1. 06 Apr, 2017 1 commit
    • Yue Chen's avatar
      Disable OBMC + compound · 8636da61
      Yue Chen authored
      Disable OBMC mode if the current block uses compound prediction.
      Results if we only disable this combination for small
      block sizes are also listed.
      
      Loss of gain (lowres/midres/AWCYHL)
      no compound :         0.230%/0.105%/0.15%
      no compound <= 8x8:   0.043%/-0.016%/0%
      no compound <= 16x16: 0.153%/0.029%/0.05%
      
      
      Change-Id: I94c9d704172dd02e53f3b2004a8032a4f1265057
      8636da61
  2. 05 Apr, 2017 1 commit
  3. 02 Apr, 2017 1 commit
    • Yue Chen's avatar
      Use 1 sample per neighbor for local warping model estimation · 5558e5da
      Yue Chen authored
      Only 1 sample needs to be collected. Max of 8 neighbors are
      used.
      In LS estimation, the projection samples (sx, sy)->(dx, dy) are
      intentionally smoothed by assuming 3 shifted versions
      (sx, sy+n)->(dx, dy+n), (sx+n, sy)->(dx+n, dy), (sx+n,
      sy+n)->(dx+n, dy+n) also contribute to the estimation.
      For example, instead of using A[0] = sx^2, we use the sum of
      squares of source x of four points, A[0] += 4sx^2+4*n*sx+n^2.
      But computational cost wise, it does not add much overhead. Coding
      gain is mostly same as the old version. If no smoothing is added,
      will lose 0.3% on lowres.
      
      Change-Id: I04be32cffa525f7dc8ee583c0bf211d7bdc6e609
      5558e5da
  4. 30 Mar, 2017 1 commit
    • Yue Chen's avatar
      Restrict # of neighbors in obmc blending · 1bd42be6
      Yue Chen authored
      Only blend with the first N neighbors at each side. If the size of
      one dimenstion is 8/16/32/64, the max # of neighbors to overlap
      with is 1/2/3/4.
      Previously we disable obmc mode if there are too many neighbors.
      
      Change of performance in AWCY, compared to disabling obmc if
      at any side there are more than 2 overlappable neighbors.
      HL improved by 0.02%
      LL improved by 0.09%
      
      Change-Id: I93d9a65c6c4aabf0b4a4946e2253d3e2ef21a662
      1bd42be6
  5. 22 Mar, 2017 1 commit
    • Jean-Marc Valin's avatar
      Jointly optimizing deringing and clpf · 5f5c132f
      Jean-Marc Valin authored
      We now signal joint strengths and use a greedy algorithm for the search.
      
      low-latency, cpu-used=4:
      
      ll4-cdef@2017-03-22T03:42:10.815Z -> ll4-cdef-newsearch-var-header-newlambda-refine4@2017-03-22T15:56:46.471Z
      
         PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
      -0.0792 |  0.3551 |  0.4393 |  -0.0108 | -0.1338 | -0.0141 |     0.1452
      
      Change-Id: I619ae1c7c7d7ec04fe993cabc5773b07c3f5b201
      5f5c132f
  6. 21 Mar, 2017 1 commit
    • Thomas Davies's avatar
      Align EXT_TX with EC_MULTISYMBOL and EC_ADAPT. · b1bedf5f
      Thomas Davies authored
      Do multisymbol coding for transform type.
      Load default cdf probabilities directly.
      Use CDF frame update mechanism when EC_ADAPT is
      enabled.
      
      Change-Id: Id23c927e81587b560e9df8b9bc56c0e2e3bb6f03
      b1bedf5f
  7. 20 Mar, 2017 1 commit
    • Debargha Mukherjee's avatar
      Add config flags and macros to control ext-inter · 37f6fe61
      Debargha Mukherjee authored
      Adds a dependent config flag 'interintra' to turn on/off interintra
      modes altogether.
      Adds a dependent config flag 'wedge' to turn on/off wedge compound
      for both interinter and interintra.
      
      Adds another macro to change wedge predictors to use
      only 0, 1/2, or 1 weights.
      
      From now, use
      --enable-ext-inter --enable-wedge --enable-interintra to get the
      same behavior as the old --enable-ext-inter.
      
