howto:ffmpeg

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision		 Previous revision
Last revisionBoth sides next revision
howto:ffmpeg [2020/12/06 17:22] – ffmpeg renamed to howto:ffmpeg andihowto:ffmpeg [2023/06/11 00:55] – [audio channel mapping / repair] andi
Line 36: Line 36:
 make install </code> make install </code>
  
-== audio channel mapping doku +== HD-Master mit Loudnorm
- +
-https://trac.ffmpeg.org/wiki/AudioChannelManipulation +
- +
- +
- +
-== silent fix+
 <code> <code>
-ffmpeg -ar 48000 -acodec pcm_s16le -f s16le -ac 2 -i <( dd if=/dev/zero ) -y -i not_silent/$file  -t 10.0 -c:v copy $file; +ffmpeg -y \ 
-</code> +-i /video/intros/XXX/intro.ts -analyzeduration 20000000 \ 
- +-ss 1858.48 -t 3550.88 -i /video/fuse//XXX/YYY/ZZZ/uncut.ts -analyzeduration 20000000 \ 
-Oder "schöner" und ohne knacken:+-i /video/intros/XXX/outro.ts \ 
 +-filter_complex ' 
 +  [0:a:0] loudnorm=i=-16:print_format=summary [norm-0-0]; 
 +  [1:a:0] loudnorm=i=-16:print_format=summary [norm-1-0]; 
 +  [2:a:0] loudnorm=i=-16:print_format=summary [norm-2-0];
  
-<code> +  [0:v:0] [norm-0-0] [1:v:0] [norm-1-0] [2:v:0] [norm-2-0] concat=n=3:v=1:a=1 [vd] [audio0_mix]; 
-ffmpeg -i not_silent/$file -filter_complex "aevalsrc=c=2c:s=48000:0|0 [a]" -map 0:v -c:v copy -map "[a]" -c:a $audiocodec -shortest $outputfile+  [vd] yadif=mode=0:parity=0, hqdn3d [v] 
 +' -analyzeduration 20000000/ 
 +-map '[v]' -c:v:0 libx264 -pix_fmt yuv420p -crf:0 23 -profile:0 high -level:0 4.1 
 +-map '[audio0_mix]' -c:a:0 aac -b:a:0 192k -ar:a:0 48000 \ 
 +-aspect 16:9 -movflags faststart -f mp4 /video/tmp/XXX/ZZZ-hd.mp4
 </code> </code>
-$audiocodec z.B. mp2 bei Intros 
-== metadata fix 
  
-=== map metadata and fix spezific entry+=== HD-Master mit Loudnorm, Pause rausschneiden 
 + - Zeitstempel = *Sekunden.Frames* 
 + - -t gibt Duration an, nicht Ende!
  
 <code> <code>
-ffmpeg -i br0ken_licence/$file \ +ffmpeg -y \ 
-  -map_metadata 0 \ +-i /video/intros/XXX/intro.ts -analyzeduration 20000000 \ 
-  -metadata:s:a:0 LICENSE="Licensed to the public under https://creativecommons.org/licenses/by/2.0/de/ - http://media.ccc.de" \ +-ss 1858.48 -t 3550.88 -i /video/fuse//XXX/YYY/ZZZ/uncut.ts -analyzeduration 20000000 \ 
-  -c:a copy $file +-ss 1858.48 -t 3550.88 -i /video/fuse//XXX/YYY/ZZZ/uncut.ts -analyzeduration 20000000 \ 
-</code>+-i /video/intros/XXX/outro.ts \ 
 +-filter_complex ' 
 +  [0:a:0] loudnorm=i=-16:print_format=summary [norm-0-0]; 
 +  [1:a:0] loudnorm=i=-16:print_format=summary [norm-1-0]; 
 +  [2:a:0] loudnorm=i=-16:print_format=summary [norm-2-0]; 
 +  [3:a:0] loudnorm=i=-16:print_format=summary [norm-3-0];
  
-=== copy metadata from another file +  [0:v:0] [norm-0-0] [1:v:0] [norm-1-0] [2:v:0] [norm-2-0] [3:v:0] [norm-3-0] concat=n=4:v=1:a=1 [vd] [audio0_mix]; 
- +  [vd] yadif=mode=0:parity=0, hqdn3d [v] 
-<code> +' -analyzeduration 20000000/ 
-ffmpeg -i $metadata_source -i $video_source -map 1 -c copy -map_metadata 0 $output_file+-map '[v]' -c:v:0 libx264 -pix_fmt yuv420p -crf:0 23 -profile:0 high -level:0 4.1 
 +-map '[audio0_mix]' -c:a:0 aac -b:a:0 192k -ar:a:0 48000 \ 
 +-aspect 16:9 -movflags faststart -f mp4 /video/tmp/XXX/ZZZ-hd.mp4
 </code> </code>
  
