Recherche avancée

Sur d’autres sites (6180)

• Revision 7fd5d8e6a4 : Fix horz loopfilter loops If count was greater than 1 the src pointer would be

  18 juillet 2013, par Frank Galligan

  Changed Paths :
   Modify /vp9/common/arm/neon/vp9_loopfilter_neon.asm

  
  
  Fix horz loopfilter loops

  If count was greater than 1 the src pointer would be off on
  the second loop.

  Change-Id : I8e09037e68dc4ae92076a8067f7b6dacbbef8263

• google function : TypeError : Cannot read property 'name' of undefined

  9 décembre 2019, par Juggernaught

  I’m trying to use Google cloud function to transcode videos from a storage bucket and outputs into another bucket using code posted online but with a few adjustments.

  but I get the following error :

  TypeError : Cannot read property ’name’ of undefined at transcodeVideo
  (/srv/index.js:17:56) at /worker/worker.js:825:24 at at
  process._tickDomainCallback (internal/process/next_tick.js:229:7)

  Index.js

  const {Storage} = require('@google-cloud/storage');
  
  const projectId = 'cc18-223318';
  
  const storage = new Storage({
  
      projectId: projectId,
  
  });
  
  const ffmpegPath = require('@ffmpeg-installer/ffmpeg').path;
  
  const ffmpeg = require('fluent-ffmpeg');
  
  
  
  const transcodedBucket = storage.bucket('2400p');
  
  const uploadBucket = storage.bucket('inpuut');
  
  ffmpeg.setFfmpegPath(ffmpegPath);
  
  
  
  exports.transcodeVideo = function transcodeVideo(event, callback) {
  
    const file = event.data;
  
    // Ensure that you only proceed if the file is newly created, and exists.
  
    if (file.metageneration !== '1' || file.resourceState !== 'exists') {
  
      callback();
  
      return;
  
    }
  
  
  
    // Open write stream to new bucket, modify the filename as needed.
  
    const remoteWriteStream = transcodedBucket.file(file.name.replace('.webm', '.mp4'))
  
      .createWriteStream({
  
        metadata: {
  
          metadata: file.metadata, // You may not need this, my uploads have associated metadata
  
          contentType: 'video/mp4', // This could be whatever else you are transcoding to
  
        },
  
      });
  
  
  
    // Open read stream to our uploaded file
  
    const remoteReadStream = uploadBucket.file(file.name).createReadStream();
  
  
  
    // Transcode
  
    ffmpeg()
  
      .input(remoteReadStream)
  
      .outputOptions('-c:v libx264') // Change these options to whatever suits your needs
  
      .outputOptions('-c:a copy')
  
      .outputOptions('-vf "scale=4800:2400"')
  
      .outputOptions('-b:a 160k')
  
      .outputOptions('-b:a 160k')
  
      .outputOptions('-x264-params mvrange=511')
  
      .outputOptions('-f mp4')
  
      .outputOptions('-preset slow')
  
      .outputOptions('-movflags frag_keyframe+empty_moov')
  
      .outputOptions('-crf 18')
  
      // https://github.com/fluent-ffmpeg/node-fluent-ffmpeg/issues/346#issuecomment-67299526 
  
      .on('start', (cmdLine) => {
  
        console.log('Started ffmpeg with command:', cmdLine);
  
      })
  
      .on('end', () => {
  
        console.log('Successfully re-encoded video.');
  
        callback();
  
      })
  
      .on('error', (err, stdout, stderr) => {
  
        console.error('An error occured during encoding', err.message);
  
        console.error('stdout:', stdout);
  
        console.error('stderr:', stderr);
  
        callback(err);
  
      })
  
      .pipe(remoteWriteStream, { end: true }); // end: true, emit end event when readable stream ends
  
  };

• how to us google ML kit Selfie segmentation in flutter to overlay segmented user on top of chosen image or video ? [closed]

  4 avril 2024, par Amr

  I'm working on implementing a virtual background feature using the Google ML Kit selfie segmentation Flutter plugin. The goal is to separate the user from the background live from the camera feed, overlay the user on a chosen image or video, and then create a new video combining these elements.

  


  Currently, I'm successfully obtaining the mask using the plugin. However, I'm concerned about performance issues if I directly copy pixels to achieve the overlay effect, especially in real-time scenarios.

  


  I'm looking for alternative approaches or best practices to efficiently achieve this functionality without sacrificing performance. Any insights or suggestions on how to approach this would be greatly appreciated.

  


  I thought about using ffmpeg but not sure how to take the user pixels from the camera feed, I am not a flutter developer so not sure how to do so.