feat: Caption services (608/708) metadata (#1138)

Add an option for caption services metadata in case the user wants to specify labels for 608/708 captions, override the ones provided in the manifest, or needs to add more information like character encoding (this isn't currently available but will be added some time in the future).
For HLS, an EXT-X-MEDIA tag can be specified with an INSTREAM-ID attribute. We already support this. https://datatracker.ietf.org/doc/html/draft-pantos-hls-rfc8216bis-08#section-4.4.6.1
This PR updated mpd-parser which uses the ANSI 214 supplemental spec section 7.2 to parse out the same information from MPD files. videojs/mpd-parser#131.

Adds a property called captionServices which has properties of the caption service IDs like CC1 or SERVICE1 and allows a user to specify a language and label.

Author: gkatsev

README.md

@@ -58,6 +58,9 @@ Video.js Compatibility: 6.0, 7.0
       - [cacheEncryptionKeys](#cacheencryptionkeys)
       - [handlePartialData](#handlepartialdata)
       - [liveRangeSafeTimeDelta](#liverangesafetimedelta)
+      - [captionServices](#captionservices)
+        - [Format](#format)
+        - [Example](#example)
   - [Runtime Properties](#runtime-properties)
     - [vhs.playlists.master](#vhsplaylistsmaster)
     - [vhs.playlists.media](#vhsplaylistsmedia)
@@ -471,6 +474,47 @@ This option defaults to `false`.
 * Default: [`SAFE_TIME_DELTA`](https://github.com/videojs/http-streaming/blob/e7cb63af010779108336eddb5c8fd138d6390e95/src/ranges.js#L17)
 * Allow to  re-define length (in seconds) of time delta when you compare current time and the end of the buffered range.
 
+##### captionServices
+* Type: `object`
+* Default: undefined
+* Provide extra information, like a label or a language, for instream (608 and 708) captions.
+
+The captionServices options object has properties that map to the caption services. Each property is an object itself that includes several properties, like a label or language.
+
+For 608 captions, the service names are `CC1`, `CC2`, `CC3`, and `CC4`. For 708 captions, the service names are `SERVICEn` where `n` is a digit between `1` and `63`.
+###### Format
+```js
+{
+  vhs: {
+    captionServices: {
+      [serviceName]: {
+        language: String, // optional
+        label: String, // optional
+        default: boolean // optional
+      }
+    }
+  }
+}
+```
+###### Example
+```js
+{
+  vhs: {
+    captionServices: {
+      CC1: {
+        language: 'en',
+        label: 'English'
+      },
+      SERVICE1: {
+        langauge: 'kr',
+        label: 'Korean',
+        default: true
+      }
+    }
+  }
+}
+```
+
 ### Runtime Properties
 Runtime properties are attached to the tech object when HLS is in
 use. You can get a reference to the VHS source handler like this:

package-lock.json

@@ -61,7 +61,7 @@
     "aes-decrypter": "3.1.2",
     "global": "^4.4.0",
     "m3u8-parser": "4.7.0",
-    "mpd-parser": "0.16.0",
+    "mpd-parser": "0.17.0",
     "mux.js": "5.11.0",
     "video.js": "^6 || ^7"
   },

package.json

@@ -638,23 +638,39 @@ export const initialize = {
       for (const variantLabel in mediaGroups[type][groupId]) {
         const properties = mediaGroups[type][groupId][variantLabel];
 
-        // We only support CEA608 captions for now, so ignore anything that
-        // doesn't use a CCx INSTREAM-ID
-        if (!properties.instreamId.match(/CC\d/)) {
+        // Look for either 608 (CCn) or 708 (SERVICEn) caption services
+        if (!/^(?:CC|SERVICE)/.test(properties.instreamId)) {
           continue;
         }
 
+        const captionServices = tech.options_.vhs && tech.options_.vhs.captionServices || {};
+
+        let newProps = {
+          label: variantLabel,
+          language: properties.language,
+          instreamId: properties.instreamId,
+          default: properties.default && properties.autoselect
+        };
+
+        if (captionServices[newProps.instreamId]) {
+          newProps = videojs.mergeOptions(newProps, captionServices[newProps.instreamId]);
+        }
+
+        if (newProps.default === undefined) {
+          delete newProps.default;
+        }
+
         // No PlaylistLoader is required for Closed-Captions because the captions are
         // embedded within the video stream
         groups[groupId].push(videojs.mergeOptions({ id: variantLabel }, properties));
 
