Add support for reading and writing RF64 WAVE files

dev/convert.html

@@ -21,7 +21,7 @@
 			chunked: true,
 			chunkSize: 2**20
 		});
-		const outputFormat = new Mediabunny.Mp4OutputFormat();
+		const outputFormat = new Mediabunny.WavOutputFormat({ large: true });
 
 		const button = document.createElement('button');
 		button.textContent = 'Cancel';
@@ -38,8 +38,8 @@
 				target
 			}),
 			audio: {
-				codec: 'opus',
-				bitrate: 128000,
+				//codec: 'opus',
+				//bitrate: 128000,
 				//numberOfChannels: 1,
 				//sampleRate: 4000
 				//discard: true

docs/.vitepress/config.mts

@@ -4,11 +4,13 @@ import tailwindcss from '@tailwindcss/vite';
 import llmstxt from 'vitepress-plugin-llms';
 import { HeadConfig } from 'vitepress';
 
+const DESCRIPTION = 'A JavaScript library for reading, writing, and converting media files. Directly in the browser,'
+	+ ' and faster than anybunny else.';
+
 // https://vitepress.dev/reference/site-config
 export default withMermaid({
 	title: 'Mediabunny',
-	description: 'A JavaScript library for reading, writing, and converting media files. Directly in the browser, and'
-		+ ' faster than anybunny else.',
+	description: DESCRIPTION,
 	cleanUrls: true,
 	head: [
 		['link', { rel: 'icon', type: 'image/png', href: '/mediabunny-logo.png' }],
@@ -16,9 +18,13 @@ export default withMermaid({
 		['meta', { property: 'og:type', content: 'website' }],
 		['meta', { property: 'og:site_name', content: 'Mediabunny' }],
 		['meta', { property: 'og:url', content: 'https://mediabunny.dev/' }],
-		['meta', { property: 'og:image', content: '/mediabunny-og-image.png' }],
+		['meta', { property: 'og:image', content: 'https://mediabunny.dev/mediabunny-og-image.png' }],
 		['meta', { property: 'og:locale', content: 'en-US' }],
+		['meta', { property: 'og:description', content: DESCRIPTION }],
+		['meta', { name: 'twitter:image', content: 'https://mediabunny.dev/mediabunny-og-image.png' }],
 		['meta', { name: 'twitter:card', content: 'summary_large_image' }],
+		['meta', { name: 'twitter:site', content: '@vanilagy' }],
+		['meta', { name: 'twitter:description', content: DESCRIPTION }],
 	],
 	themeConfig: {
 		logo: '/mediabunny-logo.svg',
@@ -114,6 +120,7 @@ export default withMermaid({
 
 		((pageData.frontmatter['head'] ??= []) as HeadConfig[]).push(
 			['meta', { property: 'og:title', content: title }],
+			['meta', { property: 'twitter:title', content: title }],
 		);
 	},
 });

docs/guide/output-formats.md

@@ -227,8 +227,11 @@ const output = new Output({
 The following options are available:
 ```ts
 type WavOutputFormatOptions = {
+	large?: boolean;
 	onHeader?: (data: Uint8Array, position: number) => unknown;
 };
 ```
+- `large`\
+	When enabled, an RF64 file be written, allowing for file sizes to exceed 4 GiB, which is otherwise not possible for regular WAVE files.
 - `onHeader`\
 	Will be called once the file header is written. The header consists of the RIFF header, the format chunk, and the start of the data chunk (with a placeholder size of 0).

docs/index.md

@@ -95,6 +95,7 @@ const sponsors = {
 		{ image: 'https://avatars.githubusercontent.com/u/84167135', name: 'Memenome', url: 'https://github.com/memenome' },
 		{ image: 'https://avatars.githubusercontent.com/u/30229596', name: 'Pablo Bonilla', url: 'https://github.com/devPablo' },
 		{ image: 'https://avatars.githubusercontent.com/u/58149663', name: 'H7GhosT', url: 'https://github.com/H7GhosT' },
+		{ image: 'https://avatars.githubusercontent.com/u/91711202', name: 'ihasq', url: 'https://github.com/ihasq' },
 	],
 };
 </script>

