Skip to content

Commit

Permalink
fix(TTML): Correctly handle multiple samples in a segment (#8088)
Browse files Browse the repository at this point in the history 
Fixes #8087

Implements handling of multiple samples in a MP4/ISOBMFF/DASH TTML
segment/fragment. Such segments are allowed by ISO14496-12 and
ISO23000-19. gpac creates such segments. The prior code just treated the
full MDAT as one TTML XML document and tried to parse it in whole
without accounting for sample(s). A testcase is included which was
created by taking the testdata from ttml-segment.mp4 and splitting the
subtitles into two independent TTML-XML documents, which then were put
as individual samples.

The testdata for the prior existing multiple MDAT testcase was invalid.
It was created by taking the same ttml-segment.mp4 as a source and just
duplicating the MDAT box, but without then also fixing the TRUN box. The
duplicated data was thus not referenced. The test case still worked,
because the prior code did not look at the TRUN box and the sample
specification at all and just handled any full MDAT box = 1 sample. The
testdata was replaced with a new file, which is basically the same as
for the multiple samples case, but with the two samples split into two
MDAT boxes.
  • Loading branch information
@julijane @avelad
julijane authored and avelad committed Feb 17, 2025
1 parent 33037d1 commit 4446d4d
Show file tree
Hide file tree
Showing 6 changed files with 110 additions and 34 deletions.
1 change: 1 addition & 0 deletions AUTHORS
View file
Original file line number Diff line number Diff line change
Expand Up @@ -56,6 +56,7 @@ Jesper Haug Karsrud <[email protected]>
Johan Sundström <[email protected]>
Jonas Birmé <[email protected]>
Jozef Chúťka <[email protected]>
Juliane Holzt <[email protected]>
Jun Hong Chong <[email protected]>
Jürgen Kartnaller <[email protected]>
Justin Swaney <[email protected]>
Expand Down
1 change: 1 addition & 0 deletions CONTRIBUTORS
View file
Original file line number Diff line number Diff line change
Expand Up @@ -86,6 +86,7 @@ Jono Ward <[email protected]>
Jozef Chúťka <[email protected]>
Juan Manuel Tomás <[email protected]>
Julian Domingo <[email protected]>
Juliane Holzt <[email protected]>
Jun Hong Chong <[email protected]>
Jürgen Kartnaller <[email protected]>
Justin Swaney <[email protected]>
Expand Down
110 changes: 81 additions & 29 deletions lib/text/mp4_ttml_parser.js
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,10 +6,12 @@

goog.provide('shaka.text.Mp4TtmlParser');

goog.require('goog.asserts');
goog.require('shaka.text.TextEngine');
goog.require('shaka.text.TtmlTextParser');
goog.require('shaka.util.BufferUtils');
goog.require('shaka.util.Error');
goog.require('shaka.util.Mp4BoxParsers');
goog.require('shaka.util.Mp4Parser');
goog.require('shaka.util.Uint8ArrayUtils');

Expand Down Expand Up @@ -80,67 +82,117 @@ shaka.text.Mp4TtmlParser = class {
parseMedia(data, time, uri) {
const Mp4Parser = shaka.util.Mp4Parser;

let sawMDAT = false;
let payload = [];
let defaultSampleSize = null;

/** @type {!Array<Uint8Array>} */
const mdats = [];

/* @type {!Map<number,!Array<number>>} */
const subSampleSizesPerSample = new Map();

/** @type {!Array<number>} */
let subSizes = [];
const sampleSizes = [];

const parser = new Mp4Parser()
.box('moof', Mp4Parser.children)
.box('traf', Mp4Parser.children)
.fullBox('tfhd', (box) => {
goog.asserts.assert(
box.flags != null,
'A TFHD box should have a valid flags value');
const parsedTFHDBox = shaka.util.Mp4BoxParsers.parseTFHD(
box.reader, box.flags);
defaultSampleSize = parsedTFHDBox.defaultSampleSize;
})
.fullBox('trun', (box) => {
goog.asserts.assert(
box.version != null,
'A TRUN box should have a valid version value');
goog.asserts.assert(
box.flags != null,
'A TRUN box should have a valid flags value');

