Bump patch

.github/workflows/release.yml

@@ -5,6 +5,7 @@ on:
     types: [published]
 
 permissions:
+  id-token: write  # Required for OIDC
   contents: read
 
 jobs:
@@ -14,7 +15,7 @@ jobs:
     permissions: write-all
     steps:
       - name: Checkout repository
-        uses: actions/checkout@08eba0b27e820071cde6df949e0beb9ba4906955 # v4.3.0
+        uses: actions/checkout@v4
         with:
           fetch-depth: 0
 
@@ -27,12 +28,15 @@ jobs:
           git push origin release
 
       - name: Set up Node.js
-        uses: actions/setup-node@49933ea5288caeca8642d1e84afbd3f7d6820020 # v4.4.0
+        uses: actions/setup-node@v4
         with:
           node-version: 22
           cache: 'npm'
           registry-url: https://registry.npmjs.org
 
+      - name: Update npm
+        run: npm install -g npm@latest
+
       - name: Get package.json version
         id: package-json-version
         run: echo "version=v$(cat package.json | jq '.version' --raw-output)" >> $GITHUB_OUTPUT
@@ -58,11 +62,7 @@ jobs:
         run: gh release upload ${{ github.event.release.tag_name }} dist/bundles/mediabunny.cjs dist/bundles/mediabunny.min.cjs dist/bundles/mediabunny.mjs dist/bundles/mediabunny.min.mjs dist/mediabunny.d.ts packages/mp3-encoder/dist/bundles/mediabunny-mp3-encoder.js packages/mp3-encoder/dist/bundles/mediabunny-mp3-encoder.min.js packages/mp3-encoder/dist/bundles/mediabunny-mp3-encoder.mjs packages/mp3-encoder/dist/bundles/mediabunny-mp3-encoder.min.mjs packages/mp3-encoder/dist/mediabunny-mp3-encoder.d.ts
 
       - name: Publish Mediabunny to npm
-        run: npm publish --provenance --access public
-        env:
-          NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}
+        run: npm publish --access public
 
       - name: Publish workspace packages to npm
-        run: npm publish --provenance --access public --workspaces 
-        env:
-          NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}
+        run: npm publish --access public --workspaces 

dev/demux.html

@@ -14,9 +14,13 @@
 			source: new Mediabunny.BlobSource(file),
 		});
 
-		const audioTrack = await input.getPrimaryAudioTrack();
-		const decoderConfig = await audioTrack.getDecoderConfig();
-		console.log(decoderConfig);
+		const videoTrack = await input.getPrimaryVideoTrack();
+		const sink = new Mediabunny.EncodedPacketSink(videoTrack);
+
+		for await (const packet of sink.packets()) {
+			console.log(packet);
+		}
+
 
 		/*
 		const videoTrack = await input.getPrimaryVideoTrack();

dev/live.html

@@ -1,6 +1,7 @@
 <button>Go</button>
 
 <script src="../dist/bundles/mediabunny.cjs"></script>
+<script src="../packages/mp3-encoder/dist/bundles/mediabunny-mp3-encoder.js"></script>
 
 <script type="module">
     function download(blob, filename) {
@@ -12,6 +13,8 @@
 		URL.revokeObjectURL(url);
 	}
 
+	MediabunnyMp3Encoder.registerMp3Encoder();
+
     const button = document.querySelector('button');
     button.addEventListener('click', async () => {
         const stream = await navigator.mediaDevices.getUserMedia({ video: true, audio: true });
@@ -34,7 +37,7 @@
         }
         if (audioTrack) {
             const source = new Mediabunny.MediaStreamAudioTrackSource(audioTrack, {
-                codec: 'aac',
+                codec: 'mp3',
                 bitrate: Mediabunny.QUALITY_MEDIUM
             });
 

