Vvenc Versions Save

bugfix release:

• fix performance regression on ARM by reverting SIMDe to use SSE41 rather than AVX2 code

bugfix release:

• fix performance regression on ARM by reverting SIMDe to use SSE41 rather than AVX2 code

Over v1.10.0:

• finalized inter-frame parallelization (previously FPP)
• added vectorized DQ implementation
• improved SCC detection by reducing false positives
• added an affine speedup based on Pejman et al., ICIP 2023.
• overall speedups:
  • 6% for fast/slow/slower
  • 12% for medium, 25%+ for multi-threaded execution with long intra periods

Over v1.11.0-rc1:

• fixed some valgrind issues
• marked some MSVC issues as fixed in 19.39

• finalized inter-frame parallelization (previously FPP)
• added vectorized DQ implementation
• improved SCC detection by reducing false positives
• added an affine speedup based on Pejman et al., ICIP 2023.
• overall speedups:
  • 6% for fast/slow/slower
  • 12% for medium, 25%+ for multi-threaded execution with long intra periods

• first IDR per default at POC31 instead of POC0 to align with other GOPs
• improved RC, especially for high rates and low activity scenes, especially for HDR
• enabled spatial sub-sampling for 2pRC in preset faster
• number of MCTF references adaptively increased
• reduced the efficiency gap for dependent frame parallelism (still experimental, as not harmonized with RC)
• around 0.5% BDR gains for all presets
• many other minor cleanups and changes

• first IDR per default at POC31 instead of POC0 to align with other GOPs
• improved RC, especially for high rates and low activity scenes, especially for HDR
• enabled spatial sub-sampling for 2pRC in preset faster
• number of MCTF references adaptively increased
• reduced the efficiency gap for dependent frame parallelism (still experimental, as not harmonized with RC)
• around 0.5% BDR gains for all presets
• many other minor cleanups and changes

Bugfix release:

• fixed problems with 8-bit encoding, especially for 10-bit input
• improvements to rate control, especially spatial sub-sampling for first pass
• added medium_lowDecEnergy as a built-in preset
• many other fixes

Bugfix release:

• fixed problems with 8-bit encoding, especially for 10-bit input
• improvements to rate control, especially spatial sub-sampling for first pass
• added medium_lowDecEnergy as a built-in preset
• many other fixes

• added rate capping during rate control (must be at least 1.5x of target rate)
• extensive improvements and cleanups to rate control, especially single pass (improved efficiency by 1-2%)
• speedup of first pass, especially for preset faster (around 15% overall runtime), using spatio-temporal downsampling
• unified MCTF activation for low- and high-rate encoding
• various speedups, fixes, and cleanups

Over 1.9.0-rc2:

• reduce single pass rate control deviation from given target bit-rate, when intra period is not a multiple of gop size (see #296)

Over 1.9.0-rc1:

• propagating missing data required for efficient RC in fast/faster presets when using QPA (bugfix)

Over 1.8.0:

• added rate capping during rate control (must be at least 1.5x of target rate)
• extensive improvements and cleanups to rate control, especially single pass (improved efficiency by 1-2%)
• speedup of first pass, especially for preset faster (around 15% overall runtime), using spatio-temporal downsampling
• unified MCTF activation for low- and high-rate encoding
• various speedups, fixes, and cleanups