Skip to content

SDK / exoplayer

Go to a project
Commit 48fc9635 by Oliver Woodman
Options

webm_extractor: Add support for Cues present at the end of file 

This CL adds support for WebM files which have Cues present at the end of the file (i.e.) after Clusters. The file referenced in the bug can now be played back using the demo app. It adds a new flag to WebmExtractor which should be set to true only when being used through ExtractorSampleSource. All others (e.g. DASH) should not set it.

Reference file: http://demos.webmproject.org/exoplayer/glass_vp9_vorbis_cues_at_end.webm
parent 6cccfc1c
Inline Side-by-side
Showing with 96 additions and 13 deletions
library/src/main/java/com/google/android/exoplayer/extractor/webm/WebmExtractor.java
......@@ -32,6 +32,7 @@ import com.google.android.exoplayer.util.ParsableByteArray;
import java.io.IOException;
import java.nio.ByteBuffer;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.concurrent.TimeUnit;
/**
......@@ -49,6 +50,10 @@ public final class WebmExtractor implements Extractor {
private static final int SAMPLE_STATE_HEADER = 1;
private static final int SAMPLE_STATE_DATA = 2;
private static final int CUES_STATE_NOT_BUILT = 0;
private static final int CUES_STATE_BUILDING = 1;
private static final int CUES_STATE_BUILT = 2;
private static final String DOC_TYPE_WEBM = "webm";
private static final String CODEC_ID_VP9 = "V_VP9";
private static final String CODEC_ID_VORBIS = "A_VORBIS";
......@@ -65,6 +70,10 @@ public final class WebmExtractor implements Extractor {
private static final int ID_DOC_TYPE = 0x4282;
private static final int ID_DOC_TYPE_READ_VERSION = 0x4285;
private static final int ID_SEGMENT = 0x18538067;
private static final int ID_SEEK_HEAD = 0x114D9B74;
private static final int ID_SEEK = 0x4DBB;
private static final int ID_SEEK_ID = 0x53AB;
private static final int ID_SEEK_POSITION = 0x53AC;
private static final int ID_INFO = 0x1549A966;
private static final int ID_TIMECODE_SCALE = 0x2AD7B1;
private static final int ID_DURATION = 0x4489;
......@@ -108,7 +117,8 @@ public final class WebmExtractor implements Extractor {
private final EbmlReader reader;
private final VarintReader varintReader;
private final ParsableByteArray sampleHeaderScratch;
private ParsableByteArray vorbisNumPageSamples;
private final ParsableByteArray vorbisNumPageSamples;
private final ParsableByteArray seekEntryIdBytes;
private long segmentContentPosition = UNKNOWN;
private long segmentContentSize = UNKNOWN;
......@@ -121,7 +131,14 @@ public final class WebmExtractor implements Extractor {
private boolean sentDrmInitData;
// Cue related elements.
// Master seek entry related elements.
private int seekEntryId;
private long seekEntryPosition;
private boolean seekForCues;
private long cuesContentPosition = UNKNOWN;
private long seekPositionAfterBuildingCues = UNKNOWN;
private int cuesState = CUES_STATE_NOT_BUILT;
private long clusterTimecodeUs = UNKNOWN;
private LongArray cueTimesUs;
private LongArray cueClusterPositions;
......@@ -148,6 +165,8 @@ public final class WebmExtractor implements Extractor {
this.reader.init(new InnerEbmlReaderOutput());
varintReader = new VarintReader();
sampleHeaderScratch = new ParsableByteArray(4);
vorbisNumPageSamples = new ParsableByteArray(ByteBuffer.allocate(4).putInt(-1).array());
seekEntryIdBytes = new ParsableByteArray(4);
}
@Override
......@@ -167,17 +186,22 @@ public final class WebmExtractor implements Extractor {
public int read(ExtractorInput input, PositionHolder seekPosition) throws IOException,
InterruptedException {
sampleRead = false;
boolean inputHasData = true;
while (!sampleRead && inputHasData) {
inputHasData = reader.read(input);
boolean continueReading = true;
while (continueReading && !sampleRead) {
continueReading = reader.read(input);
if (continueReading && maybeSeekForCues(seekPosition, input.getPosition())) {
return Extractor.RESULT_SEEK;
}
}
return inputHasData ? Extractor.RESULT_CONTINUE : Extractor.RESULT_END_OF_INPUT;
return continueReading ? Extractor.RESULT_CONTINUE : Extractor.RESULT_END_OF_INPUT;
}
/* package */ int getElementType(int id) {
switch (id) {
case ID_EBML:
case ID_SEGMENT:
case ID_SEEK_HEAD:
case ID_SEEK:
case ID_INFO:
case ID_CLUSTER:
case ID_TRACKS:
......@@ -195,6 +219,7 @@ public final class WebmExtractor implements Extractor {
return EbmlReader.TYPE_MASTER;
case ID_EBML_READ_VERSION:
case ID_DOC_TYPE_READ_VERSION:
case ID_SEEK_POSITION:
case ID_TIMECODE_SCALE:
case ID_TIME_CODE:
case ID_PIXEL_WIDTH:
......@@ -215,6 +240,7 @@ public final class WebmExtractor implements Extractor {
case ID_DOC_TYPE:
case ID_CODEC_ID:
return EbmlReader.TYPE_STRING;
case ID_SEEK_ID:
