Recherche avancée

Sur d’autres sites (11404)

• lavc/vvc_refs : Define VVC_FRAME_FLAG* to h header

  28 octobre 2024, par Fei Wang

  lavc/vvc_refs : Define VVC_FRAME_FLAG* to h header
  So that hardware decoder can use the flags too.
  Signed-off-by : Fei Wang <fei.w.wang@intel.com>
  

  • [DH] libavcodec/vvc/refs.c
  • [DH] libavcodec/vvc/refs.h

• Adding trimming option to Youtube-dl audio script

  22 septembre 2020, par Jim Jamil

  This is the current script, it's a Windows batch file that prompts for a Youtube url and then downloads the best audio in m4a. It's basically cobbled together and uses aria2 to manage the download.

  


  @echo off
SETLOCAL ENABLEDELAYEDEXPANSION
(set /p var1="Url? " && youtube-dl -f bestaudio[ext=m4a] --external-downloader aria2c --external-downloader-args "-j 16 -s 16 -x 16 -k 5M" --restrict-filenames -o "%%(title)s.%%(ext)s" --add-metadata --embed-thumbnail !Var1!)
ENDLOCAL
pause


  


  After asking for the url, I want to also prompt the user to input the start and end times to trim the audio, which would be done by ffmpeg post download.

  


  Something like :

  


  ffmpeg -i file.m4a -ss 00:00:20 -to 00:00:40 -c copy file-2.m4a


  


  Based on this : https://unix.stackexchange.com/questions/182602/trim-audio-file-using-start-and-stop-times/302469#302469

  


  The beginning and end times would need to be variables set by user input in 00:00:00 format, but not sure how to add the ffmpeg post-processing at the end or how it would all fit together. I want to add this trimming feature to remove some of the preamble on podcasts and get straight to the guest part of the show.

  


  The --embed-thumbnail is optional, and won't work anyway unless Atomic Parsley is present. FFmpeg often has trouble with Album Art anyway so I usually just use -vn on the final output file.

  


• FFmpeg output stopping early when piping through stdin

  6 juillet 2022, par Null User

  I am receiving audio data encoded using Ogg Vorbis through a stream, and piping that data to FFmpeg's stdin in order to convert it into Opus. Here is the command I'm using :

  


  ffmpeg -hide_banner -i pipe: -acodec libopus -f opus -ar 48000 -ac 2 -b:a 64000 -filter_complex volume=0.2 pipe:1


  


  Unfortunately, the output audio stops a few seconds before the end of the file.

  


  Looking at the FFmpeg logs, I can see this warning appear many times during the conversion :

  


  [libopus @ 0000025190d06a00] Queue input is backward in time
    Last message repeated 9 times
size=     728kB time=00:00:45.05 bitrate= 132.4kbits/s speed=1.26x    
[libopus @ 0000025190d06a00] Queue input is backward in time
    Last message repeated 11 times
size=     793kB time=00:00:49.07 bitrate= 132.4kbits/s speed=1.24x    
[libopus @ 0000025190d06a00] Queue input is backward in time
    Last message repeated 8 times


  


  My guess is that for some reason the inputs I'm receiving are reversed, and on the last chunk of data the "end of stream" vorbis packet comes first, therefore FFmpeg stops processing the rest of the data.

  


  Saving the stream to a file and encoding the file instead of the stream doesn't cause this to happen. This seems to be exclusive to stdin streams.

  


  Is there a way to tell FFmpeg to ignore the "end of stream" signal ?