Merge pull request videojs#42 from imbcmdth/caption-id3-parsing

Caption ID3 Parsing Improvements

lib/codecs/h264.js

@@ -152,6 +152,7 @@ H264Stream = function() {
       break;
     case 0x06:
       event.nalUnitType = 'sei_rbsp';
+      event.escapedRBSP = discardEmulationPreventionBytes(data.subarray(1));
       break;
     case 0x07:
       event.nalUnitType = 'seq_parameter_set_rbsp';

lib/m2ts/caption-stream.js

@@ -20,38 +20,62 @@
   // payload type field to indicate how they are to be
   // interpreted. CEAS-708 caption content is always transmitted with
   // payload type 0x04.
-  var USER_DATA_REGISTERED_ITU_T_T35 = 4;
+  var USER_DATA_REGISTERED_ITU_T_T35 = 4,
+      RBSP_TRAILING_BITS = 128;
 
   /**
    * Parse a supplemental enhancement information (SEI) NAL unit.
+   * Stops parsing once a message of type ITU T T35 has been found.
    *
    * @param bytes {Uint8Array} the bytes of a SEI NAL unit
    * @return {object} the parsed SEI payload
    * @see Rec. ITU-T H.264, 7.3.2.3.1
    */
   var parseSei = function(bytes) {
-    var result = {
-      payloadType: -1,
-      payloadSize: 0,
-    }, i;
-
-    // parse the payload type
-    // if the payload type is not user_data_registered_itu_t_t35,
-    // don't bother parsing any further
-    if (bytes[1] !== USER_DATA_REGISTERED_ITU_T_T35) {
-      return result;
-    }
-    result.payloadType = USER_DATA_REGISTERED_ITU_T_T35;
+    var
+      i = 0,
+      result = {
+        payloadType: -1,
+        payloadSize: 0,
+      },
+      payloadType = 0,
+      payloadSize = 0;
+
+    // go through the sei_rbsp parsing each each individual sei_message
+    while (i < bytes.byteLength) {
+      // stop once we have hit the end of the sei_rbsp
+      if (bytes[i] === RBSP_TRAILING_BITS) {
+        break;
+      }
 
-    // parse the payload size
-    for (i = 2; i < bytes.length && bytes[i] === 0xff; i++) {
-      result.payloadSize += 255;
-    }
-    result.payloadSize <<= 8;
-    result.payloadSize |= bytes[i];
-    i++;
+      // Parse payload type
+      while (bytes[i] === 0xFF) {
+        payloadType += 255;
+        i++;
+      }
+      payloadType += bytes[i++];
 
-    result.payload = bytes.subarray(i, i + result.payloadSize);
+      // Parse payload size
+      while (bytes[i] === 0xFF) {
+        payloadSize += 255;
+        i++;
+      }
+      payloadSize += bytes[i++];
+
+      // this sei_message is a 608/708 caption so save it and break
+      // there can only ever be one caption message in a frame's sei
+      if (!result.payload && payloadType === USER_DATA_REGISTERED_ITU_T_T35) {
+        result.payloadType = payloadType;
+        result.payloadSize = payloadSize;
+        result.payload = bytes.subarray(i, i + payloadSize);
+        break;
+      }
+
+      // skip the payload and parse the next message
+      i += payloadSize;
+      payloadType = 0;
+      payloadSize = 0;
+    }
 
     return result;
   };
@@ -115,21 +139,28 @@
   };
 
   var CaptionStream = function() {
+    var self = this;
     CaptionStream.prototype.init.call(this);
 
+    this.captionPackets_ = [];
+
     this.field1_ = new Cea608Stream();
+
+    // forward data and done events from field1_ to this CaptionStream
     this.field1_.on('data', this.trigger.bind(this, 'data'));
+    this.field1_.on('done', this.trigger.bind(this, 'done'));
   };
   CaptionStream.prototype = new muxjs.utils.Stream();
   CaptionStream.prototype.push = function(event) {
-    var sei, userData, captionPackets, i;
+    var sei, userData, captionPackets;
 
     // only examine SEI NALs
     if (event.nalUnitType !== 'sei_rbsp') {
       return;
     }
+
     // parse the sei
-    sei = parseSei(event.data);
+    sei = parseSei(event.escapedRBSP);
 
     // ignore everything but user_data_registered_itu_t_t35
     if (sei.payloadType !== USER_DATA_REGISTERED_ITU_T_T35) {
@@ -144,22 +175,35 @@
       return;
     }
 
-    // parse out CC data packets
-    captionPackets = parseCaptionPackets(event.pts, userData);
+    // parse out CC data packets and save them for later
+    this.captionPackets_ = this.captionPackets_.concat(parseCaptionPackets(event.pts, userData));
+  };
 
-    // send the data to the appropriate field
-    for (i = 0; i < captionPackets.length; i++) {
-      if (captionPackets[i].type === 0) {
-        this.field1_.push(captionPackets[i]);
-      }
+  CaptionStream.prototype.flush = function () {
+    // make sure we actually parsed captions before proceeding
+    if (!this.captionPackets_.length) {
+      this.field1_.flush();
+      return;
     }
-  };
 