      Change-Id: I2e787e6994163b6b859a9d6431b87c4217834ddc
      37f6fe61
  8. 17 Mar, 2017 3 commits
    • Sarah Parker's avatar
      Macro to disable warped/obmc_causal with global motion · 19234cc2
      Sarah Parker authored
      Enabling SEPARATE_GLOBAL_MOTION will remove the ability for
      a block that uses zeromv with global motion to pick warped_causal
      or obmc_causal as the motion mode. When this is enabled there is:
      
      0.05% drop on lowres for global + warped enabled
      0.15% drop on midres for global + warped enabled
      
      0.12% drop on lowres with global + motion var enabled
      0.07% drop on midres with global + motion var enabled
      No performance change for global, warped, or motion var individually.
      
      Change-Id: Idbfb8dd7a93da14902438504b06a08e5212e48cb
      19234cc2
    • Debargha Mukherjee's avatar
      Fix config flags among var-tx/ext-tx/rect-tx · 428bbb2b
      Debargha Mukherjee authored
      BUG=aomedia:396
      
      Change-Id: I1b7f6dddb06eac9e341518a95a0d71465e613e9e
      428bbb2b
    • Steinar Midtskogen's avatar
      Merge dering/clpf rdo and filtering · a9d41e88
      Steinar Midtskogen authored
      * Dering and clpf were merged into a single pass.
      * 32x32 and 128x128 filter block sizes for clpf were removed.
      * RDO for dering and clpf merged and improved:
        - "0" no longer required to be in the strength selection
        - Dering strength can now be 0, 1 or 2 bits per block
      
                    LL    HL
      PSNR:       -0.04 -0.01
      PSNR HVS:   -0.27 -0.18
      SSIM:       -0.15 +0.01
      CIEDE 2000: -0.11 -0.03
      APSNR:      -0.03 -0.00
      MS SSIM:    -0.18 -0.11
      
      Change-Id: I9f002a16ad218eab6007f90f1f176232443495f0
      a9d41e88
  9. 16 Mar, 2017 1 commit
    • Angie Chiang's avatar
      Draft of av1_update_txb_context · 0397eda0
      Angie Chiang authored
      This draft version only pass compiling check, it's not working yet.
      
      The following goal is to use new coding system when doing bitstream
      packing but keep old coding system in RD loop.
      
      Change-Id: I224a1581d1cc5c67d73e71558fb77d9faf9c2470
      0397eda0
  10. 15 Mar, 2017 1 commit
    • Sarah Parker's avatar
      Fix broken assert when rect-tx is enabled on its own · 076437f4
      Sarah Parker authored
      The function is_rect_tx previously was only defined when
      both ext-tx and rect-tx are enabled together. This function
      needs to be defined when rect-tx is enabled alone to fix
      failing asserts.
      
      Change-Id: Ic8153a55b768cb49caac07d6aeaee109d63410a9
      076437f4
  11. 04 Mar, 2017 1 commit
  12. 03 Mar, 2017 1 commit
    • Yue Chen's avatar
      Restrict the number of neighbors in obmc mode · 5329a2bf
      Yue Chen authored
      Enable obmc mode only when there are <= 2 left neighbors and <=2
      above neighbors. Also disable it when there is no overlappable
      neighbors.
      
      Gain in AWCY test: 1.60%, was 1.64% when there is no restriction
      
      Change-Id: I2d82ef4fb4daa9b0843ac8844f99b9f412c4f379
      5329a2bf
  13. 02 Mar, 2017 2 commits
    • Jingning Han's avatar
      Avoid the use of undefined marco value · b83e64ba
      Jingning Han authored
      Always define USE_TXTYPE_SEARCH_FOR_SUB8X8_IN_CB4X4 to avoid the
      use of undefined value.
      
