Recherche avancée

Sur d’autres sites (2461)

• FFmpeg recorded video is dark with Xvfb and chrome headless on Centos 7

  15 juin 2023, par narsy4

  I am trying to do a video recording of headless chrome session on Centos 7 (Amazon EC2 instance) using ffmpeg. I have installed ffmpeg, Xvfb and google chrome on the machine. I started Xvfb on :99, verified the display using xdpyinfo and started chrome. However when I run the ffmpeg cmd to capture the video (no errors in ffmpeg debug logs), the output is dark with a X sign in the centre of video screen. What am I doing wrong here ? Any help is appreciated.

  


  **Xvfb and chrome commands**
Xvfb :99 -screen 0 1920x1080x24 &
export DISPLAY=:99
google-chrome --headless --disable-gpu --no-sandbox --start-maximized --window-size=1920x1080 https://www.google.com



  


  **FFmpeg command**
ffmpeg -video_size 1920x1080 -framerate 30 -f x11grab -i :99 -loglevel debug -pix_fmt yuv420p /tmp/video.mp4



  


  dark video

  


  Searched for and read a few threads on this, but couldn't get it to work.

  


• Installing FFMPEG on my EC2 instance takes too long ; what am I doing wrong ? [closed]

  24 août 2023, par Shaban Khawar

  I found a good article on how to download onto my EC2 instance : link
Here's the issue. It takes forever ! It takes around 15 minutes to install.

  


  I'm running my EC2 instance on Amazon Linux 2023, so no apt-get for me.
Also my EC2 instance is created by my EB environment, so if auto-scaling occurs, my instance will be terminated and a new one will be created. I can set .ebextensions to run all the commands to install FFMPEG again but as mentioned above, it takes forever ! That means for 15-20 minutes, my server is down because of one dependancy. I feel like I'm going about this the wrong way, so any advice is appreciated.

  


• Extracting Thumbnail URL from Video URL

  29 août 2023, par Deepak Sangle

  So, I have a video URL in an Amazon S3 bucket. I must extract the thumbnail image (preview image) from the video URL. The catch is that I do not have the luxury to download or upload the complete video since it takes time. My approach is something like this.
I first created a blob object containing the first few seconds of that video using something like this.

  


    const chunkSize = 1024*1024; // 1MB
  const start = 0;
  const end = chunkSize - 1;
  const rangeHeader = `bytes=${start}-${end}`;
  
  const VideoResponse = await axios.get(url, { 
    responseType: 'arraybuffer',
    headers: { Range: rangeHeader }
  });

  const chunkData = VideoResponse.data; 
  
  const videoBlob = new Blob([chunkData], { type: 'video/mp4' });
  const videoUrl = URL.createObjectURL(videoBlob);
  console.log({chunkData, videoBlob, videoUrl});


  


  Console gives me something like this

  


  {&#xA;  chunkData: <buffer 00="00" 1c="1c" 66="66" 74="74" 79="79" 70="70" 6d="6d" 34="34" 32="32" 69="69" 73="73" 6f="6f" 61="61" 76="76" 63="63" 31="31" 02="02" bd="bd" ab="ab" 6c="6c" 68="68" 64="64" c7="c7" 1048526="1048526" more="more" bytes="bytes">,&#xA;  videoBlob: Blob { size: 1048576, type: &#x27;video/mp4&#x27; },&#xA;  videoUrl: &#x27;blob:nodedata:764fce87-792f-47e8-bc3e-15921ee5787f&#x27;&#xA;}&#xA;</buffer>

  


  Now, there are many options after this. I am trying to download this blob object by converting it into a file. After I successfully convert it into an mp4 file, I can easily take a screenshot of it using fluent-ffmpeg package. However, I couldn't convert it into a file. If I try to do something like this

  


    const BlobResponse = await axios({
    method: 'GET',
    url: videoUrl,
    responseType: 'stream'
  });

  const file = fs.createWriteStream(`video.mp4`);
  BlobResponse.data.pipe(file);
  
  file.on("error", (error) => {
    console.log(`There was an error writing the file. Details: $ {error}`);
  });

  file.on('finish', () => {
    file.close();
    console.log("File downloaded successfully");
  });


  


  But it throws me the error AxiosError: Unsupported protocol blob:

  


  I also tried to directly convert the blob object to a jpeg object but to no avail.

  


  I have searched complete StackOverflow, and I have not found a single solution working so please let me know why those are not working or if there is any new solution to do this.