Recherche avancée

Sur d’autres sites (14779)

• How to Generate accurate thumbnails with FFMPEG

  5 juillet 2023, par Caloux Fld

  I am using FFmpeg for generating thumbnails map.

  


  I need to generate a thumbnail for every second. Here is code i use on Windows :

  


  ffmpeg -i "C:\temp\MyVideo.mp4" -vf "drawtext=fontfile=../Fonts/calibri.ttf: text=%%`
`{n}: r=25: x=(w-tw)/2: y=h-lh-122: fontsize=295: fontcolor=Red:   box=0:` `boxcolor=Black,fps=1/1" -s 400x300 -vcodec ppm -f image2pipe - | montage -  -`
    `tile 10x -geometry 200x+0+0 C:\temp\Thumbnailer.jpg`


  


  Code Work's Fine , generate a thumbnails map with piping FFMPEG+Imagemagick (montage).
In ffmpeg i burn the frame number to check accuracy.

  


  Only problem is FFMPEG start generating at frames 12, 37, 62, 87...(video is 25 fps).

  


  How to start generating thumbnails at first frame #1 accurate ?

  


  In FFMPEG i try options like -ss 00:00:00.0 but FFMPEG start always at frame 12.
Only solution i found is to generate all frames (25fps) and post process to keep frames number 1,25,50... but very slow post process...

  


  My FFMPEG version : ffmpeg version N-88153-ga4743d2574-Reino

  


• Ffmpeg outputting empty mp4 files for video concatenation command [closed]

  25 janvier 2024, par nonsequiter

  When I run the following Ffmpeg command it produces empty mp4 files as an output.

  


  The command takes segments of a video as inputs, crops each input at particular locations, and then concatenates them (for complex filter see appendices) :

  


  ffmpeg -ss 551.389000 -t 47.292417 -i ./testo.mp4 -ss 598.681417 -t 34.217583 -i ./testo.mp4 
-y -filter_complex "" -map "[video-output]" -map "[audio-output]"
-vcodec libx264 -r 59.94005994005994 -b:v 6890k -movflags faststart ./out.mp4


  


  I have used this command for thousands of videos, but for this particular video it causes this empty mp4 files. Depending on the specific regions of the video that I specify as inputs, sometimes the command works without issues. The inputs supplied for this video are valid, as the video is 15mins or so.

  


  Here are the relevant output logs for the video (see appendices for all other input/output) :

  


  [vost#0:0/libx264 @ 0xfae9410] No filtered frames for output stream, trying to initialize anyway.
[aost#0:1/aac @ 0xfaeab60] No filtered frames for output stream, trying to initialize anyway.
frame=    0 fps=0.0 q=0.0 size=       0kB time=N/A bitrate=N/A speed=N/A    
[mp4 @ 0xfae88d0] Starting second pass: moving the moov atom to the beginning of the file
[out#0/mp4 @ 0xfae87d0] video:0kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: unknown
[out#0/mp4 @ 0xfae87d0] Output file is empty, nothing was encoded(check -ss / -t / -frames parameters if used)
frame=    0 fps=0.0 q=0.0 Lsize=       0kB time=N/A bitrate=N/A speed=N/A    
[libx264 @ 0xfae9790] final ratefactor: 17.79
[aac @ 0xfaeaee0] Qavg: nan


  


  Main insights from outputs :

  


  

  • The output tbn is 1000k (usually 60k)

  


  • "No filtered frames for output stream" for both video and audio outputs.

  


  


  Things I have tried :

  


  

  • I have a feeling its to do with the video frame-rate, so I tried using several permutations of commands like fps_mode, video_track_timescale, copytb, and enc_time_base.

