1. 24 Aug, 2013 1 commit
  2. 23 Aug, 2013 6 commits
    • Yaowu Xu's avatar
      Limit mv range to be based on partition size · 13930cf5
      Yaowu Xu authored
      Previous change c4048dbd limits the mv search range assuming max block
      size of 64x64, this commit change the search range using actual block
      size instead.
      Change-Id: Ibe07ab02b62bf64bd9f8675d2b997af20a2c7e11
    • Dmitry Kovalev's avatar
      Fixing display size setting problem. · 11e3ac62
      Dmitry Kovalev authored
      Fix of https://code.google.com/p/webm/issues/detail?id=608. We could have
      used invalid display size equal to the previous frame size (not to the
      current frame size).
      Change-Id: I91b576be5032e47084214052a1990dc51213e2f0
    • Dmitry Kovalev's avatar
      Cleanup in mvref_common.{h, c}. · 21d8e859
      Dmitry Kovalev authored
      Making code more compact, adding consts, removing redundant arguments,
      adding do/while(0) for macros.
      Change-Id: Ic9ec0bc58cee0910a5450b7fb8cfbf35fa9d0d16
    • Yaowu Xu's avatar
      Added border extension · 656632b7
      Yaowu Xu authored
      To the source buffer to be encoded as an alt ref frame. This is to fix
      the problem of using uninitialized memory in encoder.
      See https://code.google.com/p/webm/issues/detail?id=605
      Change-Id: I97618a2fc207e08abcf5301b734aa9e3ad695e2c
    • Adrian Grange's avatar
      Fix bug in convolution functions (filter selection) · 3f108313
      Adrian Grange authored
      (In response to Issue 604:
      There were bugs in the convolution code for two cases:
      1. Where the filter table was assumed to be aligned to a
         256 byte boundary. The offset of the pixel in the
         source buffer was computed incorrectly.
      2. Where no such alignment assumption was made. An
         incorrect address for the filter table base was used.
      To fix both problems, I now assume that the filter table is
      256-byte aligned and modify the pixel offset calculation to
      A later patch should remove the restriction that the filter
      table is aligned to a 256-byte boundary.
      There was also a bug in the ConvolveTest unit test
      (Bug & initial fix suggestion submitted by Tero Rintaluoma
      and Sami Pietilä).
      Change-Id: I71985551e62846e55e40de9e7e3959d4805baa82
    • Jingning Han's avatar
      Fix rectangular partition check flag · 84f3b76e
      Jingning Han authored
      Put rectangular partition check flag change according to the rd
      costs of NONE and SPLIT partition types under the speed feature.
      Change-Id: If681e1e078a8d43d86961ea4b748da5cd1b6c331
  3. 22 Aug, 2013 14 commits
    • hkuang's avatar
      Add neon optimize vp9_short_idct10_16x16_add. · 4082bf9d
      hkuang authored
      vp9_short_idct10_16x16_add is used to handle the block that only have valid data
      at top left 4x4 block. All the other datas are 0. So we could cut many
      unnecessary calculations in order to save instructions.
      Change-Id: I6e30a3fee1ece5af7f258532416d0bfddd1143f0
    • Dmitry Kovalev's avatar
      vp9_encodeframe.c cleanup. · 604022d4
      Dmitry Kovalev authored
      Removing unused get_sbuv_perpixel_variance function, using has_second_ref/
      is_inter_block functions, organizing includes.
