vtt clean content (#132)

Azure · Aug 15, 2023 · 2a52b87 · 2a52b87
1 parent 03e6e0d
commit 2a52b87
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 20 deletions.
diff --git a/pipes/MultiFileStream.cs b/pipes/MultiFileStream.cs
@@ -183,8 +183,7 @@ private void GenerateVttContent(IList<Box> inputBoxes, MP4Writer mp4Writer)
 
                 if (vttText != null)
                 {
-                    byte[] contentBytes = Encoding.UTF8.GetBytes(RemoveXmlControlCharacters(vttText));
-                    mp4Writer.Write(contentBytes);
+                    mp4Writer.Write(vttText);
                 }
                 else
                 {
@@ -197,22 +196,6 @@ private void GenerateVttContent(IList<Box> inputBoxes, MP4Writer mp4Writer)
             }
         }
 
-        private string RemoveXmlControlCharacters(string input)
-        {
-            StringBuilder output = new StringBuilder();
-
-            foreach (char c in input)
-            {
-                // Exclude XML control characters (0x00 to 0x1F, except for whitespace characters)
-                if (c >= 0x20 || char.IsWhiteSpace(c))
-                {
-                    output.Append(c);
-                }
-            }
-
-            return output.ToString();
-        }
-
         private async Task DownloadClearBlobContent(BlockBlobClient sourceBlob, Stream outputStream, CancellationToken cancellationToken)
         {
             using var tmpStream = new MemoryStream();            

diff --git a/transform/CTtml2WebVttConv.cs b/transform/CTtml2WebVttConv.cs
@@ -7,7 +7,7 @@
 
 public static class TtmlToVttConverter
 {
-    public static string? Convert(byte[]? ttmlText)
+    public static byte[]? Convert(byte[]? ttmlText)
     {
         if (ttmlText == null)
         {
@@ -127,7 +127,12 @@ public static class TtmlToVttConverter
 
              webVttContentRes= webVttContent.ToString();
         }
-        return !string.IsNullOrEmpty(webVttContentRes) ? webVttContentRes: null; //webVttContentRes;Regex.Replace(webVttContentRes, @"[\x00-\x1F\x7F]", "\n")
+        byte[]? contentBytes = null;
+        if (!string.IsNullOrEmpty(webVttContentRes))
+        {
+            contentBytes = Encoding.UTF8.GetBytes(webVttContentRes);
+        }
+        return contentBytes; 
     }
 
     private static void ParseRegionAttributes(XmlReader pReader, out string strStartCue, out string strCueSize, out string strAlign)