         if (typeof tracks[variantLabel] === 'undefined') {
           const track = tech.addRemoteTextTrack({
-            id: properties.instreamId,
+            id: newProps.instreamId,
             kind: 'captions',
-            default: properties.default && properties.autoselect,
-            language: properties.language,
-            label: variantLabel
+            default: newProps.default,
+            language: newProps.language,
+            label: newProps.label
           }, false).track;
 
           tracks[variantLabel] = track;

src/media-groups.js

@@ -16,20 +16,45 @@ export const createCaptionsTrackIfNotExists = function(inbandTextTracks, tech, c
   if (!inbandTextTracks[captionStream]) {
     tech.trigger({type: 'usage', name: 'vhs-608'});
     tech.trigger({type: 'usage', name: 'hls-608'});
-    const track = tech.textTracks().getTrackById(captionStream);
+
+    let instreamId = captionStream;
+
+    // we need to translate SERVICEn for 708 to how mux.js currently labels them
+    if (/^cc708_/.test(captionStream)) {
+      instreamId = 'SERVICE' + captionStream.split('_')[1];
+    }
+
+    const track = tech.textTracks().getTrackById(instreamId);
 
     if (track) {
       // Resuse an existing track with a CC# id because this was
       // very likely created by videojs-contrib-hls from information
       // in the m3u8 for us to use
       inbandTextTracks[captionStream] = track;
     } else {
+      // This section gets called when we have caption services that aren't specified in the manifest.
+      // Manifest level caption services are handled in media-groups.js under CLOSED-CAPTIONS.
+      const captionServices = tech.options_.vhs && tech.options_.vhs.captionServices || {};
+      let label = captionStream;
+      let language = captionStream;
+      let def = false;
+      const captionService = captionServices[instreamId];
+
+      if (captionService) {
+        label = captionService.label;
+        language = captionService.language;
+        def = captionService.default;
+      }
+
       // Otherwise, create a track with the default `CC#` label and
       // without a language
       inbandTextTracks[captionStream] = tech.addRemoteTextTrack({
         kind: 'captions',
-        id: captionStream,
-        label: captionStream
+        id: instreamId,
+        // TODO: investigate why this doesn't seem to turn the caption on by default
+        default: def,
+        label,
+        language
       }, false).track;
     }
   }

src/util/text-tracks.js

@@ -45,6 +45,7 @@ export const LoaderCommonHooks = {
     this.fakeVhs = {
       xhr: xhrFactory(),
       tech_: {
+        options_: {},
         paused: () => this.paused,
         playbackRate: () => this.playbackRate,
         currentTime: () => this.currentTime,

test/loader-common.js

@@ -3104,7 +3104,7 @@ QUnit.test('parses codec from muxed fmp4 init segment', function(assert) {
 });
 
 QUnit.test(
-  'adds only CEA608 closed-caption tracks when a master playlist is loaded',
+  'adds CEA608 closed-caption tracks when a master playlist is loaded',
   function(assert) {
     this.requests.length = 0;
     this.player.dispose();
@@ -3143,15 +3143,77 @@ QUnit.test(
       .map(cap => Object.assign({name: cap.id}, cap));
 