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "mediabunny",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "mediabunny",
-      "version": "1.1.0",
+      "version": "1.2.0",
       "license": "MPL-2.0",
       "dependencies": {
         "@types/dom-mediacapture-transform": "^0.1.11",

package.json

@@ -1,7 +1,7 @@
 {
   "name": "mediabunny",
   "author": "Vanilagy",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "description": "Pure TypeScript media toolkit for reading, writing, and converting media files, directly in the browser.",
   "type": "module",
   "main": "./dist/bundles/mediabunny.cjs",

src/input-format.ts

@@ -291,7 +291,7 @@ export class WaveInputFormat extends InputFormat {
 
 		const riffReader = new RiffReader(input._mainReader);
 		const riffType = riffReader.readAscii(4);
-		if (riffType !== 'RIFF' && riffType !== 'RIFX') {
+		if (riffType !== 'RIFF' && riffType !== 'RIFX' && riffType !== 'RF64') {
 			return false;
 		}
 

src/matroska/matroska-demuxer.ts

@@ -220,7 +220,7 @@ export class MatroskaDemuxer extends Demuxer {
 
 			const fileSize = await this.input.source.getSize();
 
-			while (this.metadataReader.pos < fileSize - MIN_HEADER_SIZE) {
+			while (this.metadataReader.pos <= fileSize - MIN_HEADER_SIZE) {
 				await this.metadataReader.reader.loadRange(
 					this.metadataReader.pos,
 					this.metadataReader.pos + MAX_HEADER_SIZE,
@@ -557,7 +557,7 @@ export class MatroskaDemuxer extends Demuxer {
 	readContiguousElements(reader: EBMLReader, totalSize: number) {
 		const startIndex = reader.pos;
 
-		while (reader.pos - startIndex < totalSize) {
+		while (reader.pos - startIndex <= totalSize - MIN_HEADER_SIZE) {
 			this.traverseElement(reader);
 		}
 	}
@@ -1468,7 +1468,7 @@ abstract class MatroskaTrackBacking implements InputTrackBacking {
 					}
 
 					const endPos = dataStartPos + size;
-					if (endPos >= segment.elementEndPos - MIN_HEADER_SIZE) {
+					if (endPos > segment.elementEndPos - MIN_HEADER_SIZE) {
 						// No more elements fit in this segment
 						break;
 					} else {

src/output-format.ts

@@ -546,6 +546,12 @@ export class Mp3OutputFormat extends OutputFormat {
  * @public
  */
 export type WavOutputFormatOptions = {
+	/**
+	 * When enabled, an RF64 file be written, allowing for file sizes to exceed 4 GiB, which is otherwise not possible
+	 * for regular WAVE files.
+	 */
+	large?: boolean;
+
 	/**
 	 * Will be called once the file header is written. The header consists of the RIFF header, the format chunk, and the
 	 * start of the data chunk (with a placeholder size of 0).
@@ -565,6 +571,9 @@ export class WavOutputFormat extends OutputFormat {
 		if (!options || typeof options !== 'object') {
 			throw new TypeError('options must be an object.');
 		}
+		if (options.large !== undefined && typeof options.large !== 'boolean') {
+			throw new TypeError('options.large, when provided, must be a boolean.');
+		}
 		if (options.onHeader !== undefined && typeof options.onHeader !== 'function') {
 			throw new TypeError('options.onHeader, when provided, must be a function.');
 		}

src/wave/riff-reader.ts

@@ -35,6 +35,21 @@ export class RiffReader {
 		return view.getUint32(offset, this.littleEndian);
 	}
 
