Recherche avancée

Sur d’autres sites (6806)

• Stream real-time video flux in HTML video tag

  26 septembre 2018, par c.censier

  I want to stream a real-time video flux that come from udp into a HTML video tag.
  I made some research but I got a lot of informations and I struggle to have a clear overview of what I can do and what I can’t.

  The video flux use H.264 and AAC codecs, MP4 container and has a 3840x2160 (4K) resolution. I’d like to play it on Chrome (latest version).

  As I understand from now, HTML video tag can natively read H.264/AAC videos. I made it work with the video direclty on my server (I’m using Meteor JS + React).

  I learnt to use FFmpeg to stream an udp flux read by VLC player, and then I used FFserver (I know it’s deprecated) to create an HTTP flux also read by VLC but not by the HTML video tag.

  So... my question is : is HTML video can natively read video stream from HTTP ?

  I’ve seen a lot of discussions about HLS and DASH, but I didn’t understand if (and why) they’re mandatory.

  I read a post about someone creating a HLS m3u8 using only FFmpeg, is it a viable solution ?

  FFserver configuration

  HTTPPort                        8090
  
  HTTPBindAddress                 0.0.0.0
  
  MaxHTTPConnections              20
  
  MaxClients                      10
  
  MaxBandwidth                    100000
  
  
  
  <feed>
  
    File                          /tmp/feed.ffm
  
    FileMaxSize                   1g
  
    ACL allow                     127.0.0.1
  
  </feed>
  
  
  
  <stream>
  
    Feed                          feed.ffm
  
    Format                        mpeg
  
    AudioCodec                    aac
  
    AudioBitRate                  256
  
    AudioChannels                 1
  
    VideoCodec                    libx264
  
    VideoBitRate                  10000      // Total random here
  
    VideoBitRateRange             5000-15000 // And here...
  
    VideoFrameRate                30
  
    VideoQMin                     1
  
    VideoQMax                     50
  
    VideoSize                     3840x2160
  
    VideoBufferSize               20000      // Not sure either
  
    AVOptionVideo                 flags +global_header
  
  </stream>

  I had to specify QMin and QMax to avoid error message but I don’t really understand what is it.

  FFmpeg command line

  ffmpeg -re -i bbb_sunflower_2160p_30fps_normal.mp4 -strict -2 -r 30 -vcodec libx264 http://localhost:8090/feed.ffm

  This work with VLC. I’m working with a file on my computer before moving to an udp stream.

• Pipe video frames from ffmpeg to canvas without loading the entire video into memory

  1er janvier 2024, par Aviato

  I am working on a project that involves frame manipulation and I decided to choose node canvas API for that. I used to work with OpenCV Python and there was a cv2.VideoCapture class that takes a video as input and prepares to read the frames of the video and we can loop through the frames one at a time without having to load all the frames at once in memory.
Now I tried a lot of ways to replicate the same using ffmpeg, i.e. trying to load frames from a video in an ordered, but "on-demand," fashion.

  


  I tried using ffmpeg as a child process to process frames and standout the frames.

  


  const spawnProcess = require('child_process').spawn,
    ffmpeg = spawnProcess('ffmpeg', [
        '-i', 'test.mp4',
        '-vcodec', 'png',
        '-f', 'rawvideo',
        '-s', '1920*1080', // size of one frame
        'pipe:1'
    ]);
ffmpeg.stdout.on('data', (data) => {
    try {
        // console.log(tf.node.decodeImage(data).shape)
        console.log(`${++i} frames read`)
        //context.drawImage(data, 0, 0, width, height)
        
        
    } catch(e) {
        console.log(e)
    } 
})


  


  The value in the console shows something around 4000 + console logs, but the video only had 150 frames, after much investigating and console logging the data, I found that was buffer data, and it's not processing it for each frame. The on-data function returns the buffer data in an unstructured way
I want to read frames from a video and process each one at a time in memory, I don't want to hold all the frames at once in memory or in the filesystem.

  


  

  


  I also want to pipe the frames in a format that could be rendered on top of a canvas using drawImage

  


• ffmpeg black screen issue for video video generation from a list of frames

  11 mai 2023, par arlaine

  I used a video to generate a list of frames from it, then I wanted to create multiple videos from this list of frames.
I've set starting and ending frames indexes for each "sub video", so for example,
indexes = [[0, 64], [64, 110], [110, 234], [234, 449]], and those indexes will help my code generate 4 videos of various durations. The idea is to decompose the original video into multiple sub videos. My code is working just fine, the video generated.

  


  But every sub video start with multiple seconds of black screen, only the first generated video (so the one using indexes[0] for starting and ending frames) is generated without this black screen part. I've tried changing the frame rate for each sub_video, according to the number of frames and things like that, but I didn't work. You can find my code below

  


  for i, (start_idx, end_idx) in enumerate(self.video_frames_indexes):
    if end_idx - start_idx > 10:
        shape = cv2.imread(f'output/video_reconstitution/{video_name}/final/frame_{start_idx}.jpg').shape
        os.system(f'ffmpeg -r 30 -s {shape[0]}x{shape[1]} -i output/video_reconstitution/{video_name}/final/frame_%d.JPG'
              f' -vf "select=between(n\,{start_idx}\,{end_idx})" -vcodec libx264 -crf 25'
              f' output/video_reconstitution/IMG_7303/sub_videos/serrage_{i}.mp4')


  


  Just the ffmpeg command

  


  ffmpeg -r 30 -s {shape[0]}x{shape[1]} -i output/video_reconstitution/{video_name}/final/frame_%d.JPG -vf "select=between(n\,{start_idx}\,{end_idx})" -vcodec libx264 -crf 25 output/video_reconstitution/IMG_7303/sub_videos/serrage_{i}.mp4