fix: 708 captions multi-byte char fix (#439)

videojs · Oct 12, 2023 · ec31749 · ec31749
1 parent 21e55aa
commit ec31749
Show file tree

Hide file tree

Showing 2 changed files with 48 additions and 7 deletions.
diff --git a/lib/m2ts/caption-stream.js b/lib/m2ts/caption-stream.js
@@ -688,18 +688,32 @@ Cea708Stream.prototype.handleText = function(i, service, options) {
   var char;
   var charCodeArray;
 
+  // Converts an array of bytes to a unicode hex string.
+  function toHexString(byteArray) {
+    return byteArray.map((byte) => {
+      return ('0' + (byte & 0xFF).toString(16)).slice(-2);
+    }).join('');
+  };
+
+  if (isMultiByte) {
+    charCodeArray = [currentByte, nextByte];
+    i++;
+  } else {
+    charCodeArray = [currentByte];
+  }
+
   // Use the TextDecoder if one was created for this service
   if (service.textDecoder_ && !isExtended) {
+    char = service.textDecoder_.decode(new Uint8Array(charCodeArray));
+  } else {
+    // We assume any multi-byte char without a decoder is unicode.
     if (isMultiByte) {
-      charCodeArray = [currentByte, nextByte];
-      i++;
+      const unicode = toHexString(charCodeArray);
+      // Takes a unicode hex string and creates a single character.
+      char = String.fromCharCode(parseInt(unicode, 16));
     } else {
-     charCodeArray = [currentByte];
+      char = get708CharFromCode(extended | currentByte);
     }
-
-    char = service.textDecoder_.decode(new Uint8Array(charCodeArray));
-  } else {
-    char = get708CharFromCode(extended | currentByte);
   }
 
   if (win.pendingNewLine && !win.isEmpty()) {

diff --git a/test/caption-stream.test.js b/test/caption-stream.test.js
@@ -3051,6 +3051,33 @@ QUnit.test('Decodes multibyte characters if valid encoding option is provided an
   }
 });
 
+QUnit.test('Decodes multi-byte characters as unicode if no valid encoding option is provided', function(assert) {
+  var captions = [];
+
+  cea708Stream = new m2ts.Cea708Stream({
+    captionServices: {
+      SERVICE1: {}
+    }
+  });
+
+  cea708Stream.on('data', function(caption) {
+    captions.push(caption);
+  });
+
+  cc708Korean.forEach(cea708Stream.push, cea708Stream);
+
+  cea708Stream.flushDisplayed(4721138662, cea708Stream.services[1]);
+
+  assert.equal(captions.length, 1, 'parsed single caption correctly');
+
+  assert.notOk(cea708Stream.services[1].textDecoder_, 'TextDecoder was not created');
+  assert.equal(
+    captions[0].text,
+    '듏낡 ',
+    'parsed multibyte characters correctly'
+  );
+});
+
 QUnit.test('Creates TextDecoder only if valid encoding value is provided', function(assert) {
   var secondCea708Stream;