-== one input, multiple format outputs  
- 
-1. h264/aac flv rtmp output 
-2. theora/vorbis ogg output 
- 
-<code> 
-ffmpeg -re -i 29c3-5266-de-en-proximax_telex_flashproxy_h264.mp4 -threads 0 -pix_fmt yuv420p -profile:v baseline -preset fast -tune zerolatency -c:v libx264 -strict -2 -c:a aac \ 
--f flv rtmp://localhost:1935/stream/saal1 \ 
--map 0 -c:v libtheora -c:a libvorbis -f ogg - | oggfwd localhost 8000 … 
-</code> 
- 
-== Audiospur via Backup Audioaufnahme reparieren und in MP4 packen 
- 
-Sprache deu-backup im Tracker anlegen. 
-Nochmal neu rausrendern lassen. 
- 
-einzelne Audiopuren vor oder nach Auphnic so umwandeln das Audio-Schnitt Tool wie z.B. Reaper was damit anfangen kann: 
- 
-  ffmpeg -i 5100-hd-audio2-auphonic.m4a -vn -c:a copy -bsf:a aac_adtstoasc 5100-hd-audio2-auphonic-echt.m4a 
-  ffmpeg -i 5100-hd-audio2-auphonic.m4a -vn -c:a copy -bsf:a aac_adtstoasc 5100-hd-audio2-auphonic-echt.m4a 
- 
- 
-Fertige Audiospur in Reaper rausrechnen, z.B. als Flac 
- 
-Audiospur wieder in Vortrag MP4 einpacken: 
- 
-<code> 
-ffmpeg -threads 0 -i /video/encoded/fossgis16/5100-hd-primary-audio-only.mp4 -i 5100-audio-fixed.flac -map 0:0 -map 1:0 -map_metadata 0 -c:v copy -c:a:0 libfdk_aac -vbr 4 -ar:a:0 48000  -metadata:s:a:0 language=deu 5100-hd-fixed2.mp4  
-</code> 
- 
-Alte Version 
-<code> 
-ffmpeg -threads 0 -i 5100-hd.mp4  -i 5100-audio-fixed.flac -map 0:0 -map 1:0 -map_metadata 1 5100-hd-fixed.mp4 
-</code> 
- 
- 
-Anmerkung: Geht auch mit MTS Datei (Transportstream mit H264 Video) der so aus der Consumer-Kamera auf die SD Karte fällt. 
- 
-Siehe auch: https://trac.ffmpeg.org/wiki/AudioChannelManipulation 
-== Audiospur(en) und Videospur vertauscht 
-<code> 
-ffmpeg -i 5115-hd-broken.mp4 -map 0:v -map 0:a -c copy -map_metadata 0 5115-hd.mp4 
-</code> 
- 
-== Translated Transcoding 
- 
-<code> 
-ffmpeg -y -i rtmp://live.lan.c3voc.de/stream/mastermaster -aspect 16:9 -threads:0 0 -c:v libx264 -filter_complex '[0:v] yadif, scale=1920:1080 [v]; [0:a:0] pan=stereo|c0=FL [native]'  -map '[native]'  -map '[v]' -maxrate:v:0 2800k -bufsize:v:0 8192k -crf:0 18 -pix_fmt:0 yuv420p -profile:v:0 main -g:v:0 25 -preset:v:0 veryfast -ac 1 -y -f flv rtmp://live.lan.c3voc.de/push/master_translated_hd 
-</code> 
 == encoding snippets == encoding snippets
  
