Recherche avancée

Sur d’autres sites (7608)

• Upgrading an existing h.264 solution of Level-4.1 to support 4K(Level-5.2)

  11 août 2015, par Codec Guy

  I am currently studying to support 4K(3840x2160) on H.264 Encoder.
  When I studied the specs of H.264, it says 4K resolution is supported in Level-5.2(XAVC).
  In order to check the implementation I encoded my full HD(1920x1080)bitstream using x264 in FFMPEG for three different levels(Level-4,4.1 and 5.2).

  Here is what I noticed
   All the three Encoded bitstreams, encoded for different levels matched pixel by pixel.
   I was wondering if the encoding is done for different levels(Level-4,4.1 and 5.2), how come there is no change in bitstream.

  • I have a solution of high 422 Level-4.1, are there any changes in the core modules like Transformation, Quantazition, CAVLC to support 4K(3840 x 2160)/XAVC.

  • or the processing of all Level-5.2 is same as Level-4.1.

  It would be very helpful, if somebody can refer me to any documents which explains the diffrence in processing of H.264/AVC and XAVC

• Revision 7418b176ce : Improve the second-level sub-pixel motion search Re-investigated the second-lev

  5 août 2015, par Yunqing Wang

  Changed Paths :
   Modify /vp9/encoder/vp9_mcomp.c

  
  
  Improve the second-level sub-pixel motion search

  Re-investigated the second-level sub-pixel motion search. Improved the
  way of choosing search points. Rewrote the second-level search code.

  At speed 0, the borg tests showed :
  1. for stdhd set, Avg PSNR gain : 0.216% ; Overall PSNR gain : 0.196% ;
  SSIM gain : 0.206%. Only 1 out of 15 clips showed PSNR loss.
  2. for derf set, Avg PSNR gain : 0.171% ; Overall PSNR gain : 0.192% ;
  SSIM gain : 0.207%. Only 3 out of 30 clips showed PSNR losses.
  Added the condition for third-point checking, namely, less points
  were checked. Speed tests showed no speed loss(Avg 0.3% speedup at
  speed 0).

  Change-Id : I6284ebb3fa7ba63be8528184c49e06757211a7f1

• Revision c3e884f5be : intrabc : Add a frame level flag For the combination of this and removing NEWDV

  20 mai 2015, par Alex Converse

  Changed Paths :
   Modify /vp9/common/vp9_entropymode.h

  
   Modify /vp9/common/vp9_onyxc_int.h

  
   Modify /vp9/decoder/vp9_decodeframe.c

  
   Modify /vp9/decoder/vp9_decodemv.c

  
   Modify /vp9/encoder/vp9_bitstream.c

  
   Modify /vp9/encoder/vp9_encodeframe.c

  
   Modify /vp9/encoder/vp9_encoder.c

  
   Modify /vp9/encoder/vp9_rdopt.c

  
  
  intrabc : Add a frame level flag

  For the combination of this and removing NEWDV from the tree :
  derflr : -0.101 screen_content : +0.053

  The bulk of the decline in screen content effecincy is from the liquify
  clip. These should be recoverable by further entropy tweaks.

  Change-Id : I9d80152b8492e60a0367c31797fb6932fb09bba9