const parsedTRUNBox = shaka.util.Mp4BoxParsers.parseTRUN(
box.reader, box.version, box.flags);

for (const sample of parsedTRUNBox.sampleData) {
const sampleSize =
sample.sampleSize || defaultSampleSize || 0;
sampleSizes.push(sampleSize);
}
})
.fullBox('subs', (box) => {
subSizes = [];
const reader = box.reader;
const entryCount = reader.readUint32();
let currentSampleNum = -1;
for (let i = 0; i < entryCount; i++) {
reader.readUint32(); // sample_delta
const sampleDelta = reader.readUint32();
currentSampleNum += sampleDelta;
const subsampleCount = reader.readUint16();
const subsampleSizes = [];
for (let j = 0; j < subsampleCount; j++) {
if (box.version == 1) {
subSizes.push(reader.readUint32());
subsampleSizes.push(reader.readUint32());
} else {
subSizes.push(reader.readUint16());
subsampleSizes.push(reader.readUint16());
}
reader.readUint8(); // priority
reader.readUint8(); // discardable
reader.readUint32(); // codec_specific_parameters
}
subSampleSizesPerSample.set(currentSampleNum, subsampleSizes);
}
})
.box('mdat', Mp4Parser.allData((data) => {
sawMDAT = true;
// Join this to any previous payload, in case the mp4 has multiple
// mdats.
if (subSizes.length) {
const contentData =
shaka.util.BufferUtils.toUint8(data, 0, subSizes[0]);
const images = [];
let offset = subSizes[0];
for (let i = 1; i < subSizes.length; i++) {
const imageData =
shaka.util.BufferUtils.toUint8(data, offset, subSizes[i]);
const raw =
shaka.util.Uint8ArrayUtils.toStandardBase64(imageData);
images.push('data:image/png;base64,' + raw);
offset += subSizes[i];
}
payload = payload.concat(
this.parser_.parseMedia(contentData, time, uri, images));
} else {
payload = payload.concat(
this.parser_.parseMedia(data, time, uri, /* images= */ []));
}
// We collect all of the mdats first, before parsing any of them.
// This is necessary in case the mp4 has multiple mdats.
mdats.push(data);
}));
parser.parse(data, /* partialOkay= */ false);

if (!sawMDAT) {
if (mdats.length == 0) {
throw new shaka.util.Error(
shaka.util.Error.Severity.CRITICAL,
shaka.util.Error.Category.TEXT,
shaka.util.Error.Code.INVALID_MP4_TTML);
}

const fullData =
shaka.util.Uint8ArrayUtils.concat(...mdats);

let sampleOffset = 0;
for (let sampleNum = 0; sampleNum < sampleSizes.length; sampleNum++) {
const sampleData =
shaka.util.BufferUtils.toUint8(fullData, sampleOffset,
sampleSizes[sampleNum]);
sampleOffset += sampleSizes[sampleNum];

const subSampleSizes = subSampleSizesPerSample.get(sampleNum);

if (subSampleSizes && subSampleSizes.length) {
const contentData =
shaka.util.BufferUtils.toUint8(sampleData, 0, subSampleSizes[0]);
const images = [];
let subOffset = subSampleSizes[0];
for (let i = 1; i < subSampleSizes.length; i++) {
const imageData =
shaka.util.BufferUtils.toUint8(data, subOffset,
subSampleSizes[i]);
const raw =
shaka.util.Uint8ArrayUtils.toStandardBase64(imageData);
images.push('data:image/png;base64,' + raw);
subOffset += subSampleSizes[i];
}
payload = payload.concat(
this.parser_.parseMedia(contentData, time, uri, images));
} else {
payload = payload.concat(
this.parser_.parseMedia(sampleData, time, uri,
/* images= */ []));
}
}

