fix(test): avoid false PCM failures in TTS tests

plugins/test/package.json

@@ -39,6 +39,7 @@
     "zod": "^3.25.76"
   },
   "dependencies": {
+    "@ffmpeg-installer/ffmpeg": "^1.1.0",
     "fastest-levenshtein": "^1.0.16",
     "vitest": "^4.0.17"
   },

plugins/test/src/tts.ts

@@ -1,16 +1,63 @@
 // SPDX-FileCopyrightText: 2024 LiveKit, Inc.
 //
 // SPDX-License-Identifier: Apache-2.0
+import ffmpegInstaller from '@ffmpeg-installer/ffmpeg';
 import type { stt } from '@livekit/agents';
 import { type AudioBuffer, initializeLogger, tokenize, tts as ttslib } from '@livekit/agents';
-import type { AudioFrame } from '@livekit/rtc-node';
+import { type AudioFrame, combineAudioFrames } from '@livekit/rtc-node';
 import { distance } from 'fastest-levenshtein';
+import { spawn } from 'node:child_process';
 import { ReadableStream } from 'stream/web';
 import { describe, expect, it } from 'vitest';
 
 const TEXT =
   'The people who are crazy enough to think they can change the world are the ones who do.';
 
+const compressedFormats = new Set([
+  'mp3',
+  'aac',
+  'ogg',
+  'flac',
+  'wav',
+  'mov,mp4,m4a,3gp,3g2,mj2',
+  'matroska,webm',
+  'mpeg',
+]);
+
+const assertPCM = async (frames: AudioFrame[]) => {
+  const frame = combineAudioFrames(frames);
+  const data = new Uint8Array(frame.data.buffer, frame.data.byteOffset, frame.data.byteLength);
+
+  const container = await new Promise<string | undefined>((resolve, reject) => {
+    const ffmpeg = spawn(ffmpegInstaller.path, [
+      '-hide_banner',
+      '-probesize',
+      '32',
+      '-analyzeduration',
+      '0',
+      '-i',
+      'pipe:0',
+    ]);
+    let stderr = '';
+
+    ffmpeg.stderr.setEncoding('utf8');
+    ffmpeg.stderr.on('data', (chunk: string) => {
+      stderr += chunk;
+    });
+    ffmpeg.stdin.on('error', () => {});
+    ffmpeg.on('error', reject);
+    ffmpeg.on('close', () => {
+      resolve(stderr.match(/Input #0, (.*), from 'pipe:0':/)?.[1]);
+    });
+
+    ffmpeg.stdin.end(data);
+  });
+
+  if (container && compressedFormats.has(container)) {
+    throw new Error(`Audio data isn't PCM (detected ${container})`);
+  }
+};
+
 const validate = async (frames: AudioBuffer, stt: stt.STT, text: string, threshold: number) => {
   const event = await stt.recognize(frames);
   const eventText = event.alternatives![0].text.toLowerCase().replace(/\s/g, ' ').trim();
@@ -64,6 +111,7 @@ export const tts = async (
         frames.push(event.frame);
       }
 
+      await assertPCM(frames);
       await validate(frames, stt, TEXT, 0.2);
       stream.close();
     });