libaom-3.12.0

[Douglas R. Reno]

New minor version

I discovered this while looking at Arch's package updates this morning.

[zeckma]

This update skips 3.11.0 and 3.10.0.

Changes:

2025-02-10 v3.12.0
  This release includes new codec interfaces, compression efficiency and
  perceptual improvements, speedup and memory optimizations, and bug
  fixes. This release is ABI compatible with the last release.

  Five internal functions (aom_free, aom_malloc, aom_wb_bytes_written,
  aom_wb_write_bit, aom_wb_write_literal) that were exported by mistake
  are no longer exported from the libaom shared library. The removal of
  these internal functions from the ABI is a bug fix and does not break
  ABI compatibility.

  Acknowledgments: The image quality optimizations in the new tuning
  mode AOM_TUNE_IQ were originally developed for SVT-AV1-PSY by
  Cole Ogaard, Gianni Rosato, Julio Barba, and Zakaria Djebrouni.

  - New Features
    * New tuning mode AOM_TUNE_IQ (image quality) for the
      AOME_SET_TUNING codec control (--tune=iq) in all-intra mode. The
      feature detection macro AOM_HAVE_TUNE_IQ, if defined, indicates
      that AOM_TUNE_IQ is available. The image quality optimizations in
      AOM_TUNE_IQ were developed by using the SSIMULACRA 2 metric for
      guidance and validated with subjective visual quality checks.
    * New value 6 for the AV1E_SET_DELTAQ_MODE codec control
      (--deltaq-mode): use modulation for all intra using Variance
      Boost. Variance Boost is a variance adaptive quantization
      implementation that modulates qindex depending on the ratio of
      low-variance to high-variance 8x8 subblocks within a 64x64
      superblock, as well as the actual variance of the subblocks
      themselves.
    * New value 3 for the AV1E_SET_ENABLE_CDEF codec control
      (--enable-cdef): Enable CDEF adaptively based on frame qindex.
    * In all-intra mode, the AOME_SET_SHARPNESS codec control now also
      sets the loop_filter_sharpness syntax element in the bitstream.
      Larger values increasingly reduce how much the filtering can
      change the sample values on block edges to favor perceived
      sharpness.
    * In all-intra mode, the default value of the AV1E_SET_QM_MIN codec
      control is decreased to 4, and the default value of the
      AV1E_SET_QM_MAX codec control is increased to 10. The default
      values in good-quality and realtime modes remain unchanged (5 and
      9, respectively).

  - Compression Efficiency Improvements
    * Tuning mode AOM_TUNE_IQ improves image compression efficiency on
      the CLIC dataset by up to 12% for the same SSIMULACRA 2 score, up
      to 14% for the same DSSIM score, and up to 17% for the same
      Butteraugli score.
    * ~3% BD-rate gains for speed 11 VGA camera mode.
    * ~5% BD-rate gains for speed 11 on scroll clips screen mode.

  - Perceptual Quality Improvements
    * Adjust temporal filter strength for better visual quality.
    * RTC screen: visual quality improvements for scrolling and for
      scene/slide changes.
    * RTC camera mode: visual quality improvements for speed 11 VGA.

  - Speedup and Memory Optimizations
    * Optimize the Arm Neon implementation of the loop filter functions
      with an average uplift of 15 - 25% in microbenchmarks.
    * Add the CDEF optimization for RISC-V.
    * Help the compiler generate better vectorized code for variance
      calculation and warped motion in generic CPU builds.
    * Make several arrays const.

  - Other Improvements
    * Binary size reduction: 1 - 2% compared with last release, with
      CONFIG_REALTIME_ONLY enabled, CONFIG_AV1_DECODER and
      CONFIG_AV1_HIGHBITDEPTH disabled.
    * Build: compile source files in parallel under MSVC.

  - Bug Fixes
    * Fix bug where metadata added with aom_img_add_metadata was lost
      when frame scaling was used.
    * Bug b:383306740: RTC: Fix to issues with scrolling for screen
      content.
    * Bug b:382465458: RTC: Fix to artifact for grayscale input.
    * Bug b:380247338: RTC: Fix to encode_time spikes on scene/slide
      changes.
    * RTC: Fix to rate correction factor update for VBR screen mode.
      https://groups.google.com/a/aomedia.org/g/av1-discuss/c/nJxECdg-7P8
    * Bug b:378401081: RTC: Fix to cyclic refresh update for external RC
      (rate control).

2024-10-24 v3.11.0
  This release includes perceptual quality improvements, binary size reduction
  under certain configurations and many bug fixes. This release changes the
  default encoder configuration for the AOM_USAGE_REALTIME mode. This release
  is ABI compatible with the last release.

  - Perceptual Quality Improvements
    * Visual quality improvements for RTC screen content
      * Higher quality on scene or slide changes
      * Faster quality ramp-up for static content after scene change
      * Quality improvements for screen content with active maps

  - Speedup
    * Added and improved Neon SIMD paths for dynamic frame scaling with ~1.5%
      overall encoding speedup.