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "mediabunny",
-  "version": "1.25.4",
+  "version": "1.25.7",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "mediabunny",
-      "version": "1.25.4",
+      "version": "1.25.7",
       "license": "MPL-2.0",
       "workspaces": [
         "packages/*"
@@ -7739,9 +7739,9 @@
       }
     },
     "node_modules/mediabunny": {
-      "version": "1.25.3",
-      "resolved": "https://registry.npmjs.org/mediabunny/-/mediabunny-1.25.3.tgz",
-      "integrity": "sha512-+LDXv/kybsElRQCmMlrdbKmPNvegeIyouGeOUzW5DJ8+M56H4G6wyEL2AJ1A45UcOB+U62qJMV7tuFO9S7yA7g==",
+      "version": "1.25.6",
+      "resolved": "https://registry.npmjs.org/mediabunny/-/mediabunny-1.25.6.tgz",
+      "integrity": "sha512-0CUgi9oHbyzPHZttn6aCQrWZLUpSAb7m1QKSSjr780LU0bgF2rgsYXAzPwt0V9ScK0t5CPuVhlmxF8MVe5BFHw==",
       "license": "MPL-2.0",
       "peer": true,
       "workspaces": [
@@ -12065,7 +12065,7 @@
     },
     "packages/mp3-encoder": {
       "name": "@mediabunny/mp3-encoder",
-      "version": "1.25.4",
+      "version": "1.25.7",
       "license": "MPL-2.0",
       "devDependencies": {
         "@types/emscripten": "^1.40.1"

package.json

@@ -1,7 +1,7 @@
 {
   "name": "mediabunny",
   "author": "Vanilagy",
-  "version": "1.25.4",
+  "version": "1.25.7",
   "description": "Pure TypeScript media toolkit for reading, writing, and converting media files, directly in the browser.",
   "type": "module",
   "workspaces": [

packages/mp3-encoder/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@mediabunny/mp3-encoder",
   "author": "Vanilagy",
-  "version": "1.25.4",
+  "version": "1.25.7",
   "description": "MP3 encoder extension for Mediabunny, based on LAME.",
   "main": "./dist/bundles/mediabunny-mp3-encoder.mjs",
   "module": "./dist/bundles/mediabunny-mp3-encoder.mjs",

packages/mp3-encoder/src/index.ts

@@ -22,7 +22,7 @@ class Mp3Encoder extends CustomAudioEncoder {
 
 	private buffer = new Uint8Array(2 ** 16);
 	private currentBufferOffset = 0;
-	private currentTimestamp = 0;
+	private currentTimestamp: number | null = null;
 	private chunkMetadata: EncodedAudioChunkMetadata = {};
 
 	static override supports(codec: AudioCodec, config: AudioDecoderConfig): boolean {
@@ -79,6 +79,11 @@ class Mp3Encoder extends CustomAudioEncoder {
 	}
 
 	async encode(audioSample: AudioSample) {
+		if (this.currentTimestamp === null) {
+			// The first sample's timestamp determines where we start
+			this.currentTimestamp = audioSample.timestamp;
+		}
+
 		const sizePerChannel = audioSample.allocationSize({
 			format: 's16-planar',
 			planeIndex: 0,
@@ -123,6 +128,8 @@ class Mp3Encoder extends CustomAudioEncoder {
 	 * these chunks and extract the MP3 frames only when they're complete.
 	 */
 	private digestOutput(bytes: Uint8Array) {
+		assert(this.currentTimestamp !== null);
+
 		const requiredBufferSize = this.currentBufferOffset + bytes.length;
 		if (requiredBufferSize > this.buffer.length) {
 			// Grow the buffer to the required size

src/conversion.ts

@@ -82,10 +82,16 @@ export type ConversionOptions = {
 
 	/** Options to trim the input file. */
 	trim?: {
-		/** The time in the input file in seconds at which the output file should start. Must be less than `end`.  */
-		start: number;
-		/** The time in the input file in seconds at which the output file should end. Must be greater than `start`. */
-		end: number;
+		/**
+		 * The time in the input file in seconds at which the output file should start. Must be less than `end`.
+		 * Defaults to 0 when omitted.
+		 */
+		start?: number;
+		/**
+		 * The time in the input file in seconds at which the output file should end. Must be greater than `start`.
+		 * Defaults to the duration of the input when omitted.
+		 */
+		end?: number;
 	};
 
