  497c0879
- avidec: Do not special case palette on big-endian · bad4aad4
  Vittorio Giovara authored Sep 25, 2016
```
This simplifies the code a bit, does not change output data in any way.
```
  bad4aad4
- pixfmt: Document alternative names for smpte 431 and 432 · 310c55f1
  Vittorio Giovara authored Sep 20, 2016
  
  310c55f1
- vaapi_h265: Add main 10 encode support · 5a5df90d
  Mark Thompson authored Sep 30, 2016
  
  5a5df90d
- hwcontext_vaapi: Enable P010 support · eaaaabf6
  Mark Thompson authored Sep 30, 2016
```
This is required for 10-bit surfaces.
```
  eaaaabf6
- vaapi_h265: Fix buffering parameters · b8cac1e8
  Mark Thompson authored Sep 30, 2016
```
A decoder may need this to be set correctly to output frames in the
right order.
```
  b8cac1e8
- vaapi_h265: Fix slice header writing · fc30a908
  Mark Thompson authored Sep 30, 2016
```
This was not observed earlier because the only syntax element which
it normally misses with the current setup is slice_qp_delta, but that
is always going to be zero (in IDR frames QP isn't varied on the
slice) which will always exp-golomb code as a single 1 bit.  The
immediately following part is the byte alignment, which is always a 1
bit followed by 0s which are ignored, so as long as the bitstream is
never aligned at that point we will never notice because the only
difference is that an ignored bit is a 1 instead of a 0.
```
  fc30a908
- vaapi_h264: Write bitstream restriction fields · ec17ab38
  Mark Thompson authored Oct 01, 2016
  
  ec17ab38
- vaapi_h264: Fix CFR mode with frame_rate set in AVCodecContext · 17a0f948
  Mark Thompson authored Oct 01, 2016
  
  17a0f948
- vaapi_encode: Decide on GOP setup before initialising sequence parameters · 314b421d
  Mark Thompson authored Sep 30, 2016
```
This was always too late; several fields related to it have been incorrectly
zero since the encoder was added.
```
  314b421d
- lavu: remove the custom atomic API · 5cc0057f
  Anton Khirnov authored Jul 29, 2016
```
It has been replaced by C11 stdatomic.h and is now unused.
```
  5cc0057f
- pthread_frame: use atomics for frame progress · 59c70227
  Anton Khirnov authored Jul 17, 2016
  
  59c70227