1. 18 Jun, 2016 1 commit
    • James Zern's avatar
      remove vp10 · 67edc5e8
      James Zern authored
      development has moved to the nextgenv2 branch and a snapshot from here
      was used to seed aomedia
      Change-Id: Iedaca11ec7870fb3a4e50b2c9ea0c2b056a0d3c0
  2. 17 Jun, 2016 1 commit
    • Tom Finegan's avatar
      Output frames in first pass for VPX_DL_REALTIME. · 5a9f21db
      Tom Finegan authored
      Since combining VPX_DL_REALTIME with VPX_RC_FIRST_PASS is basically
      nonsense, ignore the user's pass setting when this happens and
      behave as if the requested encode is a single pass encode.
      Change-Id: I5ee4c4e5838c4ca6d24988890aae490b10826db2
  3. 16 Jun, 2016 1 commit
  4. 15 Jun, 2016 1 commit
  5. 14 Jun, 2016 2 commits
    • Johann's avatar
      neon hadamard 16x16 · c516dd67
      Johann authored
      Runs about twice as fast as C
      Change-Id: I6760d99f4e22259439ca35d746194b12a81bfa71
    • Johann's avatar
      hadamard 16x16 test · 32ff4906
      Johann authored
      Change-Id: Ibe58781905b372b9fe29dace39b4bfdd33fd0f83
  6. 09 Jun, 2016 2 commits
    • James Zern's avatar
      fdct8x8_test: fix unsigned overflow · 95d2dc89
      James Zern authored
      the difference between src and dst will be signed, the error will be
      quiets -fsanitize=integer:
      unsigned integer overflow: 4294967295 * 4294967295
      Change-Id: I580813093ee46284fde7954520dfcb1188f79268
    • James Zern's avatar
      fdct4x4_test: fix unsigned overflow · 06c6e4cb
      James Zern authored
      the difference between src and dst will be signed, the error will be
      quiets -fsanitize=integer:
      unsigned integer overflow: 4294967295 * 4294967295
      Change-Id: I502fd707823c4faaa7f587c9cc0312f057e04904
  7. 03 Jun, 2016 2 commits
  8. 02 Jun, 2016 1 commit
    • Linfeng Zhang's avatar
      Slow pshufb removal in 3 intra prediction functions. · ad0646cb
      Linfeng Zhang authored
      Replaced vpx_d45_predictor_4x4_ssse3(), vpx_d45_predictor_8x8_ssse3()
      and vpx_d207_predictor_4x4_ssse3() with
      created vpx_d45_predictor_4x4_sse2(), vpx_d45_predictor_8x8_sse2()
      and vpx_d207_predictor_4x4_sse2() respectively.
      It's mostly neutral or slightly worse than ssse3 in good cases and
      better than ssse3 in the bad cases (but still worse than using the mmx
      Change-Id: Ib0237ceb71d2c57b8a93fd3170330cfed9d56bdd
  9. 01 Jun, 2016 1 commit
    • Yaowu Xu's avatar
      Fix UBSAN/IOC errors · 6382727d
      Yaowu Xu authored
      1. test/dct16x16_test.cc
      2. test/dct32x32_test.cc
      3. test/fdct8x8_test.cc
      Change-Id: I9c9315fbd65ddb3b44f688e01ba265fd22192198
  10. 27 May, 2016 2 commits
    • James Zern's avatar
      acm_random,Rand9Signed: correct cast · 13d48c42
      James Zern authored
      convert the random value to int16 before subtracting 256 from it; quiets
      a ubsan (sanitize=integer) warning
      Change-Id: Ibc2c5a21f30e112bd6c180f7d6a033327c38d0df
    • Linfeng Zhang's avatar
      Upgrade fwht4x4_mmx() to fwht4x4_sse2() for vp9 and vp10. · af7fb17c
      Linfeng Zhang authored
      Function level timing test shows about 27% time saving on
      a Xeon E5-2680 v2 desktop.
