Remove sample correlation logic in decoder pipeline

src/mp3/mp3-demuxer.ts

@@ -8,8 +8,6 @@ import { EncodedAudioSample, PLACEHOLDER_DATA } from '../sample';
 import { FrameHeader, getXingOffset, INFO, XING } from './mp3-misc';
 import { Mp3Reader } from './mp3-reader';
 
-const AUDIO_SAMPLES_PER_FRAME = 1152;
-
 type Sample = {
 	timestamp: number;
 	duration: number;
@@ -46,7 +44,7 @@ export class Mp3Demuxer extends Demuxer {
 				this.reader.pos += id3Tag.size;
 			}
 
-			let nextTimestamp = 0;
+			let nextTimestampInSamples = 0;
 
 			// Let's read all samples
 			while (true) {
@@ -71,16 +69,16 @@ export class Mp3Demuxer extends Demuxer {
 					this.firstFrameHeader = header;
 				}
 
-				const sampleDuration = AUDIO_SAMPLES_PER_FRAME / header.sampleRate;
+				const sampleDuration = header.audioSamplesInFrame / header.sampleRate;
 				const sample: Sample = {
-					timestamp: nextTimestamp,
+					timestamp: nextTimestampInSamples / header.sampleRate,
 					duration: sampleDuration,
 					dataStart: header.startPos,
 					dataSize: header.totalSize,
 				};
 
 				this.allSamples.push(sample);
-				nextTimestamp += sampleDuration;
+				nextTimestampInSamples += header.audioSamplesInFrame;
 			}
 
 			if (!this.firstFrameHeader) {
@@ -121,6 +119,11 @@ class Mp3AudioTrackBacking implements InputAudioTrackBacking {
 		return 0;
 	}
 
+	async getTimeResolution() {
+		assert(this.demuxer.firstFrameHeader);
+		return this.demuxer.firstFrameHeader.sampleRate / this.demuxer.firstFrameHeader.audioSamplesInFrame;
+	}
+
 	computeDuration() {
 		return this.demuxer.computeDuration();
 	}

src/mp3/mp3-misc.ts

@@ -44,6 +44,7 @@ export type FrameHeader = {
 	copyright: number;
 	original: number;
 	emphasis: number;
+	audioSamplesInFrame: number;
 };
 
 export const computeMp3FrameSize = (layer: number, bitrate: number, sampleRate: number, padding: number) => {
@@ -114,6 +115,19 @@ export const readFrameHeader = (word: number, reader: { pos: number; fileSize: n
 		return null;
 	}
 
+	let audioSamplesInFrame: number;
+	if (mpegVersionId === 3) {
+		audioSamplesInFrame = layer === 3 ? 384 : 1152;
+	} else {
+		if (layer === 3) {
+			audioSamplesInFrame = 384;
+		} else if (layer === 2) {
+			audioSamplesInFrame = 1152;
+		} else {
+			audioSamplesInFrame = 576;
+		}
+	}
+
 	return {
 		startPos: startPos,
 		totalSize: frameLength,
@@ -127,5 +141,6 @@ export const readFrameHeader = (word: number, reader: { pos: number; fileSize: n
 		copyright,
 		original,
 		emphasis,
+		audioSamplesInFrame,
 	};
 };