      Change-Id: I016de4af12fbbb8b4ece26a70759b2392651b095
    • Dmitry Kovalev's avatar
      check_bsize_coverage cleanup. · 335b1d36
      Dmitry Kovalev authored
      Change-Id: Ib7803857b35c00e317c9deb8630e777e25eb278f
    • Dmitry Kovalev's avatar
      Checking scale factors on access. · 3c426572
      Dmitry Kovalev authored
      It is possible to have invalid scale factors and not access them
      during decoding. Error is reported if we really try to use invalid scale
      Change-Id: Ie532d3ea7325ee0c7a6ada08269f804350c80fdf
    • James Zern's avatar
      rename LOG2_* defines to *_LOG2 · 40ae02c2
      James Zern authored
      gets rid of a mix of styles
      Change-Id: I3591d312157bc6f53a25438bf047765c671fd8a8
    • Dmitry Kovalev's avatar
      Removing useless calls to setup_{pre, dst}_planes. · 09858c23
      Dmitry Kovalev authored
      Comment is wrong, we don't initialize any xd pointers. We only initialize
      xd->planes[i]->dst and xd->planes[i]->pre[], which are actually initialized
      for every block during the decoding.
      Change-Id: If152ea872ebef1f83ca70712fa6f8df1b6855f56
    • James Zern's avatar
      vp9/encoder: fix last_frame_seg_map mem leak · a5726ac4
      James Zern authored
      remove duplicate allocation from vp9_create_compressor, it was added to
      vp9_alloc_frame_buffers in:
      d5bec522 Added resizing & initialization of last frame segment map
      Change-Id: I996723226a16a62aff8f9a52ac74e0b73cc98fdf
    • Dmitry Kovalev's avatar
      Adding vp9_is_scaled function. · 640dea4d
      Dmitry Kovalev authored
      Change-Id: Ieb7077ca3586b9491912027eed450a4f6fd38d30
    • Jingning Han's avatar
      Refactor rd_pick_partition for parameter control · 01a37177
      Jingning Han authored
      This commit changes the partition search order of superblocks from
      consistency with that of sub8x8 partition search. It enable the use
      of early termination in partition search for all block sizes.
      For ped_area_1080p 50 frames coded at 4000 kbps, it makes the runtime
      goes down from 844305ms -> 818003ms (3% speed-up) at speed 0.
      This will further move towards making the in-search partition types
      configurable, hence unifying various speed-up approaches.
      Some speed 1 and 2 features are turned off during the refactoring
      process, including:
      Stricter constraints are applied to use_square_partition_only for
      right/bottom boundary blocks. Will bring back/refine these features
      subsequently. At this point, it makes derf set at speed 1 about
      0.45% higher in compression performance, and 9% down in run-time.
      Change-Id: I3db9f9d1d1a0d6cbe2e50e49bd9eda1cf705f37c
    • hkuang's avatar
      Optimise idct4x4: rearrange the instructions a bit · 610642c1
      hkuang authored
      to improve instruction scheduling.
      Change-Id: I5ea881a6e419f9e8ed4b3b619406403b4de24134
    • Deb Mukherjee's avatar
      Fixes on feature disabling split based on variance · 8b810c7a
      Deb Mukherjee authored
      Adds a couple of minor fixes, which may be absorbed in Jingning's
      patch. Thanks to Guillaume for pointing these out.
      Also adjusts the thresholds for speed 1 and 2 to 16 and 32
      respectively, to keep quality drops small.
      derfraw300:  threshold = 16, psnr -0.082%, speedup 2-3%
                   threshold = 32, psnr -0.218%, speedup 5-6%
      stdhdraw250: threshold = 16, psnr -0.031%, speedup 2-3%
                   threshold = 32, psnr -0.273%, speedup 5-6%
      Change-Id: I4b11ae8296cca6c2a9f644be7e40de7c423b8330
    • Scott LaVarnway's avatar
      Initialize mb_skip_coeff before picking modes · 94bfbaa8
      Scott LaVarnway authored
      It appears that the above/left mb_skip_coeff used during
      the pick modes, is left over from the previously
      encode frame.  This patch initializes the flag to the default
      value of zero.
