fix(TTML): Correctly handle multiple samples in a segment (#8088)

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.

Backported to v4.9.x

Author: julijane

AUTHORS

@@ -56,6 +56,7 @@ Jesper Haug Karsrud <jesper.karsrud@gmail.com>
 Johan Sundström <oyasumi@gmail.com>
 Jonas Birmé <jonas.birme@eyevinn.se>
 Jozef Chúťka <jozefchutka@gmail.com>
+Juliane Holzt <juliane@box.fqdn.org>
 Jun Hong Chong <chongjunhong@gmail.com>
 Jürgen Kartnaller <kartnaller@lovelysystems.com>
 Justin Swaney <justin.mark.swaney@gmail.com>

CONTRIBUTORS

@@ -84,6 +84,7 @@ Jonas Birmé <jonas.birme@eyevinn.se>
 Jono Ward <jonoward@gmail.com>
 Jozef Chúťka <jozefchutka@gmail.com>
 Julian Domingo <juliandomingo@google.com>
+Juliane Holzt <juliane@box.fqdn.org>
 Jun Hong Chong <chongjunhong@gmail.com>
 Jürgen Kartnaller <kartnaller@lovelysystems.com>
 Justin Swaney <justin.mark.swaney@gmail.com>

lib/text/mp4_ttml_parser.js

@@ -6,10 +6,14 @@
 
 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');
 
 
 /**
@@ -78,25 +82,109 @@ 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>} */
+    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) => {
+          const reader = box.reader;
+          const entryCount = reader.readUint32();
+          let currentSampleNum = -1;
+          for (let i = 0; i < entryCount; i++) {
+            const sampleDelta = reader.readUint32();
+            currentSampleNum += sampleDelta;
+            const subsampleCount = reader.readUint16();
+            const subsampleSizes = [];
+            for (let j = 0; j < subsampleCount; j++) {
+              if (box.version == 1) {
+                subsampleSizes.push(reader.readUint32());
+              } else {
+                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.
-          payload = payload.concat(this.parser_.parseMedia(data, time, uri));
+          // 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]);
+        let subOffset = subSampleSizes[0];
+        for (let i = 1; i < subSampleSizes.length; i++) {
+          subOffset += subSampleSizes[i];
+        }
+        payload = payload.concat(
+            this.parser_.parseMedia(contentData, time, uri));
+      } else {
+        payload = payload.concat(
+            this.parser_.parseMedia(sampleData, time, uri));
+      }
+    }
+
     return payload;
   }
 };

test/test/assets/ttml-segment-multiple-mdat.mp4

@@ -8,7 +8,9 @@ describe('Mp4TtmlParser', () => {
   const ttmlInitSegmentUri = '/base/test/test/assets/ttml-init.mp4';
   const ttmlSegmentUri = '/base/test/test/assets/ttml-segment.mp4';
   const ttmlSegmentMultipleMDATUri =
-      '/base/test/test/assets/ttml-segment-multiplemdat.mp4';
+      '/base/test/test/assets/ttml-segment-multiple-mdat.mp4';
+  const ttmlSegmentMultipleSampleUri =
+      '/base/test/test/assets/ttml-segment-multiple-sample.mp4';
   const audioInitSegmentUri = '/base/test/test/assets/sintel-audio-init.mp4';
 
   /** @type {!Uint8Array} */
@@ -18,19 +20,23 @@ describe('Mp4TtmlParser', () => {
   /** @type {!Uint8Array} */
   let ttmlSegmentMultipleMDAT;
   /** @type {!Uint8Array} */
+  let ttmlSegmentMultipleSample;
+  /** @type {!Uint8Array} */
   let audioInitSegment;
 
   beforeAll(async () => {
     const responses = await Promise.all([
       shaka.test.Util.fetch(ttmlInitSegmentUri),
       shaka.test.Util.fetch(ttmlSegmentUri),
       shaka.test.Util.fetch(ttmlSegmentMultipleMDATUri),
+      shaka.test.Util.fetch(ttmlSegmentMultipleSampleUri),
       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]);
-    audioInitSegment = shaka.util.BufferUtils.toUint8(responses[3]);
+    ttmlSegmentMultipleSample = shaka.util.BufferUtils.toUint8(responses[3]);
+    audioInitSegment = shaka.util.BufferUtils.toUint8(responses[4]);
   });
 
   it('parses init segment', () => {
@@ -50,8 +56,24 @@ 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', () => {