+    // sort caption byte-pairs based on their PTS values
+    this.captionPackets_.sort(function(a, b) {
+      return a.pts - b.pts;
+    });
+
+    // Push each caption into Cea608Stream
+    this.captionPackets_.forEach(this.field1_.push, this.field1_);
+
+    this.captionPackets_.length = 0;
+    this.field1_.flush();
+    return;
+  };
   // ----------------------
   // Session to Application
   // ----------------------
 
   var BASIC_CHARACTER_TRANSLATION = {
+    0x2a: 0xe1,
     0x5c: 0xe9,
     0x5e: 0xed,
     0x5f: 0xf3,
@@ -168,7 +212,6 @@
     0x7c: 0xf7,
     0x7d: 0xd1,
     0x7e: 0xf1,
-    0x2a: 0xe1,
     0x7f: 0x2588
   };
 
@@ -186,8 +229,8 @@
       ROLL_UP_2_ROWS             = 0x1425,
       ROLL_UP_3_ROWS             = 0x1426,
       ROLL_UP_4_ROWS             = 0x1427,
+      RESUME_DIRECT_CAPTIONING   = 0x1429,
       CARRIAGE_RETURN            = 0x142d,
-
       // Erasure
       BACKSPACE                  = 0x1421,
       ERASE_DISPLAYED_MEMORY     = 0x142c,
@@ -217,16 +260,29 @@
     this.startPts_ = 0;
     this.displayed_ = createDisplayBuffer();
     this.nonDisplayed_ = createDisplayBuffer();
+    this.lastControlCode_ = null;
 
     this.push = function(packet) {
-      var data, swap, charCode;
+      var data, swap, char0, char1;
       // remove the parity bits
       data = packet.ccData & 0x7f7f;
 
+      // ignore duplicate control codes
+      if (data === this.lastControlCode_) {
+        this.lastControlCode_ = null;
+        return;
+      }
+
+      // Store control codes
+      if ((data & 0xf000) === 0x1000) {
+        this.lastControlCode_ = data;
+      } else {
+        this.lastControlCode_ = null;
+      }
+
       switch (data) {
       case PADDING:
         break;
-
       case RESUME_CAPTION_LOADING:
         this.mode_ = 'popOn';
         break;
@@ -275,17 +331,25 @@
       case ERASE_NON_DISPLAYED_MEMORY:
         this.nonDisplayed_ = createDisplayBuffer();
         break;
-
       default:
-        charCode = data >>> 8;
+        char0 = data >>> 8;
+        char1 = data & 0xff;
+
+        // Look for a Channel 1 Preamble Address Code
+        if (char0 >= 0x10 && char0 <= 0x17 &&
+            char1 >= 0x40 && char1 <= 0x7F &&
+            (char0 !== 0x10 || char1 < 0x60)) {
+          // Follow Safari's lead and replace the PAC with a space
+          char0 = char1 = 0x20;
+        }
 
         // ignore unsupported control codes
-        if ((charCode & 0xf0) === 0x10) {
+        if ((char0 & 0xf0) === 0x10) {
           return;
         }
 
         // character handling is dependent on the current mode
-        this[this.mode_](packet.pts, charCode, data & 0xff);
+        this[this.mode_](packet.pts, char0, char1);
         break;
       }
     };

lib/m2ts/metadata-stream.js

@@ -25,6 +25,12 @@
     parseIso88591 = function(bytes, start, end) {
       return window.unescape(percentEncode(bytes, start, end));
     },
+    parseSyncSafeInteger = function (data) {
+      return (data[0] << 21) |
+             (data[1] << 14) |
+             (data[2] << 7) |
+             (data[3]);
+    },
     tagParsers = {
       'TXXX': function(tag) {
         var i;
@@ -142,10 +148,7 @@
         // last four bytes of the ID3 header.
         // The most significant bit of each byte is dropped and the
         // results concatenated to recover the actual value.
-        tagSize = (chunk.data[6] << 21) |
-                  (chunk.data[7] << 14) |
-                  (chunk.data[8] << 7) |
-                  (chunk.data[9]);
+        tagSize = parseSyncSafeInteger(chunk.data.subarray(6, 10));
 
         // ID3 reports the tag size excluding the header but it's more
         // convenient for our comparisons to include it
@@ -176,26 +179,17 @@
       if (tag.data[5] & 0x40) {
         // advance the frame start past the extended header
         frameStart += 4; // header size field
-        frameStart += (tag.data[10] << 24) |
-                      (tag.data[11] << 16) |
-                      (tag.data[12] << 8)  |
-                      (tag.data[13]);
+        frameStart += parseSyncSafeInteger(tag.data.subarray(10, 14));
 
         // clip any padding off the end
-        tagSize -= (tag.data[16] << 24) |
-                   (tag.data[17] << 16) |
-                   (tag.data[18] << 8)  |
-                   (tag.data[19]);
+        tagSize -= parseSyncSafeInteger(tag.data.subarray(16, 20));
       }
 
       // parse one or more ID3 frames
       // http://id3.org/id3v2.3.0#ID3v2_frame_overview
       do {
         // determine the number of bytes in this frame
-        frameSize = (tag.data[frameStart + 4] << 24) |
-                    (tag.data[frameStart + 5] << 16) |
-                    (tag.data[frameStart + 6] <<  8) |
-                    (tag.data[frameStart + 7]);
+        frameSize = parseSyncSafeInteger(tag.data.subarray(frameStart + 4, frameStart + 8));
         if (frameSize < 1) {
           return console.log('Malformed ID3 frame encountered. Skipping metadata parsing.');
         }