      Change-Id: Ida4684cc99611d6e3e82628db35ed717e28ce550
    • Dmitry Kovalev's avatar
      Cleaning up foreach_transformed_block_in_plane. · 4172d7c5
      Dmitry Kovalev authored
      Change-Id: I9f45af3894c57f35cb266c255e2b904295d39c34
    • James Zern's avatar
      vp9_peek_si: add bitstream v1 support · 61673553
      James Zern authored
      currently protected by CONFIG_NON420 as v1 is still not entirely stable
      Change-Id: Id1c5081b04a2c47a842822048b8804be67d23a6d
  4. 21 Aug, 2013 10 commits
    • Dmitry Kovalev's avatar
      Cleaning up optimize_init_b function. · be60924f
      Dmitry Kovalev authored
      Change-Id: Ib2c975e1d96deefb7ac4d6b600c8c5388035d111
    • Dmitry Kovalev's avatar
      Cleaning up reset_skip_context function. · c43da352
      Dmitry Kovalev authored
      Change-Id: Ib3e72671eb8da6f2e9767a6de292ec7c7cde6bc7
    • Dmitry Kovalev's avatar
      Cleaning up sum_intra_stats function. · 048ccb28
      Dmitry Kovalev authored
      Using size_group_lookup table and better variable names.
      Change-Id: I6e67f2ce091845db43ace7d21b7ae31c6f165aec
    • Dmitry Kovalev's avatar
      Removing PLANE_TYPE argument from cost_coeffs function. · 2f1a0a0e
      Dmitry Kovalev authored
      We can determine plane_type for another function arguments.
      Change-Id: I85331877aedb357632ae916a37b5b15f22c0bb1f
    • Deb Mukherjee's avatar
      Make "good" quality 2-pass vpxenc encoding default · 0d8723f8
      Deb Mukherjee authored
      Currently, the best quality mode in VP9 is not very well developed,
      and unnecessarily makes the encode too slow. Hence the command line
      default is changed to "good" quality. Also, the number of passes
      default is changed to 2 passes as well, since 1-pass encoding is
      not very efficient in VP9.
      Besides, a number of VP9 defaults are set to the currently
      recommended settings. With these changes, vpxenc
      run with --codec=vp9 --kf-max-dist=9999 --cpu-used=0 should
      work about the same as our borg results.
      Note when the --cpu-used=0 option is dropped there will be a slight
      difference in the output, because of a difference in the cpu-used
      value for the first pass. Specifically, the default when unspecified
      is to use cpu_used=1 for the first pass and cpu_used=0 for the
      second pass. But when specified, both passes will use the cpu-used
      value specified.
      Note that this also changes the default for VP8 as being "good"
      but other options stay unchanged.
      Change-Id: Ib23c1a05ae2f36ee076c0e34403efbda518c5066
    • Dmitry Kovalev's avatar
      Removing a lot of duplicated code. · 27a984fb
      Dmitry Kovalev authored
      Adding set_contexts contexts function and call it instead of
      set_contexts_on_border. Calling txfrm_block_to_raster_xy to get aoff and
      Change-Id: I41897e344afd2cae1f923f4fdbe63daccf6fe80e
    • Dmitry Kovalev's avatar
      Adding scale factor check. · a3ae4c87
      Dmitry Kovalev authored
      We support only [1/16, 2] scale factors, enforcing this now.
      Change-Id: I0822eb7cea51720df6814e42d3f35ff340963061
    • Adrian Grange's avatar
      Fix typos and minor stylistic cleanup · ce28d0ca
      Adrian Grange authored
      Change-Id: I32e43474e8651ef2eb181d24860a8f118cfea7bf
    • James Zern's avatar
      vp9 rtcd: remove non-existent sad functions · ae455fab
      James Zern authored
      vp9_sad32x3, vp9_sad3x32
      + remove unnecessary sad include from vp9_findnearmv.c
      Change-Id: Idef2a89cadc3fec64eff82ba9be60ffff50b3468
    • Dmitry Kovalev's avatar
      Removing unused foreach_predicted_block function. · 90027be2
      Dmitry Kovalev authored
      Moving foreach_predicted_block_in_plane function to vp9_reconinter.c
      because there is only one usage.
