Recherche avancée

Sur d’autres sites (4604)

• How do I broadcast live audio in Node.js ?

  20 juin 2020, par Yousef Alaqra

  I'm trying stream live audio to a wide range of clients in a web browser.

  



  My current solution :

  



  Dotnet core 3.1 console application

  



  

  1. receive the audio data over UDP

  


  2. trimming the first 28 bytes of each received packet

  


  3. and send the processed packet over UDP.

  


  



  Node JS

  



  

  1. execute a Ffmepg as a child process to receive audio data packets
over UDP from the console app, and encode each packet to audio WAV
format

  


  2. Pipe out the result of the child process into a GET HTTP endpoint response

  


  



  Browser

  



  

  1. HTML audio element with source value equals to the node js GET
endpoint

  


  



  Problem :

  



  The solution is giving a good result, but only for one device(one to one), which is not what I want to achieve.

  



  I've tried many solutions to make it applicable to a wide range of devices, such as using working threads and forking a child process, but none of them changes the result.

  



  I believe that I've to make some changes to the node js implementation, so here I'll share it with you, hoping to get a clue to solve the problem.

  



  var express = require("express");
var app = express();
var children = require("child_process");

var port = 5001;
var host = "192.168.1.230";

app.listen(port, host, () => {
  console.log("Server running at http://" + host + ":" + port + "/");
});

app.get('/stream', (req, res) => {
  const ffmpegCommand = "ffmpeg";
  var ffmpegOptions =
    "-f s16le -ar 48000 -ac 2 -i udp://192.168.1.230:65535 -f wav -";

  var ffm = children.spawn(ffmpegCommand, ffmpegOptions.split(" "));

  res.writeHead(200, { "Content-Type": "audio/wav; codecs=PCM" });
  ffm.stdout.pipe(res);
});


  



  If someone interested to see the full implementation, please let me know.

  


• How to broadcast live audio in node js (1 to many)

  19 juin 2020, par Yousef Alaqra

  I'm trying stream live audio to a wide range of clients in a web browser.

  



  My current solution :

  



  Dotnet core 3.1 console application

  



  

  1. receive the audio data over UDP

  


  2. trimming the first 28 bytes of each received packet

  


  3. and send the processed packet over UDP.

  


  



  Node JS

  



  

  1. execute a Ffmepg as a child process to receive audio data packets
over UDP from the console app, and encode each packet to audio WAV
format

  


  2. Pipe out the result of the child process into a GET HTTP endpoint response

  


  



  Browser

  



  

  1. HTML audio element with source value equals to the node js GET
endpoint

  


  



  Problem :

  



  The solution is giving a good result, but only for one device(one to one), which is not what I want to achieve.

  



  I've tried many solutions to make it applicable to a wide range of devices, such as using working threads and forking a child process, but none of them changes the result.

  



  I believe that I've to make some changes to the node js implementation, so here I'll share it with you, hoping to get a clue to solve the problem.

  



  var express = require("express");
var app = express();
var children = require("child_process");

var port = 5001;
var host = "192.168.1.230";

app.listen(port, host, () => {
  console.log("Server running at http://" + host + ":" + port + "/");
});

app.get('/stream', (req, res) => {
  const ffmpegCommand = "ffmpeg";
  var ffmpegOptions =
    "-f s16le -ar 48000 -ac 2 -i udp://192.168.1.230:65535 -f wav -";

  var ffm = children.spawn(ffmpegCommand, ffmpegOptions.split(" "));

  res.writeHead(200, { "Content-Type": "audio/wav; codecs=PCM" });
  ffm.stdout.pipe(res);
});


  



  If someone interested to see the full implementation, please let me know.

  


• Add custom metadata tag to video file without altering anything else

  9 juin 2020, par salgarji

  I'm modifying DASH-formatted segments, and my aim is to add extra information to the video segment using metadata tags.

  



  I'm using ffmpeg to record video from my webcam and store segments named like : init-video0-0.mp4 for the initial and video-0-0-N.mp4 for the following, where N is the segment number (beginning at N=1).

  



  For ffmpeg to 'understand' video files, they must have a playable video file format, for that purpose I concat initial segment with Nth segment :

  



  cat init-video-0-mp4 video0-0-N.mp4 > video0-0-N_mod.mp4


  



  Once I have a proper/correct file, I can use ffmpeg to add tags.
I've tried the following command :

  



  ffmpeg -y -i /path/to/file.mp4 -movflags use_metadata_tags -metadata customtag='whatEverYouWantToAdd' /path/to/new/file.mp4


  



  My problem is that the ffmpeg modifies the file changing other information regarding start, duration, progressive and fragmented flags...

  



  I've used this tool : https://download.tsi.telecom-paristech.fr/gpac/mp4box.js/filereader.html, to analyze my files in a more intuitive way :

  



  

  



  Do you have any clue on how to add tags without altering DASH segments in any other way ? Even if it requires using another tool different from ffmpeg.

  



  Thank you in advance !