Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix: Fix audio properties detection #6867

Merged
merged 4 commits into from
Jun 20, 2024
Merged

fix: Fix audio properties detection #6867

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
95a50df
fix: Fix audio properties detection
avelad Jun 19, 2024
defa8f0
Merge branch 'main' into audio-detection
avelad Jun 20, 2024
01840a2
Fix typo
avelad Jun 20, 2024
a4ecad1
Simplify the code
avelad Jun 20, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
81 changes: 33 additions & 48 deletions lib/media/segment_utils.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -179,6 +179,7 @@ shaka.media.SegmentUtils = class {
break;
case 'ac-3':
case 'ec-3':
case 'ac-4':
case 'opus':
case 'flac':
audioCodecs.push(codecLC);
Expand Down Expand Up @@ -207,6 +208,19 @@ shaka.media.SegmentUtils = class {
/** @type {?string} */
let baseBox;

const genericAudioBox = (box) => {
const parsedAudioSampleEntryBox =
shaka.util.Mp4BoxParsers.audioSampleEntry(box.reader);
channelCount = parsedAudioSampleEntryBox.channelCount;
sampleRate = parsedAudioSampleEntryBox.sampleRate;
codecBoxParser(box);
};

const genericVideoBox = (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
};

new Mp4Parser()
.box('moov', Mp4Parser.children)
.box('trak', Mp4Parser.children)
Expand Down Expand Up @@ -235,9 +249,10 @@ shaka.media.SegmentUtils = class {
// AUDIO
// These are the various boxes that signal a codec.
.box('mp4a', (box) => {
const parsedMP4ABox = shaka.util.Mp4BoxParsers.parseMP4A(box.reader);
channelCount = parsedMP4ABox.channelCount;
sampleRate = parsedMP4ABox.sampleRate;
const parsedAudioSampleEntryBox =
shaka.util.Mp4BoxParsers.audioSampleEntry(box.reader);
channelCount = parsedAudioSampleEntryBox.channelCount;
sampleRate = parsedAudioSampleEntryBox.sampleRate;
if (box.reader.hasMoreData()) {
Mp4Parser.children(box);
} else {
Expand All @@ -249,54 +264,24 @@ shaka.media.SegmentUtils = class {
audioCodecs.push(parsedESDSBox.codec);
hasAudio = true;
})
.box('ac-3', codecBoxParser)
.box('ec-3', codecBoxParser)
.box('opus', codecBoxParser)
.box('Opus', codecBoxParser)
.box('fLaC', codecBoxParser)
.box('ac-3', genericAudioBox)
.box('ec-3', genericAudioBox)
.box('ac-4', genericAudioBox)
.box('Opus', genericAudioBox)
.box('fLaC', genericAudioBox)

// VIDEO
// These are the various boxes that signal a codec.
.box('avc1', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('avc3', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('hev1', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('hvc1', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('dva1', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('dvav', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('dvh1', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('dvhe', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('vp09', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('av01', (box) => {
baseBox = box.name;
Mp4Parser.visualSampleEntry(box);
})
.box('avc1', genericVideoBox)
.box('avc3', genericVideoBox)
.box('hev1', genericVideoBox)
.box('hvc1', genericVideoBox)
.box('dva1', genericVideoBox)
.box('dvav', genericVideoBox)
.box('dvh1', genericVideoBox)
.box('dvhe', genericVideoBox)
.box('vp09', genericVideoBox)
.box('av01', genericVideoBox)
.box('avcC', (box) => {
let codecBase = baseBox || '';
switch (baseBox) {
Expand Down
8 changes: 4 additions & 4 deletions lib/util/mp4_box_parsers.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -275,11 +275,11 @@ shaka.util.Mp4BoxParsers = class {
}