 	/**

src/isobmff/isobmff-demuxer.ts

@@ -2489,11 +2489,8 @@ abstract class IsobmffTrackBacking implements InputTrackBacking {
 		const timestampInTimescale = this.mapTimestampIntoTimescale(timestamp);
 
 		const sampleTable = this.internalTrack.demuxer.getSampleTableForTrack(this.internalTrack);
-		const sampleIndex = getSampleIndexForTimestamp(sampleTable, timestampInTimescale);
-		const keyFrameSampleIndex = sampleIndex === -1
-			? -1
-			: getRelevantKeyframeIndexForSample(sampleTable, sampleIndex);
-		const regularPacket = await this.fetchPacketForSampleIndex(keyFrameSampleIndex, options);
+		const sampleIndex = getKeyframeSampleIndexForTimestamp(sampleTable, timestampInTimescale);
+		const regularPacket = await this.fetchPacketForSampleIndex(sampleIndex, options);
 
 		if (!sampleTableIsEmpty(sampleTable) || !this.internalTrack.demuxer.isFragmented) {
 			// Prefer the non-fragmented packet
@@ -2913,7 +2910,45 @@ const getSampleIndexForTimestamp = (sampleTable: SampleTable, timescaleUnits: nu
 
 		const entry = sampleTable.sampleTimingEntries[index]!;
 		return entry.startIndex
-			+ Math.min(Math.floor((timescaleUnits - entry.startDecodeTimestamp) / entry.delta), entry.count - 1);
+			+ Math.min(
+				Math.floor((timescaleUnits - entry.startDecodeTimestamp) / entry.delta),
+				entry.count - 1,
+			);
+	}
+};
+
+const getKeyframeSampleIndexForTimestamp = (sampleTable: SampleTable, timescaleUnits: number) => {
+	if (!sampleTable.keySampleIndices) {
+		// Every sample is a keyframe
+		return getSampleIndexForTimestamp(sampleTable, timescaleUnits);
+	}
+
+	if (sampleTable.presentationTimestamps) {
+		const index = binarySearchLessOrEqual(
+			sampleTable.presentationTimestamps,
+			timescaleUnits,
+			x => x.presentationTimestamp,
+		);
+		if (index === -1) {
+			return -1;
+		}
+
+		// Walk the samples in presentation order until we find one that's a keyframe
+		for (let i = index; i >= 0; i--) {
+			const sampleIndex = sampleTable.presentationTimestamps[i]!.sampleIndex;
+			const isKeyFrame = binarySearchExact(sampleTable.keySampleIndices, sampleIndex, x => x) !== -1;
+
+			if (isKeyFrame) {
+				return sampleIndex;
+			}
+		}
+
+		return -1;
+	} else {
+		const sampleIndex = getSampleIndexForTimestamp(sampleTable, timescaleUnits);
+
+		const index = binarySearchLessOrEqual(sampleTable.keySampleIndices, sampleIndex, x => x);
+		return sampleTable.keySampleIndices[index] ?? -1;
 	}
 };
 
@@ -3002,15 +3037,6 @@ const getSampleInfo = (sampleTable: SampleTable, sampleIndex: number): SampleInf
 	};
 };
 
