libvpx Wigeon Duck v1.15.0-rc1

[Jerome Jiang]

Hello,

Wigeon Duck v1.15.0 has been tagged[1]. Please help testing and report issues[2].

This release is ABI compatible with the previous release.

Thanks.

[1]https://chromium.googlesource.com/webm/libvpx/+/refs/tags/v1.15.0-rc1

[2]https://g-issues.webmproject.org/issues/new?pli=1&authuser=0&component=1615215&template=2023762

Changelog:

2024-10-22 v1.15.0 "Wigeon Duck"
  This release includes new codec control for key frame filtering, more Neon
  optimizations, improvements to RTC encoding and bug fixes.

  - Upgrading:
    This release is ABI compatible with the previous release.
    Temporal filtering improvement that can be turned on with the new codec
    control VP9E_SET_KEY_FRAME_FILTERING, which gives 1+% BD-rate saving with
    minimal encoder time increase.
    libwebm is upgraded to libwebm-1.0.0.31-10-g3b63004

  - Enhancement:
    Neon optimization speed up
      1-3% speed up across speed 5 to 10 for RTC
      3% speed up for speed 0 and 1 for VoD in standard bitdepth
      3% and 7% speed up for speed 0 and 1 respectively for VoD in high bitdepth
    Scene detection is allowed for all RTC speeds (>=5)
    Support profile guided optimizations
    Delta quantization parameters for UV channels for vp8 is supported in RTC
    rate control library
    Rate control parameters are reset and maximum QP is enforced on scene
    changes in SVC when there is no inter-layer prediction

  - Bug fixes:
    Fix to Uninitialized scalar variable in `vp9_rd_pick_inter_mode_sb()`
    Fix to Integer-overflow in `resize_multistep`
    Fix to Heap-buffer-overflow in `vpx_sad64x64_avx2`
    Fix to Crash in `vpx_sad8x8_sse2`
    Fix to Assertion in `write_modes`
    Support profile guided optimizations
    Fix to Integer-overflow in `encode_frame_to_data_rate`
    Fix to Integer-overflow in `vp9_svc_check_reset_layer_rc_flag`
    Fix to core dump error from /usr/bin/tools/tiny_ssim --help
    Fix to use-of-uninitialized-value in `vp9_setup_tpl_stats`
    Fix to Undefined-shift in `vp9_cyclic_refresh_setup`
    Fix to redundant `&& __GNUC__` preproc check
    Fix to valgrind warning in EncodeAPI.OssFuzz69906
    Fix to Index-out-of-bounds in `vp8_rd_pick_inter_mode`
    Fix to Integer-overflow in `vp8_pick_frame_size`
    Fix to Use-of-uninitialized-value in `vpx_codec_peek_stream_info`
    Fix to log clutters with the message "Warning: Desired height too large"
    Fix to Integer-overflow in `vp9_svc_adjust_avg_frame_qindex`
    Fix to integer overflows caused by huge target bitrate, frame rate, or
    g_timebase numerator or denominator
    Fix to missing license headers
    Fix to build failure for Android Armv7
    Fix to integer overflows in image helpers
    Fix to Integer-overflow in `vp9_calc_iframe_target_size_one_pass_cbr`
    Fix to Heap-buffer-overflow in `vp9_pick_inter_mode`
    Fix to Segv in `vp9_multi_thread_tile_init`
    Fix to Use-of-uninitialized-value in `vp9_row_mt_sync_mem_dealloc`
    Fix to Crash in `mbloop_filter_vertical_edge_c`
    Fix to Check failed in CheckUnwind
    Fix to Heap-buffer-overflow in `write_modes_b` and `vpx_write`
    Fix to Possible signed integer overflow found in `vpx_codec_encode`
    Fix to build conflicts between Abseil and libaom/libvpx in Win ARM64 builds
    Fix to build failures on aarch64
    Fix to Data race in libvpx ARM NEON
    Fix to Heap-buffer-overflow in `scale_plane_1_to_2_phase_0`
    Fix to integer overflow in `encode_mb_row`
    Fix to Floating-point-exception in `vp8_pick_frame_size`
    Fix to Heap-buffer-overflow in `vp9_enc_setup_mi`
    Fix to build failure with --target=arm64-win64-vs17
    Fix to heap-buffer-overflow write in `vpx_img_read()`
    Fix to C vs armv8-linux-gcc encode mismatches for `y4m_360p_10bit_input`
    Fix to Null-dereference READ in `ml_predict_var_rd_partitioning`
    Fix to Heap-buffer-overflow in `vpx_scaled_2d_ssse3`
    Fix to Crash in `convolve_horiz`
    Fix to Ill in `vpx_scaled_2d_ssse3`
    Fix to Global-buffer-overflow in `cost_coeffs`