  - Other Improvements
    * Binary size reduction: 10% compared with last release, with
      CONFIG_REALTIME_ONLY enabled, CONFIG_AV1_DECODER and
      CONFIG_AV1_HIGHBITDEPTH disabled.
    * Update default_extra_cfg for CONFIG_REALTIME_ONLY to provide proper RTC
      defaults settings
    * Change the default valuess of the following encoder config options in the
      AOM_USAGE_REALTIME mode:
      *  rc_overshoot_pct and rc_undershoot_pct are changed from 25 to 50
      *  rc_buf_sz is changed from 6000 to 1000
      *  rc_buf_initial_sz is changed from 4000 to 600
      *  rc_buf_optimal_sz is changed from 5000 to 600

  - Bug Fixes
    * aomedia:363016123: rtc: Fix setting of intra-only frame for
      set_ref_frame_config and add checks
    * aomedia:42302583: rtc: Fix for artifacts for screen with active_maps
    * b:365088425: rtc: Allow for lower-QP on scene/slide changes
    * b:367285382: Fix to encoder quality max-out too early for screen
    * b:362936830: rtc: Allow QP to decrease more aggressively for static
      content
    * b:361617762: Clamp the calculation of sb64_target_rate to INT_MAX
    * chromium:362123224: rtc-svc: Reset ref_map_idx for references not used
    * chromium:367892770: Fix to possible integer overflow in reset_rc
    * webrtc:369633254: rtc-svc: Fix to reset ref_idx for svc
    * Fix exit condition in rate correction update

2024-08-27 v3.10.0
  This release includes new codec interfaces, compression efficiency and
  perceptual improvements, speedup and memory optimizations and many bug
  fixes. This release is ABI compatible with the last release.

  The definitions of the internal macros AOM_INLINE and AOM_FORCE_INLINE
  have been removed from the public header aom/aom_integer.h.

  - New Features
    * New codec controls:
      * AV1E_SET_AUTO_TILES
      * AV1E_GET_HIGH_MOTION_CONTENT_SCREEN_RTC
      * AV1E_SET_POSTENCODE_DROP_RTC: Post encode frame drop feature.
      * AV1E_SET_MAX_CONSEC_FRAME_DROP_MS_CBR
    * New key-value pair for aom_codec_set_option():
      * "auto-tiles": equivalent to the new codec control
        AV1E_SET_AUTO_TILES.

  - Deprecated Features
    * Deprecated codec control:
      * AV1E_SET_MAX_CONSEC_FRAME_DROP_CBR: Use the new codec control
        AV1E_SET_MAX_CONSEC_FRAME_DROP_MS_CBR instead.
    * The sframe_mode field in the aom_codec_enc_cfg_t struct is not
      implemented.

  - Compression Efficiency Improvements
    * BD-rate gain of 0.7 - 1.3% (by enabling global motion tool) for
      speed 5 and speed 6 with ~5% encode time increase.
    * RTC speed 11 video: ~3-5% BD-rate gain for VGA and QVGA.

  - Perceptual Quality Improvements
    * RTC quality improvements for slide changes and scrolling content.

  - Speedup and Memory Optimizations
    * RTC screen content speedups:
      * ~2x speedup for high motion content for speed 11.
      * ~2x speedup on key frame coding for speed >= 10.
    * Arm: Significant uplifts in speed in this release (vs v3.9.1) have
      come from tuning the various convolutions according to filter size
      (doing 8-tap when only 2-tap is required is inefficient) and also
      deploying Armv8.6 USMMLA instructions in 6-tap and 12-tap standard
      bitdepth convolutions.
      * Standard bitdepth RTC:
        * speed 5: +5%
        * speed 6: +4%
        * speed 7: +5%
        * speed 8: +4%
        * speed 9: +6%
        * speed 10: +6%
      * Standard bitdepth VoD:
        * speed 0: +9%
        * speed 1: +12%
        * speed 2: +9%
        * speed 3: +3%
        * speed 4: +3%
        * speed 5: -9% (expected due to global motion changes)
        * speed 6: -3% (expected due to global motion changes)
      * High bitdepth VoD:
        * speed 0: +4%
        * speed 1: +19%
        * speed 2: +23%
        * speed 3: +1%
        * speed 4: +1%
        * speed 5: -8% (expected due to global motion changes)
        * speed 6: -3% (expected due to global motion changes)
      * Standard bitdepth 2x1 horizontal super-resolution/scaling
        encoding: +101%

  - Other Improvements
    * Reduce bit rate overshoot on slide content.

  - Bug Fixes
    * rtc: Bug fix for active_maps with sb_size=128.
    * b:343429036: rtc: Fix source_sad setting near boundary.
    * Fix to QP for temporal enhancement after key frame.
    * b:343429192: rtc: Condition QP adjustment on rc->q_1/2_frame > 0.

[Douglas R. Reno]

Currency has been updated

[Douglas R. Reno]

Fixed at 170a2e3ee6374be89463358503603b492bc4f416

Test stats updated at d3ceea064bf03d77d95aada519d1102b9f0925c7