Extract AudioRecordReader

Move the logic to read from an AudioRecord and handle all corner cases
for PTS. This simplifies AudioCapture.

PR #5102 <https://github.com/Genymobile/scrcpy/pull/5102>
This commit is contained in:
Romain Vimont 2024-07-14 23:28:13 +02:00
parent 414ce4c754
commit 053bf83f58
2 changed files with 71 additions and 47 deletions

View File

@ -10,7 +10,6 @@ import android.annotation.TargetApi;
import android.content.ComponentName;
import android.content.Intent;
import android.media.AudioRecord;
import android.media.AudioTimestamp;
import android.media.MediaCodec;
import android.os.Build;
import android.os.SystemClock;
@ -24,20 +23,11 @@ public final class AudioCapture {
private static final int CHANNELS = AudioConfig.CHANNELS;
private static final int CHANNEL_MASK = AudioConfig.CHANNEL_MASK;
private static final int ENCODING = AudioConfig.ENCODING;
private static final int BYTES_PER_SAMPLE = AudioConfig.BYTES_PER_SAMPLE;
private static final int MAX_READ_SIZE = AudioConfig.MAX_READ_SIZE;
private static final long ONE_SAMPLE_US = (1000000 + SAMPLE_RATE - 1) / SAMPLE_RATE; // 1 sample in microseconds (used for fixing PTS)
private final int audioSource;
private AudioRecord recorder;
private final AudioTimestamp timestamp = new AudioTimestamp();
private long previousRecorderTimestamp = -1;
private long previousPts = 0;
private long nextPts = 0;
private AudioRecordReader reader;
public AudioCapture(AudioSource audioSource) {
this.audioSource = audioSource.value();
@ -107,6 +97,7 @@ public final class AudioCapture {
recorder = Workarounds.createAudioRecord(audioSource, SAMPLE_RATE, CHANNEL_CONFIG, CHANNELS, CHANNEL_MASK, ENCODING);
}
recorder.startRecording();
reader = new AudioRecordReader(recorder);
}
public void checkCompatibility() throws AudioCaptureException {
@ -137,41 +128,7 @@ public final class AudioCapture {
}
@TargetApi(Build.VERSION_CODES.N)
public int read(ByteBuffer directBuffer, MediaCodec.BufferInfo outBufferInfo) {
int r = recorder.read(directBuffer, MAX_READ_SIZE);
if (r <= 0) {
return r;
}
long pts;
int ret = recorder.getTimestamp(timestamp, AudioTimestamp.TIMEBASE_MONOTONIC);
if (ret == AudioRecord.SUCCESS && timestamp.nanoTime != previousRecorderTimestamp) {
pts = timestamp.nanoTime / 1000;
previousRecorderTimestamp = timestamp.nanoTime;
} else {
if (nextPts == 0) {
Ln.w("Could not get initial audio timestamp");
nextPts = System.nanoTime() / 1000;
}
// compute from previous timestamp and packet size
pts = nextPts;
}
long durationUs = r * 1000000L / (CHANNELS * BYTES_PER_SAMPLE * SAMPLE_RATE);
nextPts = pts + durationUs;
if (previousPts != 0 && pts < previousPts + ONE_SAMPLE_US) {
// Audio PTS may come from two sources:
// - recorder.getTimestamp() if the call works;
// - an estimation from the previous PTS and the packet size as a fallback.
//
// Therefore, the property that PTS are monotonically increasing is no guaranteed in corner cases, so enforce it.
pts = previousPts + ONE_SAMPLE_US;
}
previousPts = pts;
outBufferInfo.set(0, r, pts, 0);
return r;
public int read(ByteBuffer outDirectBuffer, MediaCodec.BufferInfo outBufferInfo) {
return reader.read(outDirectBuffer, outBufferInfo);
}
}

View File

@ -0,0 +1,67 @@
package com.genymobile.scrcpy.audio;
import com.genymobile.scrcpy.util.Ln;
import android.annotation.TargetApi;
import android.media.AudioRecord;
import android.media.AudioTimestamp;
import android.media.MediaCodec;
import android.os.Build;
import java.nio.ByteBuffer;
public class AudioRecordReader {
private static final long ONE_SAMPLE_US =
(1000000 + AudioConfig.SAMPLE_RATE - 1) / AudioConfig.SAMPLE_RATE; // 1 sample in microseconds (used for fixing PTS)
private final AudioRecord recorder;
private final AudioTimestamp timestamp = new AudioTimestamp();
private long previousRecorderTimestamp = -1;
private long previousPts = 0;
private long nextPts = 0;
public AudioRecordReader(AudioRecord recorder) {
this.recorder = recorder;
}
@TargetApi(Build.VERSION_CODES.N)
public int read(ByteBuffer outDirectBuffer, MediaCodec.BufferInfo outBufferInfo) {
int r = recorder.read(outDirectBuffer, AudioConfig.MAX_READ_SIZE);
if (r <= 0) {
return r;
}
long pts;
int ret = recorder.getTimestamp(timestamp, AudioTimestamp.TIMEBASE_MONOTONIC);
if (ret == AudioRecord.SUCCESS && timestamp.nanoTime != previousRecorderTimestamp) {
pts = timestamp.nanoTime / 1000;
previousRecorderTimestamp = timestamp.nanoTime;
} else {
if (nextPts == 0) {
Ln.w("Could not get initial audio timestamp");
nextPts = System.nanoTime() / 1000;
}
// compute from previous timestamp and packet size
pts = nextPts;
}
long durationUs = r * 1000000L / (AudioConfig.CHANNELS * AudioConfig.BYTES_PER_SAMPLE * AudioConfig.SAMPLE_RATE);
nextPts = pts + durationUs;
if (previousPts != 0 && pts < previousPts + ONE_SAMPLE_US) {
// Audio PTS may come from two sources:
// - recorder.getTimestamp() if the call works;
// - an estimation from the previous PTS and the packet size as a fallback.
//
// Therefore, the property that PTS are monotonically increasing is no guaranteed in corner cases, so enforce it.
pts = previousPts + ONE_SAMPLE_US;
}
previousPts = pts;
outBufferInfo.set(0, r, pts, 0);
return r;
}
}