     assert.equal(capsArr.length, 4, '4 closed-caption tracks defined in playlist');
-    assert.equal(addedCaps.length, 2, '2 CEA608 tracks added internally');
+    assert.equal(addedCaps.length, 4, '4 tracks, 2 608 and 2 708 tracks, added internally');
     assert.equal(addedCaps[0].instreamId, 'CC1', 'first 608 track is CC1');
-    assert.equal(addedCaps[1].instreamId, 'CC3', 'second 608 track is CC3');
+    assert.equal(addedCaps[2].instreamId, 'CC3', 'second 608 track is CC3');
+
+    const textTracks = this.player.textTracks();
+
+    assert.equal(
+      textTracks[1].id, addedCaps[0].instreamId,
+      'text track 1\'s id is CC\'s instreamId'
+    );
+    assert.equal(
+      textTracks[2].id, addedCaps[1].instreamId,
+      'text track 2\'s id is CC\'s instreamId'
+    );
+    assert.equal(
+      textTracks[1].label, addedCaps[0].name,
+      'text track 1\'s label is CC\'s name'
+    );
+    assert.equal(
+      textTracks[2].label, addedCaps[1].name,
+      'text track 2\'s label is CC\'s name'
+    );
+  }
+);
+
+QUnit.test(
+  'adds CEA708 closed-caption tracks when a master playlist is loaded',
+  function(assert) {
+    this.requests.length = 0;
+    this.player.dispose();
+    this.player = createPlayer();
+    this.player.src({
+      src: 'manifest/master-captions.m3u8',
+      type: 'application/vnd.apple.mpegurl'
+    });
+
+    // wait for async player.src to complete
+    this.clock.tick(1);
+
+    const masterPlaylistController = this.player.tech_.vhs.masterPlaylistController_;
+
+    assert.equal(this.player.textTracks().length, 1, 'one text track to start');
+    assert.equal(
+      this.player.textTracks()[0].label,
+      'segment-metadata',
+      'only segment-metadata text track'
+    );
+
+    // master, contains media groups for captions
+    this.standardXHRResponse(this.requests.shift());
+
+    // we wait for loadedmetadata before setting caption tracks, so we need to wait for a
+    // media playlist
+    assert.equal(this.player.textTracks().length, 1, 'only one text track after master');
+
+    // media
+    this.standardXHRResponse(this.requests.shift());
+
+    const master = masterPlaylistController.masterPlaylistLoader_.master;
+    const caps = master.mediaGroups['CLOSED-CAPTIONS'].CCs;
+    const capsArr = Object.keys(caps).map(key => Object.assign({name: key}, caps[key]));
+    const addedCaps = masterPlaylistController.mediaTypes_['CLOSED-CAPTIONS'].groups.CCs
+      .map(cap => Object.assign({name: cap.id}, cap));
+
+    assert.equal(capsArr.length, 4, '4 closed-caption tracks defined in playlist');
+    assert.equal(addedCaps.length, 4, '4 tracks, 2 608 and 2 708 tracks, added internally');
+    assert.equal(addedCaps[1].instreamId, 'SERVICE1', 'first 708 track is SERVICE1');
+    assert.equal(addedCaps[3].instreamId, 'SERVICE3', 'second 708 track is SERVICE3');
 
     const textTracks = this.player.textTracks();
 
-    assert.equal(textTracks.length, 3, '2 text tracks were added');
-    assert.equal(textTracks[1].mode, 'disabled', 'track starts disabled');
-    assert.equal(textTracks[2].mode, 'disabled', 'track starts disabled');
     assert.equal(
       textTracks[1].id, addedCaps[0].instreamId,
       'text track 1\'s id is CC\'s instreamId'

test/master-playlist-controller.test.js

@@ -1078,6 +1078,7 @@ QUnit.module('MediaGroups', function() {
         masterPlaylistLoader: {master: this.master},
         vhs: {},
         tech: {
+          options_: {},
           addRemoteTextTrack(track) {
             return { track };
           }
@@ -1258,7 +1259,18 @@ QUnit.module('MediaGroups', function() {
         en608: { language: 'en', default: true, autoselect: true, instreamId: 'CC1' },
         en708: { language: 'en', instreamId: 'SERVICE1' },
         fr608: { language: 'fr', instreamId: 'CC3' },
-        fr708: { language: 'fr', instreamId: 'SERVICE3' }
+        fr708: { language: 'fr', instreamId: 'SERVICE3' },
+        kr708: { language: 'kor', instreamId: 'SERVICE4' }
+      };
+
+      // verify that captionServices option can modify properties
+      this.settings.tech.options_.vhs = {
+        captionServices: {
+          SERVICE4: {
+            label: 'Korean',
+            default: true
+          }
+        }
       };
 
       MediaGroups.initialize[type](type, this.settings);
@@ -1268,13 +1280,23 @@ QUnit.module('MediaGroups', function() {
         {
           CCs: [
             { id: 'en608', default: true, autoselect: true, language: 'en', instreamId: 'CC1' },
-            { id: 'fr608', language: 'fr', instreamId: 'CC3' }
+            { id: 'en708', language: 'en', instreamId: 'SERVICE1' },
+            { id: 'fr608', language: 'fr', instreamId: 'CC3' },
+            { id: 'fr708', language: 'fr', instreamId: 'SERVICE3' },
+            { id: 'kr708', language: 'kor', instreamId: 'SERVICE4' }
           ]
         }, 'creates group properties'
       );
       assert.ok(this.mediaTypes[type].tracks.en608, 'created text track');
       assert.ok(this.mediaTypes[type].tracks.fr608, 'created text track');
       assert.equal(this.mediaTypes[type].tracks.en608.default, true, 'en608 track auto selected');
+      assert.deepEqual(this.mediaTypes[type].tracks.kr708, {
+        id: 'SERVICE4',
+        kind: 'captions',
+        language: 'kor',
+        label: 'Korean',
+        default: true
+      }, 'kr708 fields are overriden by the options');
     }
   );