1. 01 Mar, 2013 1 commit
    • Johann's avatar
      Use intrinsics for sse2 regular quantize · eca59cad
      Johann authored
      Remove dependency of this function on asm_offsets. ssse3/sse4 next.
      
      Change quant_shift calculation so it be done using SIMD. Pre-calculate
      as much as possible to simplify EOB selection.
      
      Take advantage of qcoeff being zero'd by tying the if statements
      together.
      
      Speed parity with previous implementation with gcc x86_64 linux
      
      Change-Id: Ife97556a1eca3a74b09def1a3d04084974dff1fb
      eca59cad
  2. 28 Feb, 2013 1 commit
  3. 27 Feb, 2013 1 commit
  4. 26 Feb, 2013 1 commit
    • Johann's avatar
      vp8 fast quantizer with intrinsics · ef887974
      Johann authored
      Reduce dependency on offsets file by using intrinsics. Disassembly shows
      improvements over previous assembly specifically in register management,
      preloading, and {pro,epi}log. Speed change is within margin of error.
      
      Change-Id: I8131b4b4d62bc092407fe847bfaa8f2c0e1384ff
      ef887974
  5. 22 Feb, 2013 4 commits
  6. 17 Feb, 2013 1 commit
  7. 01 Feb, 2013 1 commit
  8. 31 Jan, 2013 1 commit
  9. 30 Jan, 2013 1 commit
    • Scott LaVarnway's avatar
      WIP: Multiple decoder instances support · 75f647fe
      Scott LaVarnway authored
      Started adding support for multiple internal decoder instances.  Also added
      code to limit the vp8 config options available when using frame-based
      multithreading.
      
      Change-Id: I0f1ee7abcfcff59204f50162e28254b8dd6972eb
      75f647fe
  10. 29 Jan, 2013 1 commit
  11. 26 Jan, 2013 1 commit
    • Scott LaVarnway's avatar
      Bug fix: error-concealment enabled changed postproc output · 0c2919c3
      Scott LaVarnway authored
      When error concealment is enabled, it swaps the mi and prev_mi ptrs after
      each frame is decoded.  The postproc uses the mi ptr for the mode info context.
      Now the postproc will use the correct mode info context.
      
      Change-Id: I537ae5450f319c624999b44525bb52bb30047b7b
      0c2919c3
  12. 22 Jan, 2013 1 commit
  13. 11 Jan, 2013 1 commit
  14. 04 Jan, 2013 1 commit
  15. 20 Dec, 2012 1 commit
  16. 18 Dec, 2012 1 commit
    • Yaowu Xu's avatar
      Changed MAX_PSNR to 100 · 28765690
      Yaowu Xu authored
      The MAX_PSNR was used to assign a "psnr" number when the mse is close
      to zero. The direct assignment is used to prevent divide by zero in
      computation. Changing it from 60 to 100 to be consistent against what
      is being done in VP9
      
      Change-Id: I4854ffc4961e59d372ec8005a0d52ca46e3c4c1a
      28765690
  17. 11 Dec, 2012 2 commits
  18. 10 Dec, 2012 3 commits
  19. 07 Dec, 2012 1 commit
    • Scott LaVarnway's avatar
      Moved denoiser frame copy/updates out of loopfilter thread · 000c8414
      Scott LaVarnway authored
      The loopfilter thread from the previous frame can be running while
      starting the current frame.  cpi->Source will change during this time causing
      the wrong data to be copied.  The refresh_x_frame flags also change, which
      will cause incorrect updates of the denoised buffers.
      
      Change-Id: I7d982b4fcb40a0610801332aa85f3b792c64e4c3
      000c8414
  20. 06 Dec, 2012 3 commits
    • Scott LaVarnway's avatar
      added work buffer for denoiser · ef2248a2
      Scott LaVarnway authored
      The denoiser was writing to LAST_FRAME buffer.   If LAST_FRAME isn't being
      updated,  the reference frame buffers were out of sync between the encoder and the
      denoised raw buffers. This patch resolves the discrepancy by always writing to a work
      buffer (INTRA_FRAME) and then copying from that buffer to any buffers that needs to
      be updated.
      
      Change-Id: I6dd855b9749978b542bc3d515914d5f16faf25df
      ef2248a2
    • Johann's avatar
      Move vp8_scale_frame to vpx namespace · a36d9a4a
      Johann authored
      Change-Id: I92d613e89c8f1174eca0789116120bfa20c25c28
      a36d9a4a
    • Johann's avatar
      Use 'vpx_scale' consistently · 1009f765
      Johann authored
      Change-Id: I178352813d2b8702d081caf405de9dbad9af2cc3
      1009f765
  21. 05 Dec, 2012 5 commits
  22. 04 Dec, 2012 1 commit
  23. 03 Dec, 2012 2 commits
  24. 29 Nov, 2012 1 commit
  25. 28 Nov, 2012 1 commit
  26. 27 Nov, 2012 1 commit
    • Marco Paniconi's avatar
      Updates to qp-regulate and rate correction factor. · 464b1df6
      Marco Paniconi authored
      Don't use the switch to gf_rate_correction factor when
      temporal layers is used (i.e., cpi->oxcf.number_of_layers > 1).
      
      In temporal layers, we prefer to avoid this as any frame
      (e.g., base layer frame at anchor of pattern) may update
      both last and golden (and possibly alt-ref), and so we would get
      different rate correction factors within the same layer.
      
      This change will make sure one rate correction factor exists for each layer.
      
      Also, made some other code in qp-regulate that depends on
      alt/golden update specific to the 1 layer case.
      
      Change-Id: I41a6d085bd477f9307ef3b3c311695214273892c
      464b1df6
  27. 25 Nov, 2012 1 commit