将 AAC 音频文件解码为 RAW PCM 文件

Question

我有一个 AAC 格式的音频文件，我正在尝试将其转换为原始格式的 PCM 文件，以便将其与另一个音频文件混合并稍后使用 AudioTrack 播放。

经过一番研究，我发现 this library 可以正确解码我的 AAC 文件。但是，它只将解码后的字节直接传递给 AudioTrack。当尝试将解码后的字节写入输出流时，生成的文件只包含噪音。

这是我用来解码 AAC 文件的代码 -

public void AACDecoderAndPlay() {
    ByteBuffer[] inputBuffers = mDecoder.getInputBuffers();
    ByteBuffer[] outputBuffers = mDecoder.getOutputBuffers();

    BufferInfo info = new BufferInfo();

    // create an audiotrack object
    AudioTrack audioTrack = new AudioTrack(AudioManager.STREAM_MUSIC, JamboxAudioTrack.FREQUENCY,
            JamboxAudioTrack.CHANNEL_CONFIGURATION, JamboxAudioTrack.AUDIO_ENCODING,
            JamboxAudioTrack.BUFFER_SIZE, AudioTrack.MODE_STREAM);

    audioTrack.play();


    long bytesWritten = 0;
    while (!eosReceived) {
        int inIndex = mDecoder.dequeueInputBuffer(TIMEOUT_US);
        if (inIndex >= 0) {
            ByteBuffer buffer = inputBuffers[inIndex];
            int sampleSize = mExtractor.readSampleData(buffer, 0);
            if (sampleSize < 0) {
                // We shouldn't stop the playback at this point, just pass the EOS
                // flag to mDecoder, we will get it again from the
                // dequeueOutputBuffer
                Log.d(LOG_TAG, "InputBuffer BUFFER_FLAG_END_OF_STREAM");
                mDecoder.queueInputBuffer(inIndex, 0, 0, 0, MediaCodec.BUFFER_FLAG_END_OF_STREAM);

            } else {
                mDecoder.queueInputBuffer(inIndex, 0, sampleSize, mExtractor.getSampleTime(), 0);
                mExtractor.advance();
            }

            int outIndex = mDecoder.dequeueOutputBuffer(info, TIMEOUT_US);
            switch (outIndex) {
                case MediaCodec.INFO_OUTPUT_BUFFERS_CHANGED:
                    Log.d(LOG_TAG, "INFO_OUTPUT_BUFFERS_CHANGED");
                    outputBuffers = mDecoder.getOutputBuffers();
                    break;

                case MediaCodec.INFO_OUTPUT_FORMAT_CHANGED:
                    MediaFormat format = mDecoder.getOutputFormat();
                        Log.d(LOG_TAG, "New format " + format);
//                        audioTrack.setPlaybackRate(format.getInteger(MediaFormat.KEY_SAMPLE_RATE));

                        break;

                case MediaCodec.INFO_TRY_AGAIN_LATER:
                    Log.d(LOG_TAG, "dequeueOutputBuffer timed out!");
                    break;

                default:
                    ByteBuffer outBuffer = outputBuffers[outIndex];
                    Log.v(LOG_TAG, "We can't use this buffer but render it due to the API limit, " + outBuffer);

                    final byte[] chunk = new byte[info.size];
                    outBuffer.get(chunk); // Read the buffer all at once
                    outBuffer.clear(); // ** MUST DO!!! OTHERWISE THE NEXT TIME YOU GET THIS SAME BUFFER BAD THINGS WILL HAPPEN

                    audioTrack.write(chunk, info.offset, info.offset + info.size); // AudioTrack write data

                    if (info.offset > 0) {
                        Log.v(LOG_TAG, "" + info.offset);
                    }
                    try {
                        mOutputStream.write(chunk, info.offset, info.offset + info.size);
                        bytesWritten += info.offset + info.size;
                    } catch (IOException e) {
                        e.printStackTrace();
                    }
                    mDecoder.releaseOutputBuffer(outIndex, false);
                    break;
            }

            // All decoded frames have been rendered, we can stop playing now
            if ((info.flags & MediaCodec.BUFFER_FLAG_END_OF_STREAM) != 0) {
                Log.d(LOG_TAG, "OutputBuffer BUFFER_FLAG_END_OF_STREAM");
                break;
            }
        }
    }
    Log.v(LOG_TAG, "Bytes written: " + bytesWritten);

    mDecoder.stop();
    mDecoder.release();
    mDecoder = null;

    mExtractor.release();
        mExtractor = null;

        audioTrack.stop();
        audioTrack.release();
        audioTrack = null;
    }

为了播放解码文件，我使用从缓冲区读取和播放的普通 AudioTrack -

public void start() {
    new Thread(new Runnable() {
        public void run() {
            try {
                Process.setThreadPriority(Process.THREAD_PRIORITY_URGENT_AUDIO);
                InputStream inputStream = new FileInputStream(playingFile);
                BufferedInputStream bufferedInputStream = new BufferedInputStream(inputStream);
                DataInputStream dataInputStream = new DataInputStream(bufferedInputStream);

                AudioTrack track = new AudioTrack(AudioManager.STREAM_MUSIC, FREQUENCY,
                        CHANNEL_CONFIGURATION, AUDIO_ENCODING, BUFFER_SIZE, AudioTrack.MODE_STREAM);

                short[] buffer = new short[BUFFER_SIZE / 4];
                long startTime = System.currentTimeMillis();

                track.play();

                while (dataInputStream.available() > 0) {
                    int i = 0;
                    while (dataInputStream.available() > 0 && i < buffer.length

) {
                            buffer[i] = dataInputStream.readShort();
                            i++;
                        }
                        track.write(buffer, 0, buffer.length);
                        if (latency < 0) {
                            latency = System.currentTimeMillis() - startTime;
                        }
                    }
//
//            int i = 0;
//            while((i = dataInputStream.read(buffer, 0, BUFFER_SIZE)) > -1){
//                track.write(buffer, 0, i);
//            }
                    track.stop();
                    track.release();
                    dataInputStream.close();
                    inputStream.close();
                }
                catch (Exception e)
                {
                    e.printStackTrace();
                }
            }
    }).start();
}

我错过了什么？

Answer 1

您的问题似乎在于您将输出写为纯字节（尽管我没有在您的代码中看到 mOutputStream 的设置）。这些纯字节将采用您平台的本机字节序（实际上是小字节序），但以平台无关的方式（指定为大字节序）使用 DataInputStream 将其读取为短裤。

这里最简单的解决方法就是播放时用byte数组代替short数组； AudioTrack 接受字节数组和短数组，当给定一个字节数组时，它以正确的（本机）方式解释它，这与 MediaCodec 的输出相匹配。只需确保缓冲区大小为偶数字节即可。

如果您确实需要将值设为 shorts，则需要使用以小端模式读取的 reader（当前所有 Android ABI 都是小端模式） .这似乎没有任何直接可用的 API，但实际上并不太难。参见例如Java : DataInputStream replacement for endianness 中的 readLittleShort 方法，以获取有关如何执行此操作的示例。

将 AAC 音频文件解码为 RAW PCM 文件

Decoding an AAC audio file into a RAW PCM file

android

aac

audiotrack

mediaextractor

android-mediacodec