      Change-Id: I0ad90c5b5316db231e9538487bb4591dfd6a9ce7
      b83e64ba
    • Yue Chen's avatar
      Use 3-tap spatial filter in FILTER_INTRA experiment · 8d8638a1
      Yue Chen authored
      3-tap recursive intra prediction filters are added.
      Macro USE_3TAP_INTRA_FILTER is set to 1 to use 3-tap by default.
      Coding gain of FILTER_INTRA experiment in AWCY, high delay 150f
      3-tap: 0.51%
      4-tap: 0.68%
      
      Change-Id: I44192dd08bfd8155f58a9b0b5cf1de88fceb762e
      8d8638a1
  14. 28 Feb, 2017 1 commit
    • Jingning Han's avatar
      Disable compound mode in sub8x8 coding blocks · c41a549a
      Jingning Han authored
      Disable the support of compound prediction modes for sub8x8 codking
      blocks. Make the rate-distortion optimizations process account for
      such constraints.
      
      With the use 2x2 chroma prediction block, this makes the wrost case
      number of inter predictors same as vp9. It affects the coding
      gains by 0.35% for lowres, 0.17% for midres, and 0.08% for hdres.
      
      The encoding speed is up by 10%.
      
      Change-Id: Ieb2a83030676911baa403e586f1f800cbf485d81
      c41a549a
  15. 27 Feb, 2017 2 commits
    • Debargha Mukherjee's avatar
      Adds macro to test cb4x4 w/o sub8x8 txtype search · 094c9439
      Debargha Mukherjee authored
      USE_TXTYPE_SEARCH_FOR_SUB8X8_IN_CB4X4 macro added to turn
      tx_type search on/off for sub8x8 in cb4x4 mode.
      
      The purpose is mainly to analyze the coding gains from cb4x4
      but this later can be made into a speed feature as well.
      
      Change-Id: Ic22026c373eebba87f324689ac5686a2844315b6
      094c9439
    • Angie Chiang's avatar
      Prefer using get_tx_size() · 7fcfee40
      Angie Chiang authored
      Change-Id: Ifcdd3ce2953c1ecb1d0962da412a4b5ba2cda912
      7fcfee40
  16. 24 Feb, 2017 2 commits
  17. 23 Feb, 2017 2 commits
    • Sarah Parker's avatar
      Avoid sending tx type for ext_tx reduced set for sub 8x8 · 5effe3fb
      Sarah Parker authored
      This makes the handling of sub8x8 the same as when the
      ext_tx reduced set is turned off.
      
      Change-Id: I3123e473fc70469330bb49d960d02cb1aebc17cf
      5effe3fb
    • Jingning Han's avatar
      Add chroma 2x2 block experiment · 31b6a4f9
      Jingning Han authored
      Support the use of 2x2 prediction and transform block size for
      chroma components. This experiment builds on top of cb4x4. It
      brings back the loss in performance gains due to 4x4 chroma
      component constrain. The coding gains for lowres are up from 1.5%
      to 2.5%.
      
      Change-Id: I8e461e230c2d920f6df1af9d935ebe9dfa3820d9
      31b6a4f9
  18. 22 Feb, 2017 2 commits
    • Debargha Mukherjee's avatar
      Reduce transform set for 16x32/32x16 · 08542b96
      Debargha Mukherjee authored
      Previously the transform type sets used for 16x32/32x16 transform
      sizes corresponded to 16x16 transforms. Now they correspond to
      32x32 transforms with a reduced available set.
      
      BDRATE impact:
      0.05% worse for lowres, 0.13% worse for midres
      
      Change-Id: I01d405d497c13837f5404f9e0caf8c571f3b26cf
      08542b96
    • Sarah Parker's avatar
      Add default reduced set for ext-tx · e68a3e4f
      Sarah Parker authored
      Here we create a default set for the ext-tx experiment which
      includes:
      
      Intra: {DCT, ADST} X {DCT, ADST} + IDTX (identity in both directions)
      Inter: DCT_DCT + IDTX
      
      This set is used if indicated in the uncompressed header.
      
      Change-Id: Ic907ada12b4e0c98cabb7c9650ab99942d0c0859
      e68a3e4f
  19. 18 Feb, 2017 3 commits
  20. 16 Feb, 2017 1 commit
  21. 03 Feb, 2017 2 commits
  22. 26 Jan, 2017 2 commits
    • Urvang Joshi's avatar
      Palette: Don't store tokens for pixels outside image boundary. · 56ba91bb
      Urvang Joshi authored
      If part of a block falls outside right and/or bottom image boundary,
      then only store tokens for the part of it within the boundary.
      
