feat(VoiceReceiver): Proof of concept - Video recording

examples/VoiceChannel/RecordingVideo.js

@@ -0,0 +1,49 @@
+// https://v12.discordjs.guide/voice/receiving-audio.html#basic-usage
+
+/*
+Install:
+- An Opus library: @discordjs/opus or opusscript
+- An encryption packages:
+  + sodium (best performance)
+  + libsodium-wrappers
+  + @stablelib/xchacha20poly1305
+- ffmpeg (install and add to your system environment)
+*/
+
+const { Client } = require('../../src/index');
+const client = new Client();
+
+const fs = require('fs');
+
+client.on('ready', async client => {
+  console.log(`${client.user.username} is ready!`);
+
+  const channel = client.channels.cache.get('voice_id');
+  const connection = await client.voice.joinChannel(channel, {
+    selfMute: true,
+    selfDeaf: true,
+    selfVideo: false,
+  });
+
+  const connectionStream = await connection.joinStreamConnection('user_id');
+
+  const video = connectionStream.receiver.createVideoStream('user_id', {
+    portUdp: 5004,
+    output: fs.createWriteStream('video.ts'), // Output file using MPEG-TS container
+  });
+
+  video.stream.stderr.on('data', data => {
+    console.log(`FFmpeg: ${data}`);
+  });
+
+  video.on('ready', () => {
+    console.log('FFmpeg process ready!');
+  });
+
+  // After 15s
+  setTimeout(() => {
+    video.destroy();
+  }, 15_000);
+});
+
+client.login('token');

src/client/voice/receiver/FFmpegHandler.js

@@ -5,6 +5,8 @@ const { createSocket } = require('dgram');
 const { EventEmitter } = require('events');
 const { Buffer } = require('node:buffer');
 const { Writable } = require('stream');
+const find = require('find-process');
+const kill = require('tree-kill');
 const Util = require('../../../util/Util');
 const { StreamOutput } = require('../util/Socket');
 
@@ -46,11 +48,7 @@ class FFmpegHandler extends EventEmitter {
      */
     this.ready = false;
 
-    /**
+    const stream = spawn('ffmpeg', [
-     * The FFmpeg process
-     * @type {ChildProcessWithoutNullStreams}
-     */
-    this.stream = spawn('ffmpeg', [
       '-reorder_queue_size',
       '50',
       '-err_detect',
@@ -73,9 +71,15 @@ class FFmpegHandler extends EventEmitter {
       '500000',
       '-y',
       '-f', // Specify the format
-      'matroska', // MKV format
+      'mpegts', // MKV format
       isStream ? this.outputStream.url : output,
     ]);
+
+    /**
+     * The FFmpeg process
+     * @type {ChildProcessWithoutNullStreams}
+     */
+    this.stream = stream;
     this.stream.stdin.write(sdpData);
     this.stream.stdin.end();
     this.stream.stderr.once('data', data => {
@@ -101,6 +105,17 @@ class FFmpegHandler extends EventEmitter {
     const message = Buffer.from(payload);
     this.socket.send(message, 0, message.length, this.portUdp, '127.0.0.1', callback);
   }
+
+  destroy() {
+    const ffmpegPid = this.stream.pid; // But it is ppid ;-;
+    const args = this.stream.spawnargs.slice(1).join(' '); // Skip ffmpeg
+    find('name', 'ffmpeg', true).then(list => {
+      let process = list.find(o => o.pid === ffmpegPid || o.ppid === ffmpegPid || o.cmd.includes(args));
+      if (process) {
+        kill(process.pid);
+      }
+    });
+  }
 }
 
 module.exports = FFmpegHandler;

src/client/voice/receiver/Receiver.js

@@ -59,6 +59,7 @@ class VoiceReceiver extends EventEmitter {
    * @typedef {Object} ReceiveVideoStreamOptions
    * @property {number} [portUdp] The UDP port to use for the video stream (local stream).
    * @property {string} [codec='H264'] The codec to use for encoding the video. Default is 'H264'.
+   * <info>H265 supported, but not implemented</info>
    * @property {any} [output] Additional output options, as required.
    */
 
@@ -70,9 +71,10 @@ class VoiceReceiver extends EventEmitter {
    * @param {ReceiveVideoStreamOptions} options Options.
    * @returns {FFmpegHandler} The video stream for the specified user.
    */
-  createVideoStream(user, { portUdp, codec = 'H264', output } = {}) {
+  createVideoStream(user, { portUdp, codec, output } = {}) {
     user = this.connection.client.users.resolve(user);
     if (!user) throw new Error('VOICE_USER_MISSING');
+    codec = 'H264';
     const stream = this.packets.makeVideoStream(user.id, portUdp, codec, output);
     return stream;
   }

src/util/Util.js

@@ -68,7 +68,7 @@ const payloadTypes = [
     priority: 5000,
     payload_type: 109,
     rtx_payload_type: 110,
-    encode: true,
+    encode: false,
     decode: false,
   },
 ];
@@ -951,7 +951,6 @@ class Util extends null {
       payloadType = payload.payload_type;
     } else {
       const payloadType = packet[1] > 120 ? packet[1] & 0x80 : packet[1];
-      console.log('payloadType', payloadType, packet, portUdp);
       payload = payloadTypes.find(p => p.payload_type === payloadType);
     }
     let sdpData = `o=- 0 0 IN IP4 127.0.0.1

typings/index.d.ts

@@ -1109,6 +1109,7 @@ export class FFmpegHandler extends EventEmitter {
   public sendPayloadToFFmpeg(payload: Buffer): void;
   public on(event: 'ready', listener: () => void): this;
   public once(event: 'ready', listener: () => void): this;
+  public destroy(): void;
 }
 
 export class VoiceReceiver extends EventEmitter {
@@ -1116,7 +1117,7 @@ export class VoiceReceiver extends EventEmitter {
   public createStream(user: UserResolvable, options?: { mode?: 'opus' | 'pcm'; end?: 'silence' | 'manual' }): Readable;
   public createVideoStream(
     user: UserResolvable,
-    options?: { portUdp: number; codec: VideoCodec | 'H265' | 'VP9' | 'AV1'; output: Writable | string },
+    options?: { portUdp: number; codec: 'H264'; output: Writable | string },
   ): FFmpegHandler;
 
   public on(event: 'debug', listener: (error: Error | string) => void): this;