+	readU64() {
+		let low: number;
+		let high: number;
+
+		if (this.littleEndian) {
+			low = this.readU32();
+			high = this.readU32();
+		} else {
+			high = this.readU32();
+			low = this.readU32();
+		}
+
+		return high * 0x100000000 + low;
+	}
+
 	readAscii(length: number) {
 		const { view, offset } = this.reader.getViewAndOffset(this.pos, this.pos + length);
 		this.pos += length;

src/wave/riff-writer.ts

@@ -14,14 +14,20 @@ export class RiffWriter {
 
 	constructor(private writer: Writer) {}
 
+	writeU16(value: number) {
+		this.helperView.setUint16(0, value, true);
+		this.writer.write(this.helper.subarray(0, 2));
+	}
+
 	writeU32(value: number) {
 		this.helperView.setUint32(0, value, true);
 		this.writer.write(this.helper.subarray(0, 4));
 	}
 
-	writeU16(value: number) {
-		this.helperView.setUint16(0, value, true);
-		this.writer.write(this.helper.subarray(0, 2));
+	writeU64(value: number) {
+		this.helperView.setUint32(0, value, true);
+		this.helperView.setUint32(4, Math.floor(value / 2 ** 32), true);
+		this.writer.write(this.helper);
 	}
 
 	writeAscii(text: string) {

src/wave/wave-demuxer.ts

@@ -53,9 +53,13 @@ export class WaveDemuxer extends Demuxer {
 			const actualFileSize = await this.metadataReader.reader.source.getSize();
 
 			const riffType = this.metadataReader.readAscii(4);
-			this.metadataReader.littleEndian = riffType === 'RIFF';
+			this.metadataReader.littleEndian = riffType !== 'RIFX';
 
-			const totalFileSize = Math.min(this.metadataReader.readU32() + 8, actualFileSize);
+			const isRf64 = riffType === 'RF64';
+
+			const outerChunkSize = this.metadataReader.readU32();
+
+			let totalFileSize = isRf64 ? actualFileSize : Math.min(outerChunkSize + 8, actualFileSize);
 			const format = this.metadataReader.readAscii(4);
 
 			if (format !== 'WAVE') {
@@ -63,6 +67,9 @@ export class WaveDemuxer extends Demuxer {
 			}
 
 			this.metadataReader.pos = 12;
+			let chunksRead = 0;
+			let dataChunkSize: number | null = null;
+
 			while (this.metadataReader.pos < totalFileSize) {
 				await this.metadataReader.reader.loadRange(this.metadataReader.pos, this.metadataReader.pos + 8);
 
@@ -70,14 +77,28 @@ export class WaveDemuxer extends Demuxer {
 				const chunkSize = this.metadataReader.readU32();
 				const startPos = this.metadataReader.pos;
 
+				if (isRf64 && chunksRead === 0 && chunkId !== 'ds64') {
+					throw new Error('Invalid RF64 file: First chunk must be "ds64".');
+				}
+
 				if (chunkId === 'fmt ') {
 					await this.parseFmtChunk(chunkSize);
 				} else if (chunkId === 'data') {
+					dataChunkSize ??= chunkSize;
+
 					this.dataStart = this.metadataReader.pos;
-					this.dataSize = Math.min(chunkSize, totalFileSize - this.dataStart);
+					this.dataSize = Math.min(dataChunkSize, totalFileSize - this.dataStart);
+				} else if (chunkId === 'ds64') {
+					// File and data chunk sizes are defined in here instead
+
+					const riffChunkSize = this.metadataReader.readU64();
+					dataChunkSize = this.metadataReader.readU64();
+
+					totalFileSize = Math.min(riffChunkSize + 8, actualFileSize);
 				}
 
 				this.metadataReader.pos = startPos + chunkSize + (chunkSize & 1); // Handle padding
+				chunksRead++;
 			}
 
 			if (!this.audioInfo) {

src/wave/wave-muxer.ts

@@ -18,10 +18,13 @@ import { assert } from '../misc';
 
 export class WaveMuxer extends Muxer {
 	private format: WavOutputFormat;
+	private isRf64: boolean;
 	private writer: Writer;
 	private riffWriter: RiffWriter;
 	private headerWritten = false;
 	private dataSize = 0;
+	private sampleRate: number | null = null;
+	private sampleCount = 0;
 
