1. 14 Oct, 2016 7 commits
    • Nathan E. Egge's avatar
      Use Daala entropy coder to code bits. · 8043cc40
      Nathan E. Egge authored
      When building with --enable-daala_ec, calls to aom_write() and aom_read()
       use the daala entropy coder to write and read bits.
      When the probability is exactly 0.5 (128), then raw bits are used.
      
      ntt-short-1:
      
                MEDIUM (%) HIGH (%)
          PSNR -0.027556  -0.020114
       PSNRHVS -0.027401  -0.020169
          SSIM -0.027587  -0.020151
      FASTSSIM -0.027592  -0.020102
      
      subset1:
      
               RATE (%)  DSNR (dB)
          PSNR 0.03296  -0.00210
       PSNRHVS 0.03537  -0.00281
          SSIM 0.03299  -0.00161
      FASTSSIM 0.03458  -0.00111
      
      Change-Id: I48ad8eb40fc895d62d6e241ea8abc02820d573f7
      8043cc40
    • Yaowu Xu's avatar
      Clear existing format issue in the codebase · 931bc2a7
      Yaowu Xu authored
      Fix the clang-format warnings on the existing codes.
      
      Change-Id: I8e9e781b6f68f41a7fbd0a2116f6b35290d73dc8
      931bc2a7
    • Nathan E. Egge's avatar
      Add Daala entropy coder. · 1078dee5
      Nathan E. Egge authored
      Change-Id: I2849a50163268d58cc5d80aacfec1fd02299ca43
      1078dee5
    • Alex Converse's avatar
      Switch rANS to 15 bit precision, and adjust L_BASE. · 62a94a64
      Alex Converse authored
      This causes rANS to operate at the same precision as the Daala EC.
      
      aom/master stats: rans10uabs8lbase12 → rans15uabs8lbase15
      
      objective-1-fast
      PSNR YCbCr:      0.01%      0.01%      0.01%
         PSNRHVS:      0.01%
            SSIM:      0.01%
          MSSSIM:      0.01%
       CIEDE2000:      0.01%
      
      subset1
      PSNR YCbCr:     -0.01%     -0.00%     -0.00%
         PSNRHVS:     -0.01%
            SSIM:     -0.01%
          MSSSIM:     -0.01%
       CIEDE2000:     -0.01%
      
      (cherry picked from aom/master commit ddbc2e2a)
      
      Change-Id: I6ef0a4f6198784b3712a61af9f105d560a22eaea
      62a94a64
    • James Zern's avatar
      variance_avx2: sync variance functions with c-code · 8c64331a
      James Zern authored
      add missing int64 -> uint32 cast; quiets -Wshorten-64-to-32 warnings
      
      Change-Id: I4850b36e18dc8b399108342be4bfe0b684aefb78
      (cherry picked from commit 6acd061aad8cf62000cc9117390d0c94581a8591)
      8c64331a
    • Alex Converse's avatar
      Resolve -Wshorten-64-to-32 in variance. · 2176b7ac
      Alex Converse authored and James Zern's avatar James Zern committed
      The subtrahend is small enough to fit into uint32_t.
      
      Change-Id: Ic4d7128aaa665eaf6b25d562610ba8942c46137f
      (cherry picked from commit c0241664aac3a1805db9bd8e09e071ac326531e0)
      2176b7ac
    • Yi Luo's avatar
      Delete some redundant function declarations in aom_dsp_rtcd_defs.pl · 761ae880
      Yi Luo authored
      Change-Id: I4df57a7faba5800c048b2dc469ec31545406f55c
      761ae880
  2. 13 Oct, 2016 9 commits
  3. 12 Oct, 2016 3 commits
    • Steinar Midtskogen's avatar
      On x86 use _mm_set_epi32 when _mm_cvtsi64_si128 isn't available · b0748238
      Steinar Midtskogen authored
      Change-Id: Ibdbd720d4f68892da6164a9849e212e759305005
      b0748238
    • Yi Luo's avatar
      Hybrid forward transform 32x32 AVX2 optimization · fed8e1c0
      Yi Luo authored
      - av1_fht32x32 AVX2 function level time reduction ~89% compared to C.
      
      - av1_fht32x32_avx2() on DCT_DCT improves 42.62% over aom_fdct32x32_avx2()
        But function replacement must go with the corresponding inverse txfm.
      
      - No obvious user level time reduction due to 32x32 TX_TYPE selection.
      
      - Zero high 128b YMM to avoid AVX-SSE transition penalties
        (fix 16x16 case).
      
      - Added 32x32 AVX2 unit tests to verify bitexact.
      
      - AVX2 optimization summary:
        On CPU i7-6700, based on 16x16/32x32 fwd txfm optimization results:
        C to AVX2: function level time reduction, ~86-89%.
        SSE2 to AVX2: function level time reduction, ~51%.
      
      Change-Id: Idd0cd8bf066a61c7117140ef15ab6c1f8eb4b036
      fed8e1c0
    • Yaowu Xu's avatar
      minor updates · f36d0b46
      Yaowu Xu authored
      1. vp8->aom
      2. removed no-effect statements and spaces
      
      Change-Id: I367d05ff9bf1b9f3c71c517c45d8049d9d4236ec
      f36d0b46
  4. 11 Oct, 2016 4 commits
  5. 10 Oct, 2016 7 commits
  6. 06 Oct, 2016 3 commits
  7. 03 Oct, 2016 1 commit
  8. 28 Sep, 2016 3 commits
  9. 26 Sep, 2016 1 commit
  10. 22 Sep, 2016 1 commit
  11. 21 Sep, 2016 1 commit