Line 133: Line 93:
   -metadata title="<title>" \   -metadata title="<title>" \
     -metadata album="<album>" \     -metadata album="<album>" \
-    -metadata copyright="Licensed to the public under http://creativecommons.org/licenses/by-sa/3.0/" \+    -metadata copyright="This work is licensed under a Creative Commons Attribution 4.0 International License. https://creativecommons.org/licenses/by/4.0/" \
   <output-file-name>   <output-file-name>
 </code> </code>
Line 146: Line 106:
   -c:a libfdk_aac -b:a 192k -aspect 16:9 -metadata title="<title>" \   -c:a libfdk_aac -b:a 192k -aspect 16:9 -metadata title="<title>" \
     -metadata album="<album>" \     -metadata album="<album>" \
-    -metadata copyright="Licensed to the public under http://creativecommons.org/licenses/by-sa/3.0/" \+    -metadata copyright="This work is licensed under a Creative Commons Attribution 4.0 International License. https://creativecommons.org/licenses/by/4.0/" \
     -f mp4 -movflags faststart <output-file-name>     -f mp4 -movflags faststart <output-file-name>
 </code> </code>
Line 198: Line 158:
  -c:a mp3 -b:a 32k -ar:a 22050 -f mp4 -movflags faststart \  -c:a mp3 -b:a 32k -ar:a 22050 -f mp4 -movflags faststart \
  /video/tmp/..../cut-101-proxy.mp4  /video/tmp/..../cut-101-proxy.mp4
 +</code>
 +
 +
 +== one input, multiple format outputs 
 +
 +1. h264/aac flv rtmp output
 +2. theora/vorbis ogg output
 +
 +<code>
 +ffmpeg -re -i 29c3-5266-de-en-proximax_telex_flashproxy_h264.mp4 -threads 0 -pix_fmt yuv420p -profile:v baseline -preset fast -tune zerolatency -c:v libx264 -strict -2 -c:a aac \
 +-f flv rtmp://localhost:1935/stream/saal1 \
 +-map 0 -c:v libtheora -c:a libvorbis -f ogg - | oggfwd localhost 8000 …
 +</code>
 +
 +
 +== Translated Transcoding
 +
 +<code>
 +ffmpeg -y -i rtmp://live.lan.c3voc.de/stream/mastermaster -aspect 16:9 -threads:0 0 -c:v libx264 -filter_complex '[0:v] yadif, scale=1920:1080 [v]; [0:a:0] pan=stereo|c0=FL [native]'  -map '[native]'  -map '[v]' -maxrate:v:0 2800k -bufsize:v:0 8192k -crf:0 18 -pix_fmt:0 yuv420p -profile:v:0 main -g:v:0 25 -preset:v:0 veryfast -ac 1 -y -f flv rtmp://live.lan.c3voc.de/push/master_translated_hd
 +</code>
 +
 +
 +== Audiospur via Backup Audioaufnahme reparieren und in MP4 packen
 +
 +Sprache deu-backup im Tracker anlegen.
 +Nochmal neu rausrendern lassen.
 +
 +einzelne Audiopuren vor oder nach Auphnic so umwandeln das Audio-Schnitt Tool wie z.B. Reaper was damit anfangen kann:
 +
 +  ffmpeg -i 5100-hd-audio2-auphonic.m4a -vn -c:a copy -bsf:a aac_adtstoasc 5100-hd-audio2-auphonic-echt.m4a
 +  ffmpeg -i 5100-hd-audio2-auphonic.m4a -vn -c:a copy -bsf:a aac_adtstoasc 5100-hd-audio2-auphonic-echt.m4a
 +
 +
 +Fertige Audiospur in Reaper rausrechnen, z.B. als Flac
 +
 +Audiospur wieder in Vortrag MP4 einpacken:
 +
 +<code>
 +ffmpeg -threads 0 -i /video/encoded/fossgis16/5100-hd-primary-audio-only.mp4 -i 5100-audio-fixed.flac -map 0:0 -map 1:0 -map_metadata 0 -c:v copy -c:a:0 libfdk_aac -vbr 4 -ar:a:0 48000  -metadata:s:a:0 language=deu 5100-hd-fixed2.mp4 
 +</code>
 +
 +Alte Version
 +<code>
 +ffmpeg -threads 0 -i 5100-hd.mp4  -i 5100-audio-fixed.flac -map 0:0 -map 1:0 -map_metadata 1 5100-hd-fixed.mp4
 +</code>
 +
 +
 +Anmerkung: Geht auch mit MTS Datei (Transportstream mit H264 Video) der so aus der Consumer-Kamera auf die SD Karte fällt.
 +
 +Siehe auch: https://trac.ffmpeg.org/wiki/AudioChannelManipulation
 +
 +
 +== Audiospur(en) und Videospur vertauscht
 +<code>
 +ffmpeg -i 5115-hd-broken.mp4 -map 0:v -map 0:a -c copy -map_metadata 0 5115-hd.mp4
 +</code>
 +
 +
 +
 +
 +== audio channel mapping / repair
 +
 +https://trac.ffmpeg.org/wiki/AudioChannelManipulation
 +
 +
 +=== mono mixdown
 +
 +```ffmpeg -i 27-hd-orig.mp4 -codec:v copy -map_metadata 0 -ac 1 -ab 128k 27-hd.mp4 ```
 +
 +
 +=== silent fix
 +<code>
 +ffmpeg -ar 48000 -acodec pcm_s16le -f s16le -ac 2 -i <( dd if=/dev/zero ) -y -i not_silent/$file  -t 10.0 -c:v copy $file;
 +</code>
 +
 +Oder "schöner" und ohne knacken:
 +
 +<code>
 +ffmpeg -i not_silent/$file -filter_complex "aevalsrc=c=2c:s=48000:0|0 [a]" -map 0:v -c:v copy -map "[a]" -c:a $audiocodec -shortest $outputfile
 +</code>
 +$audiocodec z.B. mp2 bei Intros
 +
 +
 +== metadata fix
 +
 +=== map metadata and fix specific entry
 +
 +<code>
 +ffmpeg -i br0ken_licence/$file \
 +  -map_metadata 0 \
 +  -metadata:s:a:0 LICENSE="This work is licensed under a Creative Commons Attribution 4.0 International License. https://creativecommons.org/licenses/by/4.0/" \
 +  -c:a copy $file
 +</code>
 +
 +=== copy metadata from another file
 +
 +<code>
 +ffmpeg -i $metadata_source -i $video_source -map 1 -c copy -map_metadata 0 $output_file
 </code> </code>
  