  


  • Unsetting output fps

  


  • Changing input locations (this works)

  


  • Removing the complex filter makes no difference (details not removed for history)

  


  


  Appendices

  


  Here is the ffprobe input for the video, which was downloaded from YouTube :

  


  Input #0, mov,mp4,m4a,3gp,3g2,mj2, from '/tmp/video.mp4':
Metadata:
  major_brand     : isom
  minor_version   : 512
  compatible_brands: isomiso2avc1mp41
  encoder         : Lavf60.16.100
Duration: 00:14:31.44, start: 0.000000, bitrate: 939 kb/s
Stream #0:0[0x1](und): Video: h264 (High) (avc1 / 0x31637661), yuv420p(tv, bt709, progressive), 1920x1080 [SAR 1:1 DAR 16:9], 6848 kb/s, 59.94 fps, 59.94 tbr, 60k tbn (default)
  Metadata:
    handler_name    : ISO Media file produced by Google Inc.
    vendor_id       : [0][0][0][0]
Stream #0:1[0x2](eng): Audio: opus (Opus / 0x7375704F), 48000 Hz, stereo, fltp, 105 kb/s (default)
  Metadata:
    handler_name    : SoundHandler
    vendor_id       : [0][0][0][0]


  


  Ffmpeg output format, note the very high tbn value (usually this is 60k) :

  


  Output #0, mp4, to './out.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf60.16.100
  Stream #0:0: Video: h264 (avc1 / 0x31637661), yuv420p(progressive), 608x1080 [SAR 1:1 DAR 76:135], q=2-31, 6890 kb/s, 59.94 fps, 1000k tbn
    Metadata:
      encoder         : Lavc60.31.102 libx264
    Side data:
      cpb: bitrate max/min/avg: 0/0/6890000 buffer size: 0 vbv_delay: N/A
  Stream #0:1: Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 128 kb/s
    Metadata:
      encoder         : Lavc60.31.102 aac


  


  Complex filter :

  


  [0:a]asetpts=PTS-STARTPTS[audio-input-0]; # set pts for audio of input 1
[0:v]setpts=PTS-STARTPTS[post-label-0]; # set pts for video of input 2
[post-label-0]crop=608:1080:461:0[post-crop-0]; # crop input 1
[1:a]asetpts=PTS-STARTPTS[audio-input-1]; # set pts for audio of input 1
[1:v]setpts=PTS-STARTPTS[post-label-1]; # set pts for video of input 2
[post-label-1]crop=608:1080:539:0[post-crop-1]; # crop input 2
[post-crop-0][post-crop-1]concat=n=2:v=1:a=0[post-merge]; # concat video streams
[audio-input-0][audio-input-1]concat=n=2:v=0:a=1[audio-merged]; #concat audio streams
[post-merge]scale=608:1080,setsar=1[post-scale]; # scale video output
[post-scale]null[video-output]; # relabel
[audio-merged]anull[audio-output] # relabel


  


  Full command log can be found here : https://pastebin.com/5CpnaUvg

  


• ffmpeg overlay voice on a song with fade in and fade out

  22 octobre 2016, par Mehmed

  I have a fade in and fade out problem and used below code, but not completely resolve.
  I have a voice as voice.mp3 name with voice_length seconds length and a song that biggest from voice.
  I want mix with song from start_mix_time time.
  When voice start volume should be 0.2 and when voice end, volume return to 1.0.

  For example, if i have a voice by 10 s length and a song, song start playing and at position 3 s, starting to fade out to vol 0.2 and then, at 5 s, voice start over song and after 10 seconds, song fade in to vol 1 and play to end.

  Here is a sample :

  ffmpeg -i song1.mp3 -i voice2.mp3  -filter_complex "[0]asplit[a][b]; \
  
          [a]atrim=duration=voice_length,volume='1-max(0.25*(t-start_mix_time-2),0)':eval=frame[pre]; \
  
          [b]atrim=start=start_mix_time,asetpts=PTS-STARTPTS[song]; [song][1]amix=inputs=2:duration=first:dropout_transition=2[post]; \
  
          [pre][post]concat=n=2:v=0:a=1[mixed]" \
  
          -map "[mixed]" output.mp3

  @Mulvya