      Also, consider only the part of the block within the boundary when
      calculating the number of colors in the image, deciding the base
      colors for palette, RD calculation etc.
      
      The part of color map corresponding to pixels outside the image
      boundary is padded with color indices copied from same row/column.
      This behavior is similar to how pixels outside the boundary are padded.
      
      For screen_content set, this is improves compression performance by
      0.038 overall. One clip, in particular, has a significant gain of 0.8.
      
      Change-Id: I745ca032f313c5041aacc98c03ae4bfc33d840de
      56ba91bb
    • ltrudeau's avatar
      Convert PVQ skip variable to enum · e1c0929f
      ltrudeau authored
      Creates the PVQ_SKIP_TYPE enum to encapsulate the different types of
      skipping that can be signaled by PVQ (i.e. skip: AC, DC or both).
      
      There is no impact on the bitstream. However, the decoder will now emit
      an internal error if the decoded skip flag is out of range. The
      block_skip variable is also renamed to ac_dc_coded as it stores the same
      information.
      
      Change-Id: Ib2aadaf99dc1736ea392ae5ed8948c3cdc12da9b
      e1c0929f
  23. 24 Jan, 2017 3 commits
    • David Barker's avatar
      Make ext-inter use new rectangular intra predictor · 839467f4
      David Barker authored
      Now that https://aomedia-review.googlesource.com/#/c/6729/
      has been merged, build_intra_predictors_for_interintra() is
      now redundant, so replace it by a direct call to
      av1_predict_intra_block() and remove the old function.
      
      Reset rect_interintra back to 1.
      
      To do this, we need to make the intra predictor take a
      BLOCK_SIZE instead of a TX_SIZE. This is because we need to
      be able to predict 32x64 and 64x32 blocks, but there is no
      TX_32X64 or TX_64X32.
      
      No effect on output or performance.
      
      Change-Id: I8c185a211c97a85012cc54ec293c785a693608ed
      839467f4
    • Yaowu Xu's avatar
      Fix a couple of typos · a93e65e5
      Yaowu Xu authored
      Change-Id: Ibec40c3cd8e14343b096e406ba233cf4f131e7b9
      a93e65e5
    • Jingning Han's avatar
      Fix conflicts between ext-inter and cb4x4 modes · 61418bbd
      Jingning Han authored
      Resolve the broken coding pipeline in ext-inter experiment when
      cb4x4 mode is enabled. Turn off rectangular inter-intra mode.
      This needs some more work to hook up. Given that it gives fairly
      limited coding performance gains, disable it for the moment.
      
      BUG=aomedia:309
      
      Change-Id: I9b406df6183f75697bfd4eed5125a6e9436d84b0
      61418bbd
  24. 23 Jan, 2017 1 commit
    • Yushin Cho's avatar
      Add a new experiment, DAALA_DIST · 7a428ba2
      Yushin Cho authored
      This commit adds a new experiment, Daala's distortion function,
      which is designed to better approximate perceptual distortion
      in 8x8 pixel blocks.
      
      This experiment is expected to work best with PVQ.
      
      It measures the variance of overlapped 4x4 regions in the 8x8 area,
      then uses these variances to scale the MSE of weighted frequency domain
      distortion of 8x8 block.
      
      Since AV1 calculates distortion in blocks as small as 4x4, it is not possible to
      directly replace the existing distortion functions of AV1,
      such as dist_block() and block_rd_txf().
      Hence, there has been substantial changes in order to apply
      Daala's 8x8 distortion function.
      The daala distortion function is applied
      after all 4x4 tx blocks in a 8x8 block are encoded (during RDO),
      as in below two cases:
      1) intra/inter sub8x8 predictions and
      2) 4x4 transform with prediction size >= 8.
      
      To enable this experiment, add '--enable-daala-dist' with configure.
      
      TODO: Significant tuning of parameters is required since the function has
      originally came from Daala thus most parameters would not work
      correctly outside Daala.
      The fact that chroma distortion is added to the distortion of AV1's RDO is
      also critical since Daala's distortion function is applied to luma only
      and chroma continues to use MSE.
      
      Change-Id: If35fdd3aec7efe401f351ba1c99891ad57a3d957
      7a428ba2
  25. 20 Jan, 2017 1 commit
  26. 19 Jan, 2017 2 commits