arcodange/mediabunny
1
0
Fork 0
mirror of https://github.com/arcodange-org/mediabunny.git synced 2026-10-02 21:33:52 +02:00
Code Issues Packages Projects Releases Wiki Activity

Make ADTS demuxer spit out ADTS packets instead of AAC, make ADTS muxer accept both AAC and ADTS input

Browse Source
This commit is contained in:
Vanilagy
2026-01-14 13:52:46 +01:00
parent 06ab9c8433
commit 045df3b9a5
6 changed files with 242 additions and 61 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/adts/adts-demuxer.ts
+3 -22
View File
@@ -16,13 +16,12 @@ import {
AsyncMutex,
binarySearchExact,
binarySearchLessOrEqual,
Bitstream,
UNDETERMINED_LANGUAGE,
} from '../misc';
import { EncodedPacket, PLACEHOLDER_DATA } from '../packet';
import { readBytes, Reader } from '../reader';
import { DEFAULT_TRACK_DISPOSITION } from '../metadata';
import { AdtsFrameHeader, MAX_FRAME_HEADER_SIZE, MIN_FRAME_HEADER_SIZE, readAdtsFrameHeader } from './adts-reader';
import { AdtsFrameHeader, MIN_FRAME_HEADER_SIZE, MAX_FRAME_HEADER_SIZE, readAdtsFrameHeader } from './adts-reader';
export const SAMPLES_PER_AAC_FRAME = 1024;
@@ -95,13 +94,12 @@ export class AdtsDemuxer extends Demuxer {
const sampleRate = aacFrequencyTable[header.samplingFrequencyIndex];
assert(sampleRate !== undefined);
const sampleDuration = SAMPLES_PER_AAC_FRAME / sampleRate;
const headerSize = header.crcCheck ? MAX_FRAME_HEADER_SIZE : MIN_FRAME_HEADER_SIZE;
const sample: Sample = {
timestamp: this.nextTimestampInSamples / sampleRate,
duration: sampleDuration,
dataStart: header.startPos + headerSize,
dataSize: header.frameLength - headerSize,
dataStart: header.startPos,
dataSize: header.frameLength,
};
this.loadedSamples.push(sample);
@@ -198,27 +196,10 @@ class AdtsAudioTrackBacking implements InputAudioTrackBacking {
async getDecoderConfig(): Promise<AudioDecoderConfig> {
assert(this.demuxer.firstFrameHeader);
const bytes = new Uint8Array(3); // 19 bits max
const bitstream = new Bitstream(bytes);
const { objectType, samplingFrequencyIndex, channelConfiguration } = this.demuxer.firstFrameHeader;
if (objectType > 31) {
bitstream.writeBits(5, 31);
bitstream.writeBits(6, objectType - 32);
} else {
bitstream.writeBits(5, objectType);
}
bitstream.writeBits(4, samplingFrequencyIndex); // samplingFrequencyIndex === 15 is forbidden
bitstream.writeBits(4, channelConfiguration);
return {
codec: `mp4a.40.${this.demuxer.firstFrameHeader.objectType}`,
numberOfChannels: this.getNumberOfChannels(),
sampleRate: this.getSampleRate(),
description: bytes.subarray(0, Math.ceil((bitstream.pos - 1) / 8)),
};
}
src/adts/adts-muxer.ts
+52 -33
View File
@@ -7,7 +7,7 @@
*/
import { AacAudioSpecificConfig, parseAacAudioSpecificConfig, validateAudioChunkMetadata } from '../codec';
import { assert, Bitstream, toUint8Array } from '../misc';
import { Bitstream, toUint8Array } from '../misc';
import { Muxer } from '../muxer';
import { Output, OutputAudioTrack } from '../output';
import { AdtsOutputFormat } from '../output-format';
@@ -20,6 +20,7 @@ export class AdtsMuxer extends Muxer {
private header = new Uint8Array(7);
private headerBitstream = new Bitstream(this.header);
private audioSpecificConfig: AacAudioSpecificConfig | null = null;
private inputIsAdts: boolean | null = null;
constructor(output: Output, format: AdtsOutputFormat) {
super(output);
@@ -52,49 +53,67 @@ export class AdtsMuxer extends Muxer {
try {
this.validateAndNormalizeTimestamp(track, packet.timestamp, packet.type === 'key');
if (!this.audioSpecificConfig) {
if (this.inputIsAdts === null) {
validateAudioChunkMetadata(meta);
const description = meta?.decoderConfig?.description;
assert(description);
this.audioSpecificConfig = parseAacAudioSpecificConfig(toUint8Array(description));
// From the WebCodecs Codec Registry:
// "If description is present, it is assumed to a AudioSpecificConfig as defined in [iso14496-3] section
// 1.6.2.1, Table 1.15, and the bitstream is assumed to be in aac.
