Android 11视频音频样本提取播放异常问题求助

API 探路者

2026-7-16

Android 11视频音频提取播放异常：仅播放短时长噪音

在Android 11环境下，从15秒视频文件提取音频样本并播放时，仅能听到短时长随机噪音，getAudio()方法仅获取到487,831字节数据，以下是问题详情及解决方案：

问题代码

// 获取音频样本的方式
public ArrayList<byte[]> getAudio() {
    int index = 0;
    ArrayList<byte[]> audio = new ArrayList<>();
    ByteBuffer buffer = ByteBuffer.allocateDirect(format.getInteger(MediaFormat.KEY_MAX_INPUT_SIZE));

    for (;;) {
        int size = extractor.readSampleData(buffer, 0);
        total_size += size;
        if (size < 0) { // 无更多数据可读
            break;
        }

        for (int i = 0; i < size / 2; i++) {
            byte[] sample = new byte[2];
            for (int j = 0; j < sample.length; j++) {
                sample[j] = buffer.get();
            }
            audio.add(sample);
        }
        extractor.advance();
    }
    return audio;
}

Log.d("AudioTest", "total audio data size in bytes: " + total_size);

// 播放音频的方式
int SIZE = AudioTrack.getMinBufferSize(48000, AudioFormat.CHANNEL_OUT_MONO, AudioFormat.ENCODING_PCM_16BIT);
AudioTrack track = new AudioTrack(AudioManager.STREAM_MUSIC, 48000, AudioFormat.CHANNEL_OUT_MONO, AudioFormat.ENCODING_PCM_16BIT, SIZE, AudioTrack.MODE_STREAM);
ArrayList<byte[]> audio = getAudio();
long start = System.currentTimeMillis();

track.play();
byte[] data = new byte[audio.size() * 2];
for (int i = 0, k = 0; i < audio.size(); i++) {
    byte[] sample = audio.get(i);
    for (int j = 0; j < sample.length; j++, k++) {
        data[k] = sample[j];
    }
}
track.write(data, 0, data.length);

long duration = System.currentTimeMillis() - start;
Log.d("AudioTest", "that took: " + (duration / 1000) + "s");

视频文件信息

帧信息（视频轨道）

has-sdtp: 1
track-id: 1
level: 2048
mime: video/avc
frame-count: 461
profile: 8
language: und
color-standard: 1
display-width: 1920
track-fourcc: 828601953
csd-1: java.nio.HeapByteBuffer[pos=0 lim=9 cap=9]
color-transfer: 3
durationUs: 15,382,033
display-height: 1080
width: 1920
color-range: 2
max-input-size: 3,145,748
frame-rate: 30
height: 1080
csd-0: java.nio.HeapByteBuffer[pos=0 lim=33 cap=33]

样本信息（音频轨道）

max-bitrate: 262,664
isDMCMMExtractor: 1
sample-rate: 48,000
track-id: 2
mime: audio/mp4a-latm
profile: 2
language: und
aac-profile: 2
track-fourcc: -1
encoder-delay: 0
durationUs: 15,402,666
channel-count: 2
bits-per-sample: 16
encoder-padding: 48
max-input-size: 524,308
csd-0: java.nio.HeapByteBuffer[pos=0 lim=2 cap=2]

核心问题分析

未选择正确的音频轨道：从文件信息看，track-id=2是音频轨道，但代码中未明确让MediaExtractor切换到该轨道，可能默认读取了视频轨道的数据，导致获取的音频数据量不足。
直接播放编码后的AAC数据：音频轨道的mime类型是audio/mp4a-latm（即AAC编码），MediaExtractor读取的是编码后的压缩数据，并非可直接播放的PCM原始音频，直接用AudioTrack播放必然产生噪音。
音频参数不匹配：原音频是双声道（channel-count=2），但代码中AudioTrack使用了单声道（CHANNEL_OUT_MONO），即使解码正确也会导致播放异常。
音频数据处理逻辑错误：getAudio()方法中把编码数据按2字节分割成“样本”的操作完全错误，AAC编码数据的结构与PCM样本无关。

解决方案：解码AAC为PCM后播放

需要使用MediaCodec对AAC数据进行解码，得到PCM原始音频后再用AudioTrack播放，以下是修正后的示例代码：

private void extractAndPlayAudio(String videoPath) {
    MediaExtractor extractor = new MediaExtractor();
    MediaCodec codec = null;
    AudioTrack audioTrack = null;
    try {
        extractor.setDataSource(videoPath);
        // 找到音频轨道
        int audioTrackIndex = -1;
        MediaFormat audioFormat = null;
        for (int i = 0; i < extractor.getTrackCount(); i++) {
            MediaFormat format = extractor.getTrackFormat(i);
            String mime = format.getString(MediaFormat.KEY_MIME);
            if (mime.startsWith("audio/")) {
                audioTrackIndex = i;
                audioFormat = format;
                break;
            }
        }
        if (audioTrackIndex == -1) {
            Log.e("AudioTest", "未找到音频轨道");
            return;
        }
        extractor.selectTrack(audioTrackIndex);

