All pull requests

Pull requests

• Remove unused speed feature prune_tx_type_est_rd and related code.

  cla-signed
  #5722·urvangjoshi opened Nov 7, 2026Contributor·

  ·
  2

• Skip TIP NEWMV in the partition search dry pass at speed >= 5

  cla-signed
  #5721·xinxiong-xx opened Nov 6, 2026Contributor·

  ·
  1

• (Follow Up) TIP-first: skip more normal encodes of frames that end as TIP output

  cla-signed
  #5720·xinxiong-xx opened Nov 6, 2026Contributor·

  ·

• Prune modes in partition search dry pass

  cla-signed
  #5719·chengchen-google opened Nov 6, 2026Contributor·

  ·
  5

• Optimize AVX2 DIP matrix multiply and add AVX-512 kernels

  cla-signed
  #5715·amd-ananthkm opened Nov 6, 2026·

  ·
  3

• Skip dry-pass compound reference MV stack setup

  cla-signed
  #5714·chengchen-google opened Nov 5, 2026Contributor·

  ·
  3

• Track CI OBU presence per extended layer so G130 reordering does not …

  cla-signed
  #5712·SMPaluri opened Nov 5, 2026Contributor·

  ·

• Add NEON for fwd_stxfm and inv_stxfm

  cla-signed
  #5711·jjustiss-apple opened Nov 5, 2026Contributor·

  ·
  3

• Disable palette mode search for natural content at speed >= 1

  cla-signed
  #5710·kslu-aom opened Nov 5, 2026Contributor·

  ·
  6

• Use more aggressive speed features at speed 1

  cla-signed
  #5709·chengchen-google opened Nov 5, 2026Contributor·

  ·
  6

• Restrict IntraBC DVs to the row-mt guaranteed-ready area

  cla-signed
  #5703·gjin-rgb opened Nov 4, 2026Contributor·

  ·
  1

• Optimizations to warp affine kernels

  cla-signed
  #5701·amd-nithyavs opened Nov 4, 2026Contributor·

  ·
  2

• Add NEON for av2_highbd_dr_prediction_z1_idif, z2_idif and z3_idif

  cla-signed
  #5700·jjustiss-apple opened Nov 4, 2026Contributor·

  ·
  2

• Update reference frame info in tpl

  cla-signed
  #5698·jingninghan opened Nov 4, 2026Contributor·

  ·
  3

• Derive DOH at the encoder

  cla-signed
  #5692·jkeui opened Nov 3, 2026Contributor·

  ·
  1

  2

• Add --sframe-replace-kf encoder parameter

  cla-signed
  #5691·jkeui opened Nov 3, 2026Contributor·

  ·
  1

• Add functionality update one reference frame by switch frame

  cla-signed
  #5690·jkeui opened Nov 3, 2026Contributor·

  ·
  1

• Remove encoder-side limitation for restricted switch frames

  cla-signed
  #5688·jkeui opened Nov 3, 2026Contributor·

  ·
  2

• Limit simple-translation DRL search to one candidate

  cla-signed
  #5678·ChenFeng-Bristol opened Nov 2, 2026·

  ·
  3

• Remove redundant computation in motion, inter and tx searches

  cla-signed
  #5677·xinxiong-xx opened Nov 2, 2026Contributor·

  ·
  3

• Prune inter mode candidates with an MLP

  cla-signed
  #5670·chengchen-google opened Oct 31, 2026Contributor·

  ·
  1

• Implement VCWG-G130 properly

  cla-signed
  #5668·wantehchang opened Oct 30, 2026Member·

  ·
  12

• Prune skip mode DRL candidates by model RD

  cla-signed
  #5663·ChenFeng-Bristol opened Oct 30, 2026·

  ·

• Price the refined intra winner with its own luma mode cost

  cla-signed
  #5657·gjin-rgb opened Oct 29, 2026Contributor·

  ·
  1

• Early termination of partition search in dry pass based on best mode

  cla-signed
  #5653·jadhuran opened Oct 29, 2026Contributor·

  ·
  2