Recherche avancée

Sur d’autres sites (13343)

• Ogg file is not detected as audio mime in mediawiki (Ogg generated by ffmpeg)

  12 janvier 2015, par JoeyD

  I have an ogg file that I converted with ffmpeg from a 3gp file. The audio file is supposed to go into Mediawiki for other users to listen to. However uploading the file always gives it the mime-type application/ogg and no player is shown for that file.

  Are there specific settings that I have to follow to have it detect as audio in Mediawiki ? Or should i look to ffmpeg ? I use the -codec:a libvorbis flag for ffmpeg

• ffmpeg filter_complex doesn't work with text file input

  5 avril 2023, par Martin

  I have a working ffmpeg command that takes 4 input files
https://file.io/DllNlkCqX6J6

  


  (2 audio, 2 images) and then renders a .mkv video, using a filter_complex string which tells ffmpeg to concat the two audio files and make a slideshow for the two images.

  


  The filter_complex string relies on the order of the input files.

  


  For example : [0:a][1:a]concat=n=2:v=0:a=1[a] in filter_complex tells ffmpeg to take the index=0 input audio file, and the index=1 input audio file, and concat them.

  


  Then, the the [2:v] part tells ffmpeg to take the index=2 image file (we start counting index at zero, so index=2 is really the 3rd file, which is '3.sidea.jpg')

  


  ffmpeg 

-r 2 -i "C:\Users\martin\Documents\album\a1 - in the beginning.mp3" 
-r 2 -i "C:\Users\martin\Documents\album\a2 - stratovarious.mp3" 
-r 2 -i "C:\Users\martin\Documents\album\3 sidea.jpg" 
-r 2 -i "C:\Users\martin\Documents\album\4 sideb.jpg" 

-filter_complex "[0:a][1:a]concat=n=2:v=0:a=1[a];[2:v]scale=w=1920:h=1920:force_original_aspect_ratio=decrease,pad=1920:1920:-1:-1:color=white,setsar=1,loop=309.95:309.95[v2];[3:v]scale=w=1920:h=1920:force_original_aspect_ratio=decrease,pad=1920:1920:-1:-1:color=white,setsar=1,loop=309.95:309.95[v3];[v2][v3]concat=n=2:v=1:a=0,pad=ceil(iw/2)*2:ceil(ih/2)*2[v]" 

-map [v] -map [a] -c:a pcm_s32le -c:v libx264 -bufsize 3M -crf 18 -pix_fmt yuv420p -tune stillimage -t 309.95 
"C:\Users\martin\Documents\album\working.mkv"


  


  This command works, but I use the same format for, say, 2k+ audio input files, I reach the terminal command line char limit. So I am trying to move my input files to a text file input_files.txt :

  


  file 'C:\Users\martin\Documents\album\a1 - in the beginning.mp3'
file 'C:\Users\martin\Documents\album\a2 - stratovarious.mp3'
file 'C:\Users\martin\Documents\album\3 sidea.jpg'
file 'C:\Users\martin\Documents\album\4 sideb.jpg'



  


  The same files, in the same order. But running my new command with this input_text.txt file :

  


  ffmpeg 

-f concat -safe 0 -i input_files.txt

-filter_complex "[0:a][1:a]concat=n=2:v=0:a=1[a];[2:v]scale=w=1920:h=1920:force_original_aspect_ratio=decrease,pad=1920:1920:-1:-1:color=white,setsar=1,loop=309.95:309.95[v2];[3:v]scale=w=1920:h=1920:force_original_aspect_ratio=decrease,pad=1920:1920:-1:-1:color=white,setsar=1,loop=309.95:309.95[v3];[v2][v3]concat=n=2:v=1:a=0,pad=ceil(iw/2)*2:ceil(ih/2)*2[v]" 

-map [v] -map [a] -c:a pcm_s32le -c:v libx264 -bufsize 3M -crf 18 -pix_fmt yuv420p -tune stillimage -t 309.95 
"C:\Users\martin\Documents\album\working.mkv"


  


  Results in an error about the filter_complex index numbers :

  


  [mp3 @ 0000025d87057700] Estimating duration from bitrate, this may be inaccurate
Input #0, concat, from 'input_files.txt':
  Duration: N/A, start: 0.000000, bitrate: 320 kb/s
  Stream #0:0: Audio: mp3, 44100 Hz, stereo, fltp, 320 kb/s
Invalid file index 1 in filtergraph description [0:a][1:a]concat=n=2:v=0:a=1[a];[2:v]scale=w=1920:h=1920:force_original_aspect_ratio=decrease,pad=1920:1920:-1:-1:color=white,setsar=1,loop=309.95:309.95[v2];[3:v]scale=w=1920:h=1920:force_original_aspect_ratio=decrease,pad=1920:1920:-1:-1:color=white,setsar=1,loop=309.95:309.95[v3];[v2][v3]concat=n=2:v=1:a=0,pad=ceil(iw/2)*2:ceil(ih/2)*2[v].



  


  Almost like it's treating the -i input_files.txt as its own index=0 file. I've tried different fixes but no luck.

  


  How can I use my filter_complex string with a text file as the source of my ffmpeg command file inputs ?

  


• dnn : change .model file format to put layer number at the end of file

  29 août 2019, par Guo, Yejun

  dnn : change .model file format to put layer number at the end of file
  currently, the layer number is at the beginning of the .model file,
  
  so we have to scan twice in python script, the first scan to get the
  
  layer number. Only one scan needed after put the layer number at the
  
  end of .model file.
  Signed-off-by : Guo, Yejun <yejun.guo@intel.com>
  
  Signed-off-by : Pedro Arthur <bygrandao@gmail.com>
  

  • [DH] libavfilter/dnn/dnn_backend_native.c
  • [DH] tools/python/convert_from_tensorflow.py