        // 初始化MediaCodec解码AAC
        String mime = audioFormat.getString(MediaFormat.KEY_MIME);
        codec = MediaCodec.createDecoderByType(mime);
        codec.configure(audioFormat, null, null, 0);
        codec.start();

        // 初始化AudioTrack，匹配原音频参数
        int sampleRate = audioFormat.getInteger(MediaFormat.KEY_SAMPLE_RATE);
        int channelCount = audioFormat.getInteger(MediaFormat.KEY_CHANNEL_COUNT);
        int audioFormatType = AudioFormat.ENCODING_PCM_16BIT;
        int channelConfig = channelCount == 2 ? AudioFormat.CHANNEL_OUT_STEREO : AudioFormat.CHANNEL_OUT_MONO;
        int minBufferSize = AudioTrack.getMinBufferSize(sampleRate, channelConfig, audioFormatType);
        audioTrack = new AudioTrack(AudioManager.STREAM_MUSIC, sampleRate, channelConfig, audioFormatType, minBufferSize * 2, AudioTrack.MODE_STREAM);
        audioTrack.play();

        // 解码并播放流程
        ByteBuffer[] inputBuffers = codec.getInputBuffers();
        ByteBuffer[] outputBuffers = codec.getOutputBuffers();
        MediaCodec.BufferInfo bufferInfo = new MediaCodec.BufferInfo();
        boolean isDone = false;

        while (!isDone) {
            // 填充输入数据到MediaCodec
            int inputBufferIndex = codec.dequeueInputBuffer(10000);
            if (inputBufferIndex >= 0) {
                ByteBuffer inputBuffer = inputBuffers[inputBufferIndex];
                int sampleSize = extractor.readSampleData(inputBuffer, 0);
                if (sampleSize < 0) {
                    // 输入数据结束，发送结束标记
                    codec.queueInputBuffer(inputBufferIndex, 0, 0, 0, MediaCodec.BUFFER_FLAG_END_OF_STREAM);
                    isDone = true;
                } else {
                    long presentationTimeUs = extractor.getSampleTime();
                    codec.queueInputBuffer(inputBufferIndex, 0, sampleSize, presentationTimeUs, 0);
                    extractor.advance();
                }
            }

            // 获取解码后的PCM数据
            int outputBufferIndex = codec.dequeueOutputBuffer(bufferInfo, 10000);
            switch (outputBufferIndex) {
                case MediaCodec.INFO_OUTPUT_BUFFERS_CHANGED:
                    outputBuffers = codec.getOutputBuffers();
                    break;
                case MediaCodec.INFO_OUTPUT_FORMAT_CHANGED:
                    // 输出格式变化，这里可以重新确认参数
                    break;
                case MediaCodec.INFO_TRY_AGAIN_LATER:
                    break;
                default:
                    ByteBuffer outputBuffer = outputBuffers[outputBufferIndex];
                    // 把PCM数据写入AudioTrack播放
                    byte[] pcmData = new byte[bufferInfo.size];
                    outputBuffer.get(pcmData);
                    outputBuffer.clear();
                    audioTrack.write(pcmData, 0, pcmData.length);
                    codec.releaseOutputBuffer(outputBufferIndex, false);
                    if ((bufferInfo.flags & MediaCodec.BUFFER_FLAG_END_OF_STREAM) != 0) {
                        isDone = true;
                    }
                    break;
            }
        }
    } catch (IOException e) {
        e.printStackTrace();
    } finally {
        if (audioTrack != null) {
            audioTrack.stop();
            audioTrack.release();
        }
        if (codec != null) {
            codec.stop();
            codec.release();
        }
        extractor.release();
    }
}

关键修正点

明确选择音频轨道，避免读取视频数据
使用MediaCodec解码AAC压缩数据为PCM原始音频
确保AudioTrack的参数（采样率、声道数、编码格式）与原音频完全匹配
采用流式播放逻辑，边解码边写入AudioTrack，避免一次性加载大量数据导致的异常

内容的提问来源于stack exchange，提问作者EinStyn

Android 11视频音频样本提取播放异常问题求助

问题代码 #

视频文件信息 #

帧信息（视频轨道） #

样本信息（音频轨道） #

核心问题分析 #

解决方案：解码AAC为PCM后播放 #

关键修正点 #