      Change-Id: I9852feae43fc3cf809b817fc541d043bc5496209
  5. 20 Aug, 2013 9 commits
    • Dmitry Kovalev's avatar
      Using has_second_ref function to simplify the code. · 27de4fe9
      Dmitry Kovalev authored
      Updating implementation of vp9_get_pred_context_single_ref_p2 using
      has_second_ref function to make code easier to read.
      Change-Id: I5ba642712f59861a48aab974e73aa01640d086fe
    • Dmitry Kovalev's avatar
      vp9_filter.{h, c} cleanup + adding SUBPEL_TAPS constant. · d19ac4b6
      Dmitry Kovalev authored
      Change-Id: Ib394ea23f464591dad50b5c65c316701378d06d7
    • hkuang's avatar
      Add neon optimize vp9_short_idct10_8x8_add. · 37cda6dc
      hkuang authored
      vp9_short_idct10_8x8_add is used to handle the block that only have valid data
      at top left 4x4 block. All the other datas are 0. So we could cut several
      unnecessary calculations in order to save instructions.
      Change-Id: I34fda95e29082b789aded97c2df193991c2d9195
    • Jingning Han's avatar
      Enable zero coeff check in sub8x8 UV rd loop · 1bf14286
      Jingning Han authored
      Check the minimum rate-distortion cost of regular quantization and
      all zero coeffs cases in the sub8x8 inter prediction rd loop for
      luma components. Use this as the cumulative rdcost sent to UV rd
      Change-Id: Ia4bc7700437d5e13d7cdad4cf9ae57ab036d3e97
    • Deb Mukherjee's avatar
      Cleanup/enhancements of switchable filter search · 2ffe64ad
      Deb Mukherjee authored
      Cleans up the switchable filter search logic. Also adds a
      speed feature - a variance threshold - to disable filter search
      if source variance is lower than this value.
      Results: derfraw300
      threshold = 16, psnr -0.238%, 4-5% speedup (tested on football)
      threshold = 32, psnr -0.381%, 8-9% speedup (tested on football)
      threshold = 64, psnr -0.611%, 12-13% speedup (tested on football)
      threshold = 96, psnr -0.804%, 16-17% speedup (tested on football)
      Based on these results, the threshold is chosen as 16 for speed 1,
      32 for speed 2, 64 for speed 3 and 96 for speed 4.
      Change-Id: Ib630d39192773b1983d3d349b97973768e170c04
    • Jim Bankoski's avatar
      fix the mv_ref_idx issue · f167433d
      Jim Bankoski authored
      The following issue was reported :
      This code makes the choice and code cleaner and removes any question
      about whether the border needs to be checked.
      Change-Id: Ia7aecfb3168e340618805bd318499176c2989597
    • Paul Wilkins's avatar
      Changes to auto partition size selection. · e8923fe4
      Paul Wilkins authored
      Changes to code to auto select a partition size range
      based on data from spatial neighbors.
      Now looks at the sb_type in each 8x8 block of above
      and left SB64.
      The effect on speed 1 is now weaker giving better
      quality but less speed gain. Now also used in speed 2.
      Change-Id: Iace33a97d5c3498dd2a9a8a4067351941abcbabc
    • Dmitry Kovalev's avatar
      Adding VP9_FILTER_BITS constant. · 2612b99c
      Dmitry Kovalev authored
      constants. Using ROUND_POWER_OF_TWO for rounding.
      Change-Id: I2e8d6858dcd600a87096138209731137d7decc24
    • Dmitry Kovalev's avatar
      Adding has_second_ref function. · d8286dd5
      Dmitry Kovalev authored
      Updating implementation of vp9_get_pred_context_single_ref_p1 using
      has_second_ref function to make code easier to read.
      Change-Id: Ie8f60403a7195117ceb2c6c43176ca9a9e70b909