Line 214: Line 272:
     " -map '[v]' -c:v:0 libx264 -pix_fmt yuv420p -bufsize:0 8192k -crf:0 20 -minrate:0 100k -maxrate:0 5000k -profile:0 main -level:0 4.0 -threads 8 -map '[a]' -c:a:0 libfdk_aac -b:a:0 192k -ac:a:0 1 -ar:a:0 48000 -aspect 16:9 -f mp4 50-hd-camonly-cut.mp4     " -map '[v]' -c:v:0 libx264 -pix_fmt yuv420p -bufsize:0 8192k -crf:0 20 -minrate:0 100k -maxrate:0 5000k -profile:0 main -level:0 4.0 -threads 8 -map '[a]' -c:a:0 libfdk_aac -b:a:0 192k -ac:a:0 1 -ar:a:0 48000 -aspect 16:9 -f mp4 50-hd-camonly-cut.mp4
          
-=== HD-Master mit Loudnorm 
-<code> 
-ffmpeg -y \ 
--i /video/intros/XXX/intro.ts -analyzeduration 20000000 \ 
--ss 1858.48 -t 3550.88 -i /video/fuse//XXX/YYY/ZZZ/uncut.ts -analyzeduration 20000000 \ 
--i /video/intros/XXX/outro.ts \ 
--filter_complex ' 
-  [0:a:0] loudnorm=i=-16:print_format=summary [norm-0-0]; 
-  [1:a:0] loudnorm=i=-16:print_format=summary [norm-1-0]; 
-  [2:a:0] loudnorm=i=-16:print_format=summary [norm-2-0]; 
  
-  [0:v:0] [norm-0-0] [1:v:0] [norm-1-0] [2:v:0] [norm-2-0] concat=n=3:v=1:a=1 [vd] [audio0_mix]; 
-  [vd] yadif=mode=0:parity=0, hqdn3d [v] 
-' -analyzeduration 20000000/ 
--map '[v]' -c:v:0 libx264 -pix_fmt yuv420p -crf:0 23 -profile:0 high -level:0 4.1 
--map '[audio0_mix]' -c:a:0 aac -b:a:0 192k -ar:a:0 48000 \ 
--aspect 16:9 -movflags faststart -f mp4 /video/tmp/XXX/ZZZ-hd.mp4 
-</code> 
- 
-=== HD-Master mit Loudnorm, Pause rausschneiden 
- - Zeitstempel = *Sekunden.Frames* 
- - -t gibt Duration an, nicht Ende! 
- 
-<code> 
-ffmpeg -y \ 
--i /video/intros/XXX/intro.ts -analyzeduration 20000000 \ 
--ss 1858.48 -t 3550.88 -i /video/fuse//XXX/YYY/ZZZ/uncut.ts -analyzeduration 20000000 \ 
--ss 1858.48 -t 3550.88 -i /video/fuse//XXX/YYY/ZZZ/uncut.ts -analyzeduration 20000000 \ 
--i /video/intros/XXX/outro.ts \ 
--filter_complex ' 
-  [0:a:0] loudnorm=i=-16:print_format=summary [norm-0-0]; 
-  [1:a:0] loudnorm=i=-16:print_format=summary [norm-1-0]; 
-  [2:a:0] loudnorm=i=-16:print_format=summary [norm-2-0]; 
-  [3:a:0] loudnorm=i=-16:print_format=summary [norm-3-0]; 
- 
-  [0:v:0] [norm-0-0] [1:v:0] [norm-1-0] [2:v:0] [norm-2-0] [3:v:0] [norm-3-0] concat=n=4:v=1:a=1 [vd] [audio0_mix]; 
-  [vd] yadif=mode=0:parity=0, hqdn3d [v] 
-' -analyzeduration 20000000/ 
--map '[v]' -c:v:0 libx264 -pix_fmt yuv420p -crf:0 23 -profile:0 high -level:0 4.1 
--map '[audio0_mix]' -c:a:0 aac -b:a:0 192k -ar:a:0 48000 \ 
--aspect 16:9 -movflags faststart -f mp4 /video/tmp/XXX/ZZZ-hd.mp4 
-</code> 
 === Icecast-Audio mit ffmpeg === Icecast-Audio mit ffmpeg
  
  • howto/ffmpeg.txt
  • Last modified: 2023/06/11 00:58
  • by andi