/**
* Parses a MP4A box.
* Parses an audio sample entry box.
* @param {!shaka.util.DataViewReader} reader
* @return {!shaka.util.ParsedMP4ABox}
* @return {!shaka.util.ParsedAudioSampleEntryBox}
*/
static parseMP4A(reader) {
static audioSampleEntry(reader) {
reader.skip(6); // Skip "reserved"
reader.skip(2); // Skip "data_reference_index"
reader.skip(8); // Skip "reserved"
Expand Down Expand Up @@ -841,7 +841,7 @@ shaka.util.ParsedTENCBox;
*
* @exportDoc
*/
shaka.util.ParsedMP4ABox;
shaka.util.ParsedAudioSampleEntryBox;

/**
* @typedef {{
Expand Down
Binary file added test/test/assets/audio-ac-4.mp4
View file
Open in desktop
Binary file not shown.
34 changes: 31 additions & 3 deletions test/util/mp4_box_parsers_unit.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,23 +9,28 @@ describe('Mp4BoxParsers', () => {
const videoSegmentUri = '/base/test/test/assets/sintel-video-segment.mp4';

const audioInitSegmentXheAacUri = '/base/test/test/assets/audio-xhe-aac.mp4';
const audioInitSegmentAC4Uri = '/base/test/test/assets/audio-ac-4.mp4';

/** @type {!ArrayBuffer} */
let videoInitSegment;
/** @type {!ArrayBuffer} */
let videoSegment;
/** @type {!ArrayBuffer} */
let audioInitSegmentXheAac;
/** @type {!ArrayBuffer} */
let audioInitSegmentAC4;

beforeAll(async () => {
const responses = await Promise.all([
shaka.test.Util.fetch(videoInitSegmentUri),
shaka.test.Util.fetch(videoSegmentUri),
shaka.test.Util.fetch(audioInitSegmentXheAacUri),
shaka.test.Util.fetch(audioInitSegmentAC4Uri),
]);
videoInitSegment = responses[0];
videoSegment = responses[1];
audioInitSegmentXheAac = responses[2];
audioInitSegmentAC4 = responses[3];
});

it('parses init segment', () => {
Expand Down Expand Up @@ -177,9 +182,10 @@ describe('Mp4BoxParsers', () => {
.box('stbl', Mp4Parser.children)
.fullBox('stsd', Mp4Parser.sampleDescription)
.box('mp4a', (box) => {
const parsedMP4ABox = shaka.util.Mp4BoxParsers.parseMP4A(box.reader);
channelCount = parsedMP4ABox.channelCount;
sampleRate = parsedMP4ABox.sampleRate;
const parsedAudioSampleEntryBox =
shaka.util.Mp4BoxParsers.audioSampleEntry(box.reader);
channelCount = parsedAudioSampleEntryBox.channelCount;
sampleRate = parsedAudioSampleEntryBox.sampleRate;
if (box.reader.hasMoreData()) {
Mp4Parser.children(box);
}
Expand All @@ -193,6 +199,28 @@ describe('Mp4BoxParsers', () => {
expect(codec).toBe('mp4a.40.42');
});

it('parses ac-4 box for ac-4 segment', () => {
let channelCount;
let sampleRate;

const Mp4Parser = shaka.util.Mp4Parser;
new Mp4Parser()
.box('moov', Mp4Parser.children)
.box('trak', Mp4Parser.children)
.box('mdia', Mp4Parser.children)
.box('minf', Mp4Parser.children)
.box('stbl', Mp4Parser.children)
.fullBox('stsd', Mp4Parser.sampleDescription)
.box('ac-4', (box) => {
const parsedAudioSampleEntryBox =
shaka.util.Mp4BoxParsers.audioSampleEntry(box.reader);
channelCount = parsedAudioSampleEntryBox.channelCount;
sampleRate = parsedAudioSampleEntryBox.sampleRate;
}).parse(audioInitSegmentAC4, /* partialOkay= */ false);
expect(channelCount).toBe(10);
expect(sampleRate).toBe(48000);
});

/**
*
* Explanation on the Uint8Array:
Expand Down
Loading