Recherche avancée

Sur d’autres sites (2289)

• bash - Parse folder name, generate gif from images inside using ffmpeg

  16 janvier 2023, par BestMordaEver

  I have 150 folders with some nasty folder structure. Inside are animation frames helpfully named 1.png, 2.png and so on. I need to generate gifs from those images and have them named properly. Here's how I'd handle this case by case :

  


  Folder ./dragon/raw/bodies/dragon/green/old/5/down/idle has 25 .png files. I run the following commands :

  


  # pngs to gif
ffmpeg -framerate 12 -start_number 1 -i ./dragon/raw/bodies/dragon/green/old/5/down/idle/%d.png dragon-green-old-5-down-idle-intermediate.gif

# gif to boomerang gif
ffmpeg -i dragon-green-old-5-down-idle-intermediate.gif -filter_complex "[0]trim=start_frame=1:end_frame=29,setpts=PTS-STARTPTS,reverse[r];[0][r]concat=n=2:v=1:a=0,split[s0][s1];[s0]palettegen[p];[s1][p]paletteuse" dragon-green-old-5-down-idle.gif

# remove the intermediate gif
rm dragon-green-old-5-down-idle-intermediate.gif


  


  Note : relevant part of the name for gif is whatever's after /raw/bodies

  


  Question : how do I put this in a loop that would iterate over all the nested folders, parse their names and feed the result in the commands above ?

  


  Bonus question : is there a way to merge the two commands and create the boomerang gif immediately, removing the need in intermediate files ?

  


• How do I add a text for the 1st 30 seconds inside a filter_complex expression for each video part ?

  30 décembre 2022, par PirateApp

  I am generating a video grid using the following filter_complex command

  


  ffmpeg
     -i v_nimble_guardian.mkv -i macko_nimble_guardian.mkv -i ghost_nimble_guardian_1.mp4 -i nano_nimble_guardian.mkv
     -filter_complex "
         nullsrc=size=3840x2160 [base];
         [0:v] trim=start=39.117000,setpts=PTS-STARTPTS, scale=1920x1080 [upperleft];
         [1:v] trim=start=40.483000,setpts=PTS-STARTPTS, scale=1920x1080 [upperright];
         [2:v] trim=start=32.416471,setpts=PTS-STARTPTS, scale=1920x1080 [lowerleft];
         [3:v] trim=start=28.100000,setpts=PTS-STARTPTS, scale=1920x1080 [lowerright];
         [3:a] atrim=start=28.100000,asetpts=PTS-STARTPTS[outa];
         [base][upperleft] overlay=shortest=1 [tmp1];
         [tmp1][upperright] overlay=shortest=1:x=1920 [tmp2];
         [tmp2][lowerleft] overlay=shortest=1:y=1080 [tmp3];
         [tmp3][lowerright] overlay=shortest=1:x=1920:y=1080[v]
     "
     -map "[v]" -map "[outa]" -c:v libx264 -crf 17 -shortest -t 880 output4k.mkv


  


  How do I text to this video grid that will appear with a fade in at 10 seconds, stay for 30 seconds and then fade out ?

  


  

  


  What I tried ?

  


  ffmpeg
     -i v.mkv -i macko_nimble_guardian.mkv -i ghost_nimble_guardian_subtle_arrow_1.mp4 -i nano_nimble_guardian.mkv
     -filter_complex "
         nullsrc=size=1920x1080 [base];
         drawtext=text='Summer Video':enable='between(t,10,30)'[fg];
         [0:v] trim=start=39.117000,setpts=PTS-STARTPTS, scale=960x540 [upperleft];
         [1:v] trim=start=40.483000,setpts=PTS-STARTPTS, scale=960x540 [upperright];
         [2:v] trim=start=32.416471,setpts=PTS-STARTPTS, scale=960x540 [lowerleft];
         [3:v] trim=start=28.100000,setpts=PTS-STARTPTS, scale=960x540 [lowerright];
         [3:a] atrim=start=28.100000,asetpts=PTS-STARTPTS[outa];
         [base][upperleft] overlay=shortest=1 [tmp1];
         [tmp1][upperright] overlay=shortest=1:x=960 [tmp2];
         [tmp2][lowerleft] overlay=shortest=1:y=540 [tmp3];
         [tmp3][lowerright] overlay=shortest=1:x=960:y=540[v]
     "
     -map "[v]" -map "[outa]" -c:v libx264 -shortest -t '30' output2.mkv


  


  It gives me an error

  


  [Parsed_drawtext_1 @ 0x600002bdc420] Using "/System/Library/Fonts/Supplemental/Verdana.ttf"
Filter drawtext:default has an unconnected output


  


• how to convert an rstp ffmpeg mp4 buffer into ImageData inside nodejs (nestjs)

  23 décembre 2022, par distante

  I am reading an rtsp stream using ffmpeg in this way :

  


    const ffmpegArgs = [
    '-rtsp_transport',
    'tcp',
    '-i',
    this.rtspStreamPath,
    '-vcodec',
    'copy',
    '-f',
    'mp4',
    '-movflags',
    'frag_keyframe+empty_moov',
    '-reset_timestamps',
    '1',
    '-vsync',
    '1',
    '-flags',
    'global_header',
    '-bsf:v',
    'dump_extra',
    '-y',
    '-', // output to stdout
  ];
    const liveffmpeg = spawn(ffmpegPath, ffmpegArgs, {
      detached: false,
    });

    liveffmpeg.stdout.on('data', (data) => {
      console.log('data here!', data.toString());
    });


  


  Later I pipe the liveffmpeg.stdout to the response object of my NestJS controller so it can be consumed by the frontend in a video HTML Element.

  


  In the frontend I calculate this by doing a snapshot of the video element into a canvas, getting the ImageData from the canvasContext and doing my calculations after I get two or more updates.

  


  Now I need to do the exact same thing on the backend but I am not able to convert the data value from liveffmpeg.stdout.on('data', ) into an ImageData so I can use the same average function I use in the frontend.

  


  I already installed node-canvas to try to reproduce the frontend steps but to be able to draw something inside the canvas context I need an Image and no Video object can be constructed in nodejs.

  


  So, is there a way to convert the stdout data of the FFmpeg process into ImageData objects ? Or, to extract the images of each data update so I can manually draw them on the node-canvas element and get the ImageData ?

  


  Thanks !

  


  PDd : I think convert them into Uint8ClampedArray could also work since node-canvas has a function that uses that : createImageData(data: Uint8ClampedArray, width: number, height?: number) => ImageData