return payload;
}
};
Expand Down
Binary file modified test/test/assets/ttml-segment-multiple-mdat.mp4
View file
Binary file not shown.
Binary file added test/test/assets/ttml-segment-multiple-sample.mp4
View file
Binary file not shown.
32 changes: 27 additions & 5 deletions test/text/mp4_ttml_parser_unit.js
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,8 @@ describe('Mp4TtmlParser', () => {
const ttmlSegmentUri = '/base/test/test/assets/ttml-segment.mp4';
const ttmlSegmentMultipleMDATUri =
'/base/test/test/assets/ttml-segment-multiple-mdat.mp4';
const ttmlSegmentMultipleSampleUri =
'/base/test/test/assets/ttml-segment-multiple-sample.mp4';
const imscImageInitSegmentUri =
'/base/test/test/assets/imsc-image-init.cmft';
const imscImageSegmentUri =
Expand All @@ -22,6 +24,8 @@ describe('Mp4TtmlParser', () => {
/** @type {!Uint8Array} */
let ttmlSegmentMultipleMDAT;
/** @type {!Uint8Array} */
let ttmlSegmentMultipleSample;
/** @type {!Uint8Array} */
let imscImageInitSegment;
/** @type {!Uint8Array} */
let imscImageSegment;
Expand All @@ -33,16 +37,18 @@ describe('Mp4TtmlParser', () => {
shaka.test.Util.fetch(ttmlInitSegmentUri),
shaka.test.Util.fetch(ttmlSegmentUri),
shaka.test.Util.fetch(ttmlSegmentMultipleMDATUri),
shaka.test.Util.fetch(ttmlSegmentMultipleSampleUri),
shaka.test.Util.fetch(imscImageInitSegmentUri),
shaka.test.Util.fetch(imscImageSegmentUri),
shaka.test.Util.fetch(audioInitSegmentUri),
]);
ttmlInitSegment = shaka.util.BufferUtils.toUint8(responses[0]);
ttmlSegment = shaka.util.BufferUtils.toUint8(responses[1]);
ttmlSegmentMultipleMDAT = shaka.util.BufferUtils.toUint8(responses[2]);
imscImageInitSegment = shaka.util.BufferUtils.toUint8(responses[3]);
imscImageSegment = shaka.util.BufferUtils.toUint8(responses[4]);
audioInitSegment = shaka.util.BufferUtils.toUint8(responses[5]);
ttmlSegmentMultipleSample = shaka.util.BufferUtils.toUint8(responses[3]);
imscImageInitSegment = shaka.util.BufferUtils.toUint8(responses[4]);
imscImageSegment = shaka.util.BufferUtils.toUint8(responses[5]);
audioInitSegment = shaka.util.BufferUtils.toUint8(responses[6]);
});

it('parses init segment', () => {
Expand All @@ -62,8 +68,24 @@ describe('Mp4TtmlParser', () => {
expect(ret[0].nestedCues.length).toBe(1);
expect(ret[1].nestedCues.length).toBe(1);
// Cues.
expect(ret[0].nestedCues[0].nestedCues.length).toBe(10);
expect(ret[1].nestedCues[0].nestedCues.length).toBe(10);
expect(ret[0].nestedCues[0].nestedCues.length).toBe(5);
expect(ret[1].nestedCues[0].nestedCues.length).toBe(5);
});

it('handles media segments with multiple sample', () => {
const parser = new shaka.text.Mp4TtmlParser();
parser.parseInit(ttmlInitSegment);
const time =
{periodStart: 0, segmentStart: 0, segmentEnd: 60, vttOffset: 0};
const ret = parser.parseMedia(ttmlSegmentMultipleSample, time, null);
// Bodies.
expect(ret.length).toBe(2);
// Divs.
expect(ret[0].nestedCues.length).toBe(1);
expect(ret[1].nestedCues.length).toBe(1);
// Cues.
expect(ret[0].nestedCues[0].nestedCues.length).toBe(5);
expect(ret[1].nestedCues[0].nestedCues.length).toBe(5);
});

it('accounts for offset', () => {
Expand Down

0 comments on commit 4446d4d

Please sign in to comment.