-const getRelevantKeyframeIndexForSample = (sampleTable: SampleTable, sampleIndex: number) => {
-	if (!sampleTable.keySampleIndices) {
-		return sampleIndex;
-	}
-
-	const index = binarySearchLessOrEqual(sampleTable.keySampleIndices, sampleIndex, x => x);
-	return sampleTable.keySampleIndices[index] ?? -1;
-};
-
 const getNextKeyframeIndexForSample = (sampleTable: SampleTable, sampleIndex: number) => {
 	if (!sampleTable.keySampleIndices) {
 		return sampleIndex + 1;

src/media-sink.ts

@@ -201,9 +201,10 @@ export class EncodedPacketSink {
 		}
 
 		const packet = await this._track._backing.getKeyPacket(timestamp, options);
-		if (!packet || packet.type === 'delta') {
+		if (!packet) {
 			return packet;
 		}
+		assert(packet.type === 'key');
 
 		const determinedType = await this._track.determinePacketType(packet);
 		if (determinedType === 'delta') {
@@ -235,9 +236,10 @@ export class EncodedPacketSink {
 		}
 
 		const nextPacket = await this._track._backing.getNextKeyPacket(packet, options);
-		if (!nextPacket || nextPacket.type === 'delta') {
+		if (!nextPacket) {
 			return nextPacket;
 		}
+		assert(nextPacket.type === 'key');
 
 		const determinedType = await this._track.determinePacketType(nextPacket);
 		if (determinedType === 'delta') {
@@ -474,9 +476,6 @@ export abstract class BaseMediaSampleSink<
 			const packetSink = this._createPacketSink();
 			const keyPacket = await packetSink.getKeyPacket(startTimestamp, { verifyKeyPackets: true })
 				?? await packetSink.getFirstPacket();
-			if (!keyPacket) {
-				return;
-			}
 
 			let currentPacket: EncodedPacket | null = keyPacket;
 
@@ -498,7 +497,7 @@ export abstract class BaseMediaSampleSink<
 				}
 			}
 
-			const packets = packetSink.packets(keyPacket, endPacket);
+			const packets = packetSink.packets(keyPacket ?? undefined, endPacket);
 			await packets.next(); // Skip the start packet as we already have it
 
 			while (currentPacket && !ended && !this._track.input._disposed) {
@@ -2127,11 +2126,14 @@ export class AudioBufferSink {
 
 	/** @internal */
 	_audioSampleToWrappedArrayBuffer(sample: AudioSample): WrappedAudioBuffer {
-		return {
+		const result: WrappedAudioBuffer = {
 			buffer: sample.toAudioBuffer(),
 			timestamp: sample.timestamp,
 			duration: sample.duration,
 		};
+
+		sample.close();
+		return result;
 	}
 
 	/**

src/media-source.ts

@@ -1356,21 +1356,27 @@ class AudioEncoderWrapper {
 					(audioSample.timestamp + audioSample.duration) * audioSample.sampleRate,
 				);
 
-				if (this.lastEndSampleIndex !== null && startSampleIndex > this.lastEndSampleIndex) {
-					const sampleCount = startSampleIndex - this.lastEndSampleIndex;
-					const fillSample = new AudioSample({
-						data: new Float32Array(sampleCount * audioSample.numberOfChannels),
-						format: 'f32-planar',
-						sampleRate: audioSample.sampleRate,
-						numberOfChannels: audioSample.numberOfChannels,
-						numberOfFrames: sampleCount,
-						timestamp: this.lastEndSampleIndex / audioSample.sampleRate,
-					});
+				if (this.lastEndSampleIndex === null) {
+					this.lastEndSampleIndex = endSampleIndex;
+				} else {
+					const sampleDiff = startSampleIndex - this.lastEndSampleIndex;
 
-					await this.add(fillSample, true); // Recursive call
+					if (sampleDiff >= 64) {
+						// The gap is big enough, let's add a correction sample
+						const fillSample = new AudioSample({
+							data: new Float32Array(sampleDiff * audioSample.numberOfChannels),
+							format: 'f32-planar',
+							sampleRate: audioSample.sampleRate,
+							numberOfChannels: audioSample.numberOfChannels,
+							numberOfFrames: sampleDiff,
+							timestamp: this.lastEndSampleIndex / audioSample.sampleRate,
+						});
+
+						await this.add(fillSample, true); // Recursive call
+					}
+
+					this.lastEndSampleIndex += audioSample.numberOfFrames;
 				}
-
-				this.lastEndSampleIndex = endSampleIndex;
 			}
 
 			if (this.customEncoder) {

test/browser/flac.test.ts

@@ -19,7 +19,7 @@ test('can decode samples from a FLAC file', async () => {
 
 	const sink = new AudioSampleSink(track);
 
