Recherche avancée

Sur d’autres sites (4599)

• How to create an image using ffmpeg and PHP ?

  19 décembre 2014, par webkul

  I’m using this script, but I’m not able to create the image.

  My file is here.

• ffmpeg wasm - how to take client-side created mp4 and upload it to the same server hosting the index/js files being used

  29 juillet 2022, par John Farrell

  Ok, so Im an IT guy and kind of a noob on the dev side of the fence. But I've been able to create this ffmpeg wasm page that takes a canvas and converts it to webm / and .mp4 — what i WANT to do is take the resulting .mp4 file and upload it to the server where the page/js are being served from. is this possible ? I will include my source code which is fairly simple and straight forward, I just don't know how to manipulate the resulting mp4 file that ffmpeg spits out (i realize it is happening client side) to be able to push it up to the server (maybe with aupload.php type situation ?) the solution can be html/java/php whatever, so long as it takes the mp4 output and gets it onto the server. I'd VERY MUCH appreciate a hand here.

  


  Going to try my best to properly insert the html and js. please bear with me if i've done something wrong, i've never had to -ask- a question on here, usually just look up existing answers.

  


  


  


  const { createFFmpeg } = FFmpeg;
const ffmpeg = createFFmpeg({
  log: true
});

const transcode = async (webcamData) => {
  const message = document.getElementById('message');
  const name = 'record.webm';
  await ffmpeg.load();
  message.innerHTML = 'Start transcoding';
  await ffmpeg.write(name, webcamData);
  await ffmpeg.transcode(name,  'output.mp4');
  message.innerHTML = 'Complete transcoding';
  const data = ffmpeg.read('output.mp4');

  const video = document.getElementById('output-video');
  video.src = URL.createObjectURL(new Blob([data.buffer], { type: 'video/mp4' }));
  dl.href = video.src;
  dl.innerHTML = "download mp4"
}

fn().then(async ({url, blob})=>{
    transcode(new Uint8Array(await (blob).arrayBuffer()));
})

function fn() {
var recordedChunks = [];

var time = 0;
var canvas = document.getElementById("canvas");

return new Promise(function (res, rej) {
    var stream = canvas.captureStream(60);

    mediaRecorder = new MediaRecorder(stream, {
        mimeType: "video/webm; codecs=vp9"
    });

    mediaRecorder.start(time);

    mediaRecorder.ondataavailable = function (e) {
        recordedChunks.push(event.data);
        // for demo, removed stop() call to capture more than one frame
    }

    mediaRecorder.onstop = function (event) {
        var blob = new Blob(recordedChunks, {
            "type": "video/webm"
        });
        var url = URL.createObjectURL(blob);
        res({url, blob}); // resolve both blob and url in an object

        myVideo.src = url;
        // removed data url conversion for brevity
    }

// for demo, draw random lines and then stop recording
var i = 0,
tid = setInterval(()=>{
  if(i++ > 20) { // draw 20 lines
    clearInterval(tid);
    mediaRecorder.stop();
  }
  let canvas = document.querySelector("canvas");
  let cx = canvas.getContext("2d");
  cx.beginPath();
  cx.strokeStyle = 'green';
  cx.moveTo(Math.random()*100, Math.random()*100);
  cx.lineTo(Math.random()*100, Math.random()*100);
  cx.stroke();
},200)

});
}

  


  &#xA;&#xA;&#xA;<code class="echappe-js">&lt;script src=&quot;https://unpkg.com/@ffmpeg/ffmpeg@0.8.1/dist/ffmpeg.min.js&quot; defer&gt;&lt;/script&gt;&#xA;&lt;script src='http://stackoverflow.com/feeds/tag/canvas2mp4.js' defer&gt;&lt;/script&gt;&#xA;&#xA;&#xA;&#xA;here is a canvas
  &#xA;
  &#xA;&#xA;here is a recorded video of the canvas in webM format
  &#xA;
  &#xA;&#xA;&#xA;here is a transcoded mp4 from the webm above CLIENT SIDE using ffmpeg
  &#xA;
  &#xA;&#xA;
  
  &#xA;&#xA;
  

  


  


  



• ALSA lib pcm.c:7963 :(snd_pcm_recover) underrun occurred with MoviePy

  25 janvier 2017, par mdornfe1

  So I have a bit of an unusual use case for MoviePy. I have a 40 core computer, and on each core I’m creating a moviepy object, having that object connect to an mp4 file, then having each object decompress random images from the mp4 file. I’m doing this so I can use the images to train a neural network. It actually works really except I’m getting this warning

  ALSA lib pcm.c:7963:(snd_pcm_recover) underrun occurred

  I can’t figure out its source. Here is a basic code example

  from moviepy.editor import VideoFileClip
  
  from queue import Queue, deque
  
  import threading, multiprocessing
  
  import numpy as np
  
  
  
  def decompress_worker(training_data:Queue, frame_idxs:Queue, video_file_path:str):
  
      """
  
      Eache of these workers will pop off one integer from frame_idxs.
  
      It will then decompress that frame number from vs. Finally
  
      it will store the decompressed image in training_data.
  
      """
  
      vs = VideoFileClip(video_file_path)
  
      fps = vs.fps
  
  
  
      while frame_idxs.qsize() > 0:
  
          frame_number = frame_idxs.get()
  
          frame = vs.get_frame( frame_number / fps )
  
          training_data.put(frame)
  
  
  
  
  
  num_cores = multiprocessing.cpu_count()
  
  frame_idxs = Queue()
  
  training_data = Queue()
  
  frame_idxs.queue = deque(np.random.randint(0, 10000, size = 1000))
  
  
  
  decompress_threads = []
  
  for n in range(num_cores):
  
      decompress_thread = threading.Thread(target=decompress_worker, 
  
          args=(training_data, frame_idxs, video_file_path))
  
      decompress_threads.append(decompress_thread)
  
      decompress_thread.start()
  
  
  
  [thread.join() for thread in decompress_threads]

  I’m guessing it’s coming from me reading data from the mp4 file to fast ? It would be nice if I could just suppress the warning. Is there a way to do that ?