 	constructor(output: Output, format: WavOutputFormat) {
 		super(output);
@@ -29,6 +32,7 @@ export class WaveMuxer extends Muxer {
 		this.format = format;
 		this.writer = output._writer;
 		this.riffWriter = new RiffWriter(output._writer);
+		this.isRf64 = !!format._options.large;
 	}
 
 	async start() {
@@ -58,13 +62,22 @@ export class WaveMuxer extends Muxer {
 				assert(meta.decoderConfig);
 
 				this.writeHeader(track, meta.decoderConfig);
+				this.sampleRate = meta.decoderConfig.sampleRate;
 				this.headerWritten = true;
 			}
 
 			this.validateAndNormalizeTimestamp(track, packet.timestamp, packet.type === 'key');
 
+			if (!this.isRf64 && this.writer.getPos() + packet.data.byteLength >= 2 ** 32) {
+				throw new Error(
+					'Adding more audio data would exceed the maximum RIFF size of 4 GiB. To write larger files, use'
+					+ ' RF64 by setting `large: true` in the WavOutputFormatOptions.',
+				);
+			}
+
 			this.writer.write(packet.data);
 			this.dataSize += packet.data.byteLength;
+			this.sampleCount += Math.round(packet.duration * this.sampleRate!);
 
 			await this.writer.flush();
 		} finally {
@@ -101,10 +114,26 @@ export class WaveMuxer extends Muxer {
 		const blockSize = pcmInfo.sampleSize * channels;
 
 		// RIFF header
-		this.riffWriter.writeAscii('RIFF');
-		this.riffWriter.writeU32(0); // File size placeholder
+		this.riffWriter.writeAscii(this.isRf64 ? 'RF64' : 'RIFF');
+
+		if (this.isRf64) {
+			this.riffWriter.writeU32(0xffffffff); // Not used in RF64
+		} else {
+			this.riffWriter.writeU32(0); // File size placeholder
+		}
+
 		this.riffWriter.writeAscii('WAVE');
 
+		if (this.isRf64) {
+			this.riffWriter.writeAscii('ds64');
+			this.riffWriter.writeU32(28); // Chunk size
+			this.riffWriter.writeU64(0); // RIFF size placeholder
+			this.riffWriter.writeU64(0); // Data size placeholder
+			this.riffWriter.writeU64(0); // Sample count placeholder
+			this.riffWriter.writeU32(0); // Table length
+			// Empty table
+		}
+
 		// fmt chunk
 		this.riffWriter.writeAscii('fmt ');
 		this.riffWriter.writeU32(16); // Chunk size
@@ -117,7 +146,12 @@ export class WaveMuxer extends Muxer {
 
 		// data chunk
 		this.riffWriter.writeAscii('data');
-		this.riffWriter.writeU32(0); // Data size placeholder
+
+		if (this.isRf64) {
+			this.riffWriter.writeU32(0xffffffff); // Not used in RF64
+		} else {
+			this.riffWriter.writeU32(0); // Data size placeholder
+		}
 
 		if (this.format._options.onHeader) {
 			const { data, start } = this.writer.stopTrackingWrites();
@@ -130,13 +164,27 @@ export class WaveMuxer extends Muxer {
 
 		const endPos = this.writer.getPos();
 
-		// Write file size
-		this.writer.seek(4);
-		this.riffWriter.writeU32(this.dataSize + 36); // File size - 8
+		if (this.isRf64) {
+			// Write riff size
+			this.writer.seek(20);
+			this.riffWriter.writeU64(endPos - 8);
 
-		// Write data chunk size
-		this.writer.seek(40);
-		this.riffWriter.writeU32(this.dataSize);
+			// Write data size
+			this.writer.seek(28);
+			this.riffWriter.writeU64(this.dataSize);
+
+			// Write sample count
+			this.writer.seek(36);
+			this.riffWriter.writeU64(this.sampleCount);
+		} else {
+			// Write file size
+			this.writer.seek(4);
+			this.riffWriter.writeU32(endPos - 8);
+
+			// Write data chunk size
+			this.writer.seek(40);
+			this.riffWriter.writeU32(this.dataSize);
+		}
 
 		this.writer.seek(endPos);