-	const sample = await sink.getSample(1);
+	using sample = await sink.getSample(1);
 	assert(sample);
 	expect(sample.timestamp).toBe(0.9287981859410431);
 });

test/browser/transparency.test.ts

@@ -21,7 +21,7 @@ test('Can decode transparent video', async () => {
 	expect(await videoTrack.canBeTransparent()).toBe(true);
 
 	const sink = new VideoSampleSink(videoTrack);
-	const sample = (await sink.getSample(0.5))!;
+	using sample = (await sink.getSample(0.5))!;
 
 	expect(sample.format).toContain('A'); // Probably RGBA
 	expect(sample.hasAlpha).toBe(true);
@@ -47,10 +47,10 @@ test('Can decode faulty transparent video and behaves gracefully', async () => {
 
 	const sink = new VideoSampleSink(videoTrack);
 
-	const startSample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
+	using startSample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
 	expect(startSample.format).toContain('A');
 
-	const secondSample = (await sink.getSample(secondKeyPacket.timestamp))!;
+	using secondSample = (await sink.getSample(secondKeyPacket.timestamp))!;
 	expect(secondSample.format).not.toContain('A'); // There was no alpha key frame for this one
 	expect(secondSample.hasAlpha).toBe(false);
 });
@@ -164,7 +164,7 @@ test('Can encode transparent video', async () => {
 
 	const sink = new VideoSampleSink(videoTrack);
 
-	const firstSample = (await sink.getSample(0))!;
+	using firstSample = (await sink.getSample(0))!;
 	expect(firstSample.format).toContain('A');
 
 	probeContext.clearRect(0, 0, probeCanvas.width, probeCanvas.height);
@@ -201,7 +201,7 @@ test('Can encode video with alternating transparency', async () => {
 	await output.start();
 
 	for (let i = 0; i < 64; i++) {
-		const sample = new VideoSample(new Uint8Array(640 * 480 * 4), {
+		using sample = new VideoSample(new Uint8Array(640 * 480 * 4), {
 			format: i % 2 ? 'RGBX' : 'RGBA',
 			codedWidth: 640,
 			codedHeight: 480,
@@ -235,7 +235,7 @@ test('Can encode video with alternating transparency', async () => {
 	const sampleSink = new VideoSampleSink(videoTrack);
 
 	i = 0;
-	for await (const sample of sampleSink.samples()) {
+	for await (using sample of sampleSink.samples()) {
 		if (i % 2) {
 			expect(sample.format).not.toContain('A');
 		} else {
@@ -299,7 +299,7 @@ test('Can transmux transparent video, discards alpha by default', async () => {
 	expect(await videoTrack.canBeTransparent()).toBe(false);
 
 	const sink = new VideoSampleSink(videoTrack);
-	const sample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
+	using sample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
 	expect(sample.hasAlpha).toBe(false);
 });
 
@@ -331,7 +331,7 @@ test('Can transmux transparent video, can keep alpha', async () => {
 	expect(await videoTrack.canBeTransparent()).toBe(true);
 
 	const sink = new VideoSampleSink(videoTrack);
-	const sample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
+	using sample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
 	expect(sample.format).toContain('A');
 	expect(sample.hasAlpha).toBe(true);
 });
@@ -370,7 +370,7 @@ test('Can reencode transparent video, keeping alpha', async () => {
 	expect(videoTrack.displayWidth).toBe(320);
 
 	const sink = new VideoSampleSink(videoTrack);
-	const sample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
+	using sample = (await sink.getSample(await videoTrack.getFirstTimestamp()))!;
 	expect(sample.format).toContain('A');
 	expect(sample.hasAlpha).toBe(true);
 });