      Rename vp9_dct_sse2.c to vp9_dct_intrin_sse2.c for vp9 and
      rename dct_sse2.c to dct_intrin_sse2.c for vp10 to avoid
      duplicate basenames.
      Actually vp9_fwht4x4_mmx/sse2() and vp10_fwht4x4_mmx/sse2()
      are identical. TODO: They should be unified later if there is
      no intention to keep a duplicate.
      Change-Id: I3e537b7bbd9ba417c606cd7c68c4dbbfa583f77d
  11. 26 May, 2016 1 commit
    • Linfeng Zhang's avatar
      Upgrade vpx_lpf_{vertical,horizontal}_4 mmx to sse2 · 4b5e462d
      Linfeng Zhang authored
      Followed the code style of other lpf fuctions.
      These 2 functions put 2 rows of data in a single xmm register,
      so they have similar but not identical filter operations,
      and cannot share the same macros.
      Change-Id: I3bab55a5d1a1232926ac8fd1f03251acc38302bc
  12. 25 May, 2016 1 commit
  13. 24 May, 2016 1 commit
  14. 20 May, 2016 2 commits
  15. 18 May, 2016 2 commits
  16. 16 May, 2016 2 commits
    • Johann's avatar
      neon hadamard 8x8 · 9b54e812
      Johann authored
      Runs about 30% faster than the C
      Change-Id: I6809d6d84c3077ab619c53298296950e976bdaba
    • hui su's avatar
      Add level test for VP9 · cafbf63d
      hui su authored
      Change-Id: I99f50bdd5af3f64a029c2f5f6f5fb1ff45bad67e
  17. 13 May, 2016 1 commit
  18. 11 May, 2016 4 commits
    • Tom Finegan's avatar
      simple_encoder: Add a frame limit argument. · 7d6edc3d
      Tom Finegan authored
      - Add frame limit argument.
      - Make all arguments required.
      - Enable the VP9 simple encoder test.
      Change-Id: I11d228b358ff90c60ea92e02760cb476434ea571
    • Tom Finegan's avatar
      twopass_encoder: Add frame limit argument. · 9d47341a
      Tom Finegan authored
      - Remove twopass_encoder test TODO re frame limit.
      - Enable VP9 twopass_encoder test.
      Change-Id: I0649f15aabef79a63891e997fd20b212af5672e6
    • Linfeng Zhang's avatar
      remove mmx variance functions · d0ffae82
      Linfeng Zhang authored
      there are sse2 equivalents which is a reasonable modern baseline
      Removed mmx variance functions:
      Change-Id: Iffaf85344c6676a3dd337c0645a2dd5deb2f86a1
    • Linfeng Zhang's avatar
      remove mmx sad functions · d0e687bf
      Linfeng Zhang authored
      there are sse2 equivalents which is a reasonable modern baseline
      Change-Id: Ibbe536a5ad1c2cccef6bdcc75c13b3dde35a56ba
  19. 09 May, 2016 1 commit
  20. 07 May, 2016 1 commit
  21. 03 May, 2016 1 commit
    • Jim Bankoski's avatar
      libvpx: add a unit test for plane_add_noise. · 34d5aff7
      Jim Bankoski authored
      In so doing this fixes a couple of bugs:
      vpx_plane_add_noise.c needed to subtract a clamp instead of add.
      And the assembly (mmx sse) had assumptions that parameters were
      continuous in memory which was not true.
      Change-Id: I76f2c43cf54bfc838eb2edf8a443eaaa7565d7b5
  22. 28 Apr, 2016 1 commit
  23. 27 Apr, 2016 5 commits
  24. 26 Apr, 2016 1 commit
    • JackyChen's avatar
      vp9: Simplify the logic in denoiser SSE2 code. · 1a53c0c9
      JackyChen authored
      Block size passed into denoiser filter is always >= BLOCK_8X8 (in
      vp9_pick_inter_mode), it is not necessary to check smaller block
      size. Passed the bitexact test on clips with different resolutions and
      noise levels.
      Change-Id: I19fa3195d18c27d9e5de60dc11cff1522ef3714e
  25. 22 Apr, 2016 2 commits