// If the description is not present, the bitstream is assumed to be in adts format."
this.inputIsAdts = !description;
const { objectType, frequencyIndex, channelConfiguration } = this.audioSpecificConfig;
const profile = objectType - 1;
if (!this.inputIsAdts) {
this.audioSpecificConfig = parseAacAudioSpecificConfig(toUint8Array(description!));
this.headerBitstream.writeBits(12, 0b1111_11111111); // Syncword
this.headerBitstream.writeBits(1, 0); // MPEG Version
this.headerBitstream.writeBits(2, 0); // Layer
this.headerBitstream.writeBits(1, 1); // Protection absence
this.headerBitstream.writeBits(2, profile); // Profile
this.headerBitstream.writeBits(4, frequencyIndex); // MPEG-4 Sampling Frequency Index
this.headerBitstream.writeBits(1, 0); // Private bit
this.headerBitstream.writeBits(3, channelConfiguration); // MPEG-4 Channel Configuration
this.headerBitstream.writeBits(1, 0); // Originality
this.headerBitstream.writeBits(1, 0); // Home
this.headerBitstream.writeBits(1, 0); // Copyright ID bit
this.headerBitstream.writeBits(1, 0); // Copyright ID start
this.headerBitstream.skipBits(13); // Frame length
this.headerBitstream.writeBits(11, 0x7ff); // Buffer fullness
this.headerBitstream.writeBits(2, 0); // Number of AAC frames minus 1
// Omit CRC check
const { objectType, frequencyIndex, channelConfiguration } = this.audioSpecificConfig;
const profile = objectType - 1;
this.headerBitstream.writeBits(12, 0b1111_11111111); // Syncword
this.headerBitstream.writeBits(1, 0); // MPEG Version
this.headerBitstream.writeBits(2, 0); // Layer
this.headerBitstream.writeBits(1, 1); // Protection absence
this.headerBitstream.writeBits(2, profile); // Profile
this.headerBitstream.writeBits(4, frequencyIndex); // MPEG-4 Sampling Frequency Index
this.headerBitstream.writeBits(1, 0); // Private bit
this.headerBitstream.writeBits(3, channelConfiguration); // MPEG-4 Channel Configuration
this.headerBitstream.writeBits(1, 0); // Originality
this.headerBitstream.writeBits(1, 0); // Home
this.headerBitstream.writeBits(1, 0); // Copyright ID bit
this.headerBitstream.writeBits(1, 0); // Copyright ID start
this.headerBitstream.skipBits(13); // Frame length
this.headerBitstream.writeBits(11, 0x7ff); // Buffer fullness
this.headerBitstream.writeBits(2, 0); // Number of AAC frames minus 1
// Omit CRC check
}
}
const frameLength = packet.data.byteLength + this.header.byteLength;
this.headerBitstream.pos = 30;
this.headerBitstream.writeBits(13, frameLength);
if (this.inputIsAdts) {
// Packets are already ADTS frames, write them directly
const startPos = this.writer.getPos();
this.writer.write(packet.data);
const startPos = this.writer.getPos();
this.writer.write(this.header);
this.writer.write(packet.data);
if (this.format._options.onFrame) {
this.format._options.onFrame(packet.data, startPos);
}
} else {
// Packets are raw AAC, prepend ADTS header
const frameLength = packet.data.byteLength + this.header.byteLength;
this.headerBitstream.pos = 30;
this.headerBitstream.writeBits(13, frameLength);
if (this.format._options.onFrame) {
const frameBytes = new Uint8Array(frameLength);
frameBytes.set(this.header, 0);
frameBytes.set(packet.data, this.header.byteLength);
const startPos = this.writer.getPos();
this.writer.write(this.header);
this.writer.write(packet.data);
this.format._options.onFrame(frameBytes, startPos);
if (this.format._options.onFrame) {
const frameBytes = new Uint8Array(frameLength);
frameBytes.set(this.header, 0);
frameBytes.set(packet.data, this.header.byteLength);
this.format._options.onFrame(frameBytes, startPos);
}
}
await this.writer.flush();