Permalink
Commits on Jan 9, 2018
  1. encode: fix read of uninitialized target_bitrate value

    tmatth committed Nov 6, 2017
    Regression from commit 5eb3f15 "Add missing file rate.c"
Commits on Jun 16, 2017
  1. dump_video: Empty video files will not be dumped.

    danesh-d authored and tdaede committed Jun 16, 2017
    Empty input video files will not be processed in 'dump_video' and
    an error message will be shown instead.
  2. daala_decode_img_out: NULL pointer check

    danesh-d authored and tdaede committed Jun 16, 2017
    Added extra check in the function 'daala_decode_img_out'
    to make sure that the passed pointers are not NULL.
Commits on Jun 8, 2017
Commits on Jun 7, 2017
  1. Error handling for missing CIEDE dependencies

    luctrudeau committed Jun 6, 2017
    rd_collect will ignore CIEDE2000 when the python packages are missing.
    An error message is shown to the user indicating which dependencies are
    required.
Commits on Jun 6, 2017
  1. Remove support for vp10 from rd_collect.sh scripts.

    negge committed Jun 2, 2017
    The vp10 codec was an experimental codec that eventually became av1.
Commits on Jun 5, 2017
  1. Add support for CIEDE2000 in rd tools

    luctrudeau committed Jun 5, 2017
    rd_collect now calls dump_ciede2000.py to output CIEDE2000 metrics.
    bd_rate.sh, rd_average.sh and rd_plot.sh also support CIEDE2000
Commits on Apr 24, 2017
  1. Update the coding tools draft.

    Timothy B. Terriberry
    Timothy B. Terriberry committed Apr 14, 2017
    Describes an way to use a smaller multiply in the dyadic partition
     function, and removes the description of the non-dyadic partition
     functions, since we don't currently plan to use those, long-term.
Commits on Mar 24, 2017
  1. ci: only do make check for travis builds

    tmatth committed Mar 24, 2017
    Rationale: make distcheck does a plain old ./configure, which forces a
    dependency on libsd2-dev and libogg.
Commits on Mar 23, 2017
  1. encode: minor refactoring

    tmatth committed Mar 23, 2017
Commits on Mar 7, 2017
  1. Merging robust and nodesync

    luctrudeau committed Mar 7, 2017
    The variables robust and nodesync appear to represent the same concept.
    In order to simplify, they are merged into nodesync. This change does
    not alter the bitstream.
Commits on Feb 23, 2017
Commits on Jan 18, 2017
  1. Getting rid of the DCT in od_compute_dist_8x8()

    jmvalin committed Jan 18, 2017
    Replacing the DCT and frequency weighting by a filter
    
       PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
    -0.0995 | -0.9381 | -0.9595 |  -1.1745 | 0.5285 |  0.1386 |    -0.1813
Commits on Jan 11, 2017
  1. Simplified clipping skip to 1

    luctrudeau committed Jan 11, 2017
    Remove double negation and added a comment explaining that this is used
    for visualization.
Commits on Dec 20, 2016
  1. Adds comment to explain nb_coeff in CfL

    luctrudeau committed Dec 20, 2016
    Added a comment because nb_coeff is not the number of coefficients in
    the current band, but the number of coefficients in the current band
    plus the number of coefficients in subsequent bands until the end of the
    block.
Commits on Dec 16, 2016
Commits on Dec 15, 2016
  1. Eliminate a sqrt in od_compute_dist_8x8().

    Timothy B. Terriberry
    Timothy B. Terriberry committed Dec 8, 2016
    No reason to use two when one will do.
    
    master-2016-12-09-7098e5 -> derf-compute_dist-fewer_sqrts4
    
      PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
    0.0000 |  0.0000 |  0.0000 |   0.0000 | 0.0000 |  0.0000 |     0.0000
Commits on Dec 8, 2016
  1. pvq_encoder: tune delta_rate for k=1

    barrbrain committed Nov 22, 2016
    Fit to collected rate data from subset3, for k = 1.
    
    daala-master-2016-11-16T13-02-41-114Z -> daala-delta-rate-k1-subset1
    
       PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
    -0.1471 | -0.8594 | -0.7224 |  -0.0740 | 0.0032 |  0.0734 |    -0.3382
    
    master-2017-11-13-4403315 -> daala-delta-rate-k1-o1f
    
       PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
    -0.1484 | -0.6323 | -0.6271 |  -0.1811 | -0.0215 | -0.0829 |    -0.3850
    
    Signed-off-by: David Michael Barr <b@rr-dav.id.au>
  2. Fully order the pvq search candidates

    barrbrain committed Sep 6, 2016
    For portable and stable sorting, break ties.
    Large differences in output were observed between
    AWCY and an OS X machine because of the platform
    qsort implementation.
    
    Signed-off-by: David Michael Barr <b@rr-dav.id.au>
Commits on Nov 30, 2016
  1. Wrong offset for the red difference plane in CIEDE

    luctrudeau authored and tdaede committed Nov 24, 2016
    Replaces A // 4 with Adiv2, because A // 4 != (H // 2) * (W // 2)
Commits on Nov 29, 2016
Commits on Nov 22, 2016
  1. Add tools/y4m2yuv.c.

    tdaede committed Nov 18, 2016
  2. tools: Add 444, 444p10 support to dump_ciede2000.py

    barrbrain authored and tdaede committed Nov 16, 2016
    Support comparing different formats, just because.
    
    Signed-off-by: David Michael Barr <b@rr-dav.id.au>
Commits on Nov 16, 2016
  1. pvq: fix artificial chroma boost with --enable-float-pvq

    tmatth committed Nov 16, 2016
    Fixes regression on --enable-float-pvq from commit
    c774e25.
    
    Caught by David Michael Barr.
    
    objective-fast1 (with/without --enable-float-pvq):
    daala-fix-float-pvq-rounding2016-11-16T143315122Z -> master-2017-11-13-4403315
    
       PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
    -0.0693 | -0.0757 |  0.1661 |  -0.0222 | -0.0493 | -0.0262 |    -0.0173
    
    subset1 (with/without --enable-float-pvq):
    daala-fix-pvq-float-rounding2016-11-16T163903172Z -> daala-master-2016-11-16T13-02-41-114Z
    
       PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
    -0.1306 | -0.0399 |  0.0673 |  -0.1358 | -0.0611 | -0.0999 |    -0.0946
Commits on Nov 15, 2016
  1. pvq: cosmetics: drop OD_ROUND32 where it is a noop

    tmatth committed Nov 15, 2016
    No change in metrics
  2. pvq: in float pvq case, mag must default to 1.0

    tmatth committed Nov 14, 2016
    Regression from c774e25
    
    Change with --enable-float-pvq:
    
            PSNR  PSNR-HVS  SSIM  FASTSSIM  CIEDE 2000
    Average 0.36 -9.80     -3.93 -7.68      3.36
    
            PSNR Cb  PSNR Cr  APSNR Y  APSNR Cb  APSNR Cr  MS SSIM
    Average 14.92    14.09    0.60     15.17     14.38    -8.12
Commits on Nov 13, 2016
Commits on Nov 9, 2016
  1. Don't mutate stride when downsampling in msssim.

    tdaede committed Nov 8, 2016
    Fixes a metric calculation error introduced in:
    1c65b80
  2. pvq: fix typo

    tmatth committed Nov 9, 2016