Parse initialization data from AV1 tracks

Initialization data is not passed to `MediaCodecRenderer` as it is not required for playing AV1 video.

See: https://developer.android.com/reference/android/media/MediaCodec#CSD
PiperOrigin-RevId: 629729301

libraries/exoplayer/src/main/java/androidx/media3/exoplayer/mediacodec/MediaCodecRenderer.java

@@ -1544,6 +1544,15 @@ public abstract class MediaCodecRenderer extends BaseRenderer {
           newFormat,
           PlaybackException.ERROR_CODE_DECODING_FORMAT_UNSUPPORTED);
     }
+
+    // Remove the initialization data from the format if present when dealing with AV1, as it is not
+    // required for playing AV1 video.
+    // Reference: https://developer.android.com/reference/android/media/MediaCodec#CSD
+    if (Objects.equals(newFormat.sampleMimeType, MimeTypes.VIDEO_AV1)
+        && !newFormat.initializationData.isEmpty()) {
+      newFormat = newFormat.buildUpon().setInitializationData(null).build();
+    }
+
     setSourceDrmSession(formatHolder.drmSession);
     inputFormat = newFormat;
 

libraries/extractor/src/main/java/androidx/media3/extractor/mp4/AtomParsers.java

@@ -1227,6 +1227,12 @@ import java.util.Objects;
             ColorInfo.isoTransferCharacteristicsToColorTransfer(transferCharacteristics);
       } else if (childAtomType == Atom.TYPE_av1C) {
         mimeType = MimeTypes.VIDEO_AV1;
+
+        int childAtomBodySize = childAtomSize - Atom.HEADER_SIZE;
+        byte[] initializationDataChunk = new byte[childAtomBodySize];
+        parent.readBytes(initializationDataChunk, /* offset= */ 0, childAtomBodySize);
+        initializationData = ImmutableList.of(initializationDataChunk);
+
         parent.setPosition(childStartPosition + Atom.HEADER_SIZE);
         ColorInfo colorInfo = parseAv1c(parent);
 

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_av1c.mp4.0.dump

@@ -23,6 +23,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf60.3.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 3DFDDB0E
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_av1c.mp4.1.dump

@@ -23,6 +23,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf60.3.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 3DFDDB0E
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_av1c.mp4.2.dump

@@ -23,6 +23,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf60.3.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 3DFDDB0E
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_av1c.mp4.3.dump

@@ -23,6 +23,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf60.3.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 3DFDDB0E
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_av1c.mp4.unknown_length.dump

@@ -23,6 +23,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf60.3.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 3DFDDB0E
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_colr_mdcv_and_clli.mp4.0.dump

@@ -24,6 +24,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf58.76.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 4DF5B288
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_colr_mdcv_and_clli.mp4.1.dump

@@ -24,6 +24,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf58.76.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 4DF5B288
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_colr_mdcv_and_clli.mp4.2.dump

@@ -24,6 +24,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf58.76.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 4DF5B288
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_colr_mdcv_and_clli.mp4.3.dump

@@ -24,6 +24,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf58.76.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 4DF5B288
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/extractordumps/mp4/sample_with_colr_mdcv_and_clli.mp4.unknown_length.dump

@@ -24,6 +24,8 @@ track 0:
       lumaBitdepth = 10
       chromaBitdepth = 10
     metadata = entries=[TSSE: description=null: values=[Lavf58.76.100], Mp4Timestamp: creation time=0, modification time=0, timescale=1000]
+    initializationData:
+      data = length 20, hash 4DF5B288
   sample 0:
     time = 0
     flags = 1

libraries/test_data/src/test/assets/muxerdumps/sample_av1.mp4.dump

@@ -20,6 +20,8 @@ track 0:
       lumaBitdepth = 8
       chromaBitdepth = 8
     metadata = entries=[Mp4Timestamp: creation time=100000000, modification time=500000000, timescale=10000]
+    initializationData:
+      data = length 17, hash 54AC4E6D
   sample 0:
     time = 0
     flags = 1