case ID_CONTENT_ENCRYPTION_KEY_ID:
case ID_SIMPLE_BLOCK:
case ID_BLOCK:
......@@ -238,6 +264,10 @@ public final class WebmExtractor implements Extractor {
segmentContentPosition = contentPosition;
segmentContentSize = contentSize;
return;
case ID_SEEK:
seekEntryId = UNKNOWN;
seekEntryPosition = UNKNOWN;
return;
case ID_CUES:
cueTimesUs = new LongArray();
cueClusterPositions = new LongArray();
......@@ -245,6 +275,13 @@ public final class WebmExtractor implements Extractor {
case ID_CUE_POINT:
seenClusterPositionForCurrentCuePoint = false;
return;
case ID_CLUSTER:
// If we encounter a Cluster before building Cues, then we should try to build cues first
// before parsing the Cluster.
if (cuesState == CUES_STATE_NOT_BUILT && cuesContentPosition != UNKNOWN) {
seekForCues = true;
}
return;
case ID_CONTENT_ENCODING:
// TODO: check and fail if more than one content encoding is present.
return;
......@@ -261,8 +298,21 @@ public final class WebmExtractor implements Extractor {
/* package */ void endMasterElement(int id) throws ParserException {
switch (id) {
case ID_SEEK:
if (seekEntryId == UNKNOWN || seekEntryPosition == UNKNOWN) {
throw new ParserException("Mandatory element SeekID or SeekPosition not found");
}
if (seekEntryId == ID_CUES) {
cuesContentPosition = seekEntryPosition;
}
return;
case ID_CUES:
extractorOutput.seekMap(buildCues());
if (cuesState != CUES_STATE_BUILT) {
extractorOutput.seekMap(buildCues());
cuesState = CUES_STATE_BUILT;
} else {
// We have already built the cues. Ignore.
}
return;
case ID_CONTENT_ENCODING:
if (!trackFormat.hasContentEncryption) {
......@@ -326,6 +376,11 @@ public final class WebmExtractor implements Extractor {
throw new ParserException("DocTypeReadVersion " + value + " not supported");
}
return;
case ID_SEEK_POSITION:
// Seek Position is the relative offset beginning from the Segment. So to get absolute
// offset from the beginning of the file, we need to add segmentContentPosition to it.
seekEntryPosition = value + segmentContentPosition;
return;
case ID_TIMECODE_SCALE:
timecodeScale = value;
return;
......@@ -433,6 +488,12 @@ public final class WebmExtractor implements Extractor {
/* package */ void binaryElement(int id, int contentSize, ExtractorInput input)
throws IOException, InterruptedException {
switch (id) {
case ID_SEEK_ID:
Arrays.fill(seekEntryIdBytes.data, (byte) 0);
input.readFully(seekEntryIdBytes.data, 4 - contentSize, contentSize);
seekEntryIdBytes.setPosition(0);
seekEntryId = (int) seekEntryIdBytes.readUnsignedInt();
return;
case ID_CODEC_PRIVATE:
trackFormat.codecPrivate = new byte[contentSize];
input.readFully(trackFormat.codecPrivate, 0, contentSize);
......@@ -535,12 +596,7 @@ public final class WebmExtractor implements Extractor {
// we set it to -1). The android platform media extractor [2] does the same.
// [1] https://android.googlesource.com/platform/frameworks/av/+/lollipop-release/media/libstagefright/codecs/vorbis/dec/SoftVorbis.cpp#314
// [2] https://android.googlesource.com/platform/frameworks/av/+/lollipop-release/media/libstagefright/NuMediaExtractor.cpp#474
if (vorbisNumPageSamples == null) {
vorbisNumPageSamples =
new ParsableByteArray(ByteBuffer.allocate(4).putInt(-1).array());
} else {
vorbisNumPageSamples.setPosition(0);
}
vorbisNumPageSamples.setPosition(0);
trackOutput.sampleData(vorbisNumPageSamples, 4);
sampleSize += 4;
}
......@@ -682,6 +738,33 @@ public final class WebmExtractor implements Extractor {
}
}
/**
* Updates the position of the holder to Cues element's position if the extractor configuration
* permits use of master seek entry. After building Cues sets the holder's position back to where
* it was before.
*
* @param seekPosition The holder whose position will be updated.
* @param currentPosition Current position of the input.
* @return true if the seek position was updated, false otherwise.
*/
private boolean maybeSeekForCues(PositionHolder seekPosition, long currentPosition) {
if (seekForCues) {
seekPositionAfterBuildingCues = currentPosition;
seekPosition.position = cuesContentPosition;
cuesState = CUES_STATE_BUILDING;
seekForCues = false;
return true;
}
// After parsing Cues, Seek back to original position if available. We will not do this unless
// we seeked to get to the Cues in the first place.
if (cuesState == CUES_STATE_BUILT && seekPositionAfterBuildingCues != UNKNOWN) {
seekPosition.position = seekPositionAfterBuildingCues;
seekPositionAfterBuildingCues = UNKNOWN;
return true;
}
return false;
}
private long scaleTimecodeToUs(long unscaledTimecode) {
return TimeUnit.NANOSECONDS.toMicros(unscaledTimecode * timecodeScale);
}
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or sign in to comment