Recherche avancée

Sur d’autres sites (10294)

• Command is running different from expected when i use it trought Python

  17 septembre 2021, par Gustavo Marinho

  I have a code where i download a youtube video as 3gpp and convert it to a mp3, i need to use FFmpeg to do this, and it work well when using both cmd and powershell, but, when i tried to run the same command in Python, it didin't work at all.

  


  This is my command :

  


  ffmpeg -i  C:\YTDownloads\CurrentAudio.3gpp C:\YTDownloads\CurrentAudio.mp3

  


  I tried :

  


  subprocess.call(r'%SystemRoot%\system32\WindowsPowerShell\v1.0\powershell.exe ffmpeg -i  C:\YTDownloads\CurrentAudio.3gpp C:\YTDownloads\CurrentAudio.mp3', shell=True)

  


  subprocess.run(["ffmpeg","-i","C:\YTDownloads\CurrentAudio.3gpp","C:\YTDownloads\CurrentAudio.mp3]")


  


  os.system('powershell ffmpeg -i  C:\YTDownloads\CurrentAudio.3gpp C:\YTDownloads\CurrentAudio.mp3')


  


  subprocess.run([
    'ffmpeg',
    '-i', os.path.join(parent_dir, f"{newname}.3gpp"),
    os.path.join(parent_dir, f"{newname}.mp3")
]) 


  


  subprocess.call('C:\Windows\System32\powershell.exe ffmpeg -i  C:\YTDownloads\CurrentAudio.3gpp C:\YTDownloads\CurrentAudio.mp3', shell=True)


  


  all of them return some type of error, in some of them it returns that ffmpeg isn't a recognized as a internal command, in others it says that the system can't find the specified path, but none of them works, even thought it works perfectly when i use the exactly same command on cmd/powershell.

  


  sorry for my bad english :3

  


• Animation with gnuplot : only one cruve plotted in an animation whereas 2 expected

  26 mai 2022, par stefan

  I have the following script which is expected to produce the animations of 2 curves :

  


  #!/bin/bash

for i in {1..397}; do
gnuplot -p <<-EOFMarker
set terminal png;
set output "pic$i.png";
set title "power spectrum";
set xlabel "scale (k)";
set ylabel "P(k)";
set key top left;
set grid;
set ytics out nomirror;
set xtics out nomirror;
set logscale x;
set logscale y;
set format x "10^{%L}";
set yrange [0:30000];
plot "CAMB-1.3.5/matter_camb$i" u 1:2 w l;
replot "EFTCAMB_v3_beta/matter_eftcamb$i" u 1:2 w l;
EOFMarker
done

# Build movie with ffmpeg
ffmpeg -start_number 1 -i pic%d.png  movie.mpeg


  


  Every works fine excepted the fact that only one curve is plotted in animation (CAMB-1.3.5) :

  


  Here is an example of frame :

  


  

  


  Why the "replot" command is not taken into account in the generated image ?

  


• How to concatenate two MP4 files, which require http basic Authorization : Bearer , using ffmpeg ?

  8 juillet 2023, par Jeff Strongman

  Hello dear ffmpeg experts ! 🧠 🎯

  


  I ran the following command, which worked perfectly :

  


  ffmpeg -protocol_whitelist https,concat,tls,tcp -i "concat :https://dash.akamaized.net/akamai/bbb_30fps/bbb_30fps_1280x720_4000k/bbb_30fps_1280x720_4000k_0.m4v|https://dash.akamaized.net/akamai/bbb_30fps/bbb_30fps_1280x720_4000k/bbb_30fps_1280x720_4000k_1.m4v|https://dash.akamaized.net/akamai/bbb_30fps/bbb_30fps_1280x720_4000k/bbb_30fps_1280x720_4000k_2.m4v" -c:v copy -vframes 180 -y Movie_of_6_seconds.mp4

  


  I followed the recommended solution of the following post :
How to concatenate two MP4 files using FFmpeg ?

  


  You can execute the command on your local computer and see that it should run just fine...

  


  I used 3. concat protocol, which does indeed concat init + progressive segments

  


  However... when every segment on a server I refer to, is password protected, it fails with 401 Unauthorized, even though I added the following line :
-headers "Authorization : Bearer bas64user:password" , before specifying the -i "concat :...".

  


  It seems to me... that the headers don't pass to the concat command inside of the input of ffmpeg... and it simply ignores them. When I used the same -headers command, on a single file, without concat, it passed the authorization successfully

  


  Notes :

  


  

  • Even though every segment has a length of 120 frames (So in maximum, I could have generated 2*120 = 240 frames... I wanted a movie of 6 seconds and not 8... And by this way, to test that ffmpeg is smart enough to stop processing the whole input). To do that, I used -vframes 180, where 180 / 30 (FPS) = 6 seconds

  


  • I used -c:v copy, to get without re-encoding, only the video part (No audio !)

  


  • I used -y to override existing file...

  


  • 0.m4v, is the init file ! it is a small file, that has metadata of the original video which was produced with mpeg-dash

  


  • 1.m4v and 2.m4v, are the progressive segments

  


  


  Is there a way, to pass the http basic headers (Authorization : Bearer) to all of the chained files ?

  


  Like :

  


  

  • Via a json content type on the ffmpeg request

  


  • Or user:password@video_segment (Although... it seems to me it's not a header ?)

  


  • Somehow specify header inside the concat command ?

  


  


  I don't want to first download all files and then get rid of the password protected... as it both takes ridiculous time & other resources... and I would like to record from a segment that is "endless", meaning a camera that keeps streaming data.

  


  Thanks in advance 🙏🏻,

  


  FFmpeg noobie 🙈