Fix WAV files with embedded ID3 metadata returning empty tags (#110)

* Read ID3v2 tags from WAV file

* remove js import

Co-authored-by: David P. <[email protected]>

* revert slicing in wave-demuxer.ts and add id3.ts for further sharing

* Add `writeId3Tag` to the `WavOutputFormat` constructor options

* add more context in private and public APIs that it's non-standard

* compare actual values in the test with real file

* remove unnecesarry double slice requesting

* trim test audio file (i tried my best to preserve ID3, including artwork)

* Couple of code cleanups

* const -> using

* Fix linting errors

---------

Co-authored-by: David P. <[email protected]>

src/wave/wave-demuxer.ts

@@ -15,6 +15,7 @@ import { MetadataTags } from '../tags';
 import { assert, UNDETERMINED_LANGUAGE } from '../misc';
 import { EncodedPacket, PLACEHOLDER_DATA } from '../packet';
 import { readAscii, readBytes, Reader, readU16, readU32, readU64 } from '../reader';
+import { parseId3V2Tag, readId3V2Header } from '../id3';
 
 export enum WaveFormat {
 	PCM = 0x0001,
@@ -107,6 +108,8 @@ export class WaveDemuxer extends Demuxer {
 					totalFileSize = Math.min(riffChunkSize + 8, this.reader.fileSize ?? Infinity);
 				} else if (chunkId === 'LIST') {
 					await this.parseListChunk(startPos, chunkSize, littleEndian);
+				} else if (chunkId === 'ID3 ' || chunkId === 'id3 ') {
+					await this.parseId3Chunk(startPos, chunkSize);
 				}
 
 				currentPos = startPos + chunkSize + (chunkSize & 1); // Handle padding
@@ -271,6 +274,21 @@ export class WaveDemuxer extends Demuxer {
 		}
 	}
 
+	private async parseId3Chunk(startPos: number, size: number) {
+		// Parse ID3 tag embedded in WAV file (non-default, but used a lot in practice anyway)
+		let slice = this.reader.requestSlice(startPos, size);
+		if (slice instanceof Promise) slice = await slice;
+		if (!slice) return; // File too short
+
+		const id3V2Header = readId3V2Header(slice);
+		if (id3V2Header) {
+			// Extract the content portion (skip the 10-byte header)
+			const contentSlice = slice.slice(startPos + 10, id3V2Header.size);
+
+			parseId3V2Tag(contentSlice, id3V2Header, this.metadataTags);
+		}
+	}
+
 	getCodec(): AudioCodec | null {
 		assert(this.audioInfo);
 

src/wave/wave-muxer.ts

@@ -16,6 +16,7 @@ import { EncodedPacket } from '../packet';
 import { WavOutputFormat } from '../output-format';
 import { assert, assertNever, isIso88591Compatible, keyValueIterator } from '../misc';
 import { MetadataTags, metadataTagsAreEmpty } from '../tags';
+import { Id3V2Writer } from '../id3';
 
 export class WaveMuxer extends Muxer {
 	private format: WavOutputFormat;
@@ -159,9 +160,17 @@ export class WaveMuxer extends Muxer {
 		this.riffWriter.writeU16(blockSize);
 		this.riffWriter.writeU16(8 * pcmInfo.sampleSize);
 
+		// Metadata tags
 		if (!metadataTagsAreEmpty(this.output._metadataTags)) {
-			// Metadata exists, let's write an INFO chunk
-			this.writeInfoChunk(this.output._metadataTags);
+			const metadataFormat = this.format._options.metadataFormat ?? 'info';
+
+			if (metadataFormat === 'info') {
+				this.writeInfoChunk(this.output._metadataTags);
+			} else if (metadataFormat === 'id3') {
+				this.writeId3Chunk(this.output._metadataTags);
+			} else {
+				assertNever(metadataFormat);
+			}
 		}
 
 		// data chunk
@@ -300,6 +309,29 @@ export class WaveMuxer extends Muxer {
 		}
 	}
 
+	private writeId3Chunk(metadata: MetadataTags) {
+		const startPos = this.writer.getPos();
+
+		// Write RIFF chunk header
+		this.riffWriter.writeAscii('ID3 ');
+		this.riffWriter.writeU32(0); // Size placeholder
+
+		const id3Writer = new Id3V2Writer(this.writer);
+		const id3TagSize = id3Writer.writeId3V2Tag(metadata);
+
+		const endPos = this.writer.getPos();
+
+		// Update RIFF chunk size
+		this.writer.seek(startPos + 4);
+		this.riffWriter.writeU32(id3TagSize);
+		this.writer.seek(endPos);
+
+		// Add padding byte if chunk size is odd
+		if (id3TagSize & 1) {
+			this.writer.write(new Uint8Array(1));
+		}
+	}
+
 	async finalize() {
 		const release = await this.mutex.acquire();