1. 26 Aug, 2014 1 commit
    • Yaowu Xu's avatar
      add a new interp filter search strategy. · 1144fee3
      Yaowu Xu authored
      This commit addes a new strategy to reduce the search for optimal
      interpolation filter type. The encoder counts and store how many each
      filter type is selected and used for each of the reference frames.
      A filter type that is rarely used for all three reference frames is
      masked out to avoid computation.
      
      The impact on compression is neglectible:
      -0.02% on derf
      +0.02% on stdhd
      
      Encoding time is seen to reduce by 2~3%.
      
      Change-Id: Ibafa92291b51185de40da513716222db4b230383
      1144fee3
  2. 22 Aug, 2014 5 commits
  3. 21 Aug, 2014 1 commit
  4. 20 Aug, 2014 2 commits
  5. 19 Aug, 2014 11 commits
  6. 18 Aug, 2014 7 commits
    • Jim Bankoski's avatar
      vp9_firstpass.c : clear warning re: use of null lc · b78f8f67
      Jim Bankoski authored
      Change-Id: I80f81f6f4f15766f504c15f8c20258a84bbaa9ac
      b78f8f67
    • Yunqing Wang's avatar
      Add early termination in transform size search · ba70f160
      Yunqing Wang authored
      In the full-rd transform size search, we go through all transform
      sizes to choose the one with best rd score. In this patch, an
      early termination is added to stop the search once we see that the
      smaller size won't give better rd score than the larger size. Also,
      the search starts from largest transform size, then goes down to
      smallest size.
      
      A speed feature tx_size_search_breakout is added, which is turned off
      at speed 0, and on for other speeds. The transform size search is
      turned on at speed 1.
      
      Borg test results:
      1. At speed 1,
         derf set: psnr gain: 0.618%, ssim gain: 0.377%;
         stdhd set: psnr gain: 0.594%, ssim gain: 0.162%;
         No noticeable speed change.
      3. At speed 2,
         derf set: psnr loss: 0.157%, ssim loss: 0.175%;
         stdhd set: psnr loss: 0.090%, ssim loss: 0.101%;
         speed gain: ~4%.
      
      Change-Id: I22535cd2017b5e54f2a62bb6a38231aea4268b3f
      ba70f160
    • Jingning Han's avatar
      Speed up mode search depending on relative ref frame position · 6a464eca
      Jingning Han authored
      This commit enables the encoder to record the location of the
      center frame to generate alter reference frame. It then allows to
      skip checking prediction modes of other reference frame types when
      it comes to encode this frame.
      
      The speed 3 runtime is reduced for the test sequences:
      bus at CIF 1000 kbps, 9791 ms -> 9446 ms, i.e., 3.5% speed-up,
      pedestrian at 1080p 2000 kbps, 184043 ms -> 175730 ms, i.e., 4.5%
      speed-up.
      
      No compression performance change observed.
      
      Change-Id: Iacfde3bcc1445964e7a241f239bd6ea11cb94bd1
      6a464eca
    • Dmitry Kovalev's avatar
      Adding gf_group temp variable. · 4590c5e5
      Dmitry Kovalev authored
      Change-Id: Ifc33b1e893480b36ccf88794ea15e87e5cc97823
      4590c5e5
    • Dmitry Kovalev's avatar
      Removing unnecessary casts to int64_t. · a885e1cb
      Dmitry Kovalev authored
      Change-Id: I156419fdf68056759a4cd4206cc829a00025ef78
      a885e1cb
    • Dmitry Kovalev's avatar
      Replacing SVC oxcf->mode check with oxcf->pass check. · 17a26eb4
      Dmitry Kovalev authored
      oxcf->pass is used in all other places in similar contexts.
      
      Change-Id: I9de3a5f24311838d5a6dc3dd470d258cb189ff59
      17a26eb4
    • Yaowu Xu's avatar
      Revert "Removing duplicated code from init_config()." · beff5b6a
      Yaowu Xu authored
      This reverts commit 5509b7fd
      
      
      Observed a big drop in compression quality and speed for speed 1 for a 360p test clip, revert this now for investigation.
      
      Change-Id: If69dc8d77a225b34dc7907a9472e1a7a0a22762d
      beff5b6a
  7. 16 Aug, 2014 1 commit
    • Dmitry Kovalev's avatar
      Removing vp9_rescale(). · 1acc96fb
      Dmitry Kovalev authored
      Now everything is int64_t, so we don't need that function.
      
      Change-Id: I7dd2023677b4943ecd968d68cc06f9d4862c3c3f
      1acc96fb
  8. 15 Aug, 2014 7 commits
  9. 14 Aug, 2014 5 commits
    • Jingning Han's avatar
      Remove redundant vp9_init_plane_quantizers call · d67b608c
      Jingning Han authored
      When aq mode is on, the quantizer will be reset later in the same
      function (line 571).
      
      Change-Id: I20635db31261d136d04d5deeb881ad3957078bf1
      d67b608c
    • levytamar82's avatar
      32 Align Load bug · efdfdf57
      levytamar82 authored
      In the sub_pixel_avg_variance the parameter sec was also aligned load and
      changed to unaligned.
      
      Change-Id: I4d4966e0291059ea4d705baed1503dc58444fcb7
      efdfdf57
    • Yaowu Xu's avatar
      Mask out H_PRED and V_PRED for 32x32 blocks · 5966586a
      Yaowu Xu authored
      Change-Id: I2847af5062b5fa320629fcabb9fa6b23ba3e5513
      5966586a
    • Yaowu Xu's avatar
      Set max_intra_bsize to 32x32 · 4d6d0613
      Yaowu Xu authored
      At --good and speed 3 or above for resolution less than 720p. This
      disables the tests for 64x64 intra prediction modes. Encoding time
      reduction is about 1%.
      
      Change-Id: Ib396e3d1417fece416e3f0fee929b128acbb130f
      4d6d0613
    • Adrian Grange's avatar
      Fix bug 837: realloc mode info buffers on resize · 4e30565a
      Adrian Grange authored
      The test to determine if the mode info buffers need
      to be resized when the frame size changes was
      incorrect, as per bug 837.
      
      By storing the size of the allocated data structure,
      a simple test determines whether to allocate more
      memory when the frame size changes.
      
      Change-Id: I1544698f2882cf958fc672485614f2f46e9719bd
      4e30565a