Recherche avancée

Sur d’autres sites (5589)

• FFMPEG remove step of creating overlay video in multistep process

  19 avril 2020, par Peter Azuolas

  I have a multistep process to generate a downloadable copy of a livestream with onscreen elements being burned into the video.

  



  First command generates a transparent animated overlay from a frame list in a text file.

  



  ffmpeg -loglevel panic -f concat -safe 0 -r 30 -vcodec png  -i /tmp/bottomThird/UNKNOWN/36000/bt_files.txt -r 30 -codec:v png /tmp/bottomThird/UNKNOWN/36000/d8d55cd61873959017ed8dce8a0dcb61.mov


  



  The text file contains a sequence of frames for an animation that has "opening frames" and then loop frames, essentially :

  



  /tmp/bottomThird/UNKNOWN/img/00001.png
/tmp/bottomThird/UNKNOWN/img/00002.png
/tmp/bottomThird/UNKNOWN/img/00003.png
/tmp/bottomThird/UNKNOWN/img/00004.png
/tmp/bottomThird/UNKNOWN/img/00005.png
/tmp/bottomThird/UNKNOWN/img/00006.png
/tmp/bottomThird/UNKNOWN/img/00007.png
/tmp/bottomThird/UNKNOWN/img/00008.png
/tmp/bottomThird/UNKNOWN/img/00009.png
/tmp/bottomThird/UNKNOWN/img/00010.png
/tmp/bottomThird/UNKNOWN/img/00005.png
/tmp/bottomThird/UNKNOWN/img/00006.png
/tmp/bottomThird/UNKNOWN/img/00007.png
/tmp/bottomThird/UNKNOWN/img/00008.png
/tmp/bottomThird/UNKNOWN/img/00009.png
/tmp/bottomThird/UNKNOWN/img/00010.png
...


  



  repeated until the frame count is reached (in this case 36000 frames).

  



  The resultant MOV file is in the 2gb range as it's basically just a container with the frames images copied into.

  



  The next step intermediate file is generated using multiple overlays. The MOV above, text that gets displayed after 40 frames and stays static until end, and a second overlay "watermark" style image.

  



  ffmpeg  -i /tmp/embedGenerator/51039542/51039542.mp4  -vf "movie=/tmp/embedGenerator/51039542/bt_51039542.mov[inner];movie=/tmp/embedGenerator/51039542/bt_51039542_blank.png[overlay];movie=/tmp/embedGenerator/51039542/cover.png[cover];[in][inner] overlay=(W-w)/2:H-h-15 [out];[out][overlay] overlay=(W-w)/2:H-h-15:enable='between(n,40,36000)' [fin];[fin][cover] overlay=0:0 [final]" -shortest -b:v 5000k  -f mpegts /tmp/embedGenerator/51039542/intermediate.ts


  



  I'd like the avoid the step of creating the animated MOV file and combine both steps into a single command mainly to avoid having to store such a large file temporarily.

  



  ---

  



  One alternative I've considered but not yet tried is to generate 2 MOV files. The "entry" and then the "loop" sequence and looping the second file instead of generating an complete set of frames.

  


• aacenc_tns : redo coefficient quantization and decision making

  12 septembre 2015, par Rostislav Pehlivanov

  aacenc_tns : redo coefficient quantization and decision making
  This finally (and again) gets rid of basically everything the
  
  specifications say about how TNS should be done. The main
  
  problem used to be that a single filter was used for all
  
  coefficients which despite being explicitly recommended by
  
  the specifications usually sounds wrong, therefore it’s
  
  a corner case in the current TNS implementation.
  This commit also changes the coefficient bit size, as apparently
  
  it’s better to use lower precision in case the windows are eight
  
  short. This is apparently what fdk_aac uses, looking at the bit
  
  stream and makes sense. Also the order when 8 SHORT windows happen
  
  is important as 7 was too much and according to PSNR was worse
  
  while 5 is just about correct.
  Signed-off-by : Rostislav Pehlivanov <atomnuker@gmail.com>
  

  • [DH] libavcodec/aacenc_tns.c
  • [DH] libavcodec/aacenc_tns.h

• Revision c5443fc881 : Fix decoder mismatch in sub-pixel SSSE3 intrinsic filters In 8-tap filtering, t

  23 mai 2014, par Yunqing Wang

  Changed Paths :
   Modify /vp9/common/x86/vp9_asm_stubs.c

  
   Modify /vp9/common/x86/vp9_subpixel_8t_intrin_ssse3.c

  
  
  Fix decoder mismatch in sub-pixel SSSE3 intrinsic filters

  In 8-tap filtering, to guarantee the intermediate results fit in
  16 bits, the order of accumulating the products needs to be done
  correctly, and the largest product should be added last. This
  patch fixed the problem using the method in commit "Correct ssse3
  8/16-pixel wide sub-pixel filter calculation".

  Change-Id : I79d0ad60c057b15011ece84cda9648eee0809423