基于Camera、AudioRecord 、MediaCodec 和 MediaMuxer 录制 MP4

一.前言

在 AAC 音频编码保存和解码播放和Camera 视频采集，H264 编码保存
两篇文章中介绍了如何通过 AudioRecord 和 MediaCodec 录制 AAC 音频以及如何通过 Camera
和 MediaCodec 录制 H264 视频。本文将介绍如何通过 MediaMuxer 合成 MP4 文件。

MP4

在 音视频开发基础概念中有介绍过，MP4 (或者称 MPEG-4) 是一种标准的数字多媒体容器格式，可以存储
音频数据和视频数据。对于视频格式，常见的是 H264 和 H265; 对于音频格式通常是 AAC 。

MediaMuxer

MediaMuxer 是 Android 用来产生一个混合音频和视频多媒体文件的 API ,只支持下面几种格式。

public static final inMUXER_OUTPUT_3GPP = 2;
public static final inMUXER_OUTPUT_HEIF = 3;
public static final inMUXER_OUTPUT_MPEG_4 = 0;
public static final inMUXER_OUTPUT_OGG = 4;
public static final inMUXER_OUTPUT_WEBM = 1;

1. 初始化

mMediaMuxer = new MediaMuxer(path, MediaMuxer.OutputFormat.MUXER_OUTPUT_MPEG_4);

path 表示 MP4 文件的输出路径

2. 添加音频轨和视频轨

if (type == AAC_ENCODER) {
    mAudioTrackIndex = mMediaMuxer.addTrack(mediaFormat);
}
if (type == H264_ENCODER) {
    mVideoTrackIndex = mMediaMuxer.addTrack(mediaFormat);
}

传入 MediaFormat 对象从 MediaCodec 中获取。

3. 开始合成

mMediaMuxer.start();

4. 写入数据

mMediaMuxer.writeSampleData(avData.trackIndex, avData.byteBuffer, avData.bufferInfo);

5. 停止并释放资源

mMediaMuxer.stop();
mMediaMuxer.release();

二. 录制 MP4

AudioTrack、Camera、MediaCodec 和 MediaMuxer 录制 MP4 流程如下图所示：

1. 音频录制

音频录制使用 AudioRecord

public class AudioRecorder {

    private int mAudioSource;
    private int mSampleRateInHz;
    private int mChannelConfig;
    private int mAudioFormat;
    private int mBufferSizeInBytes;

    private AudioRecord mAudioRecord;
    private volatile boolean mIsRecording;
    private Callback mCallback;
    private byte[] mBuffer;

    public void setCallback(Callback callback) {
        mCallback = callback;
    }

    public interface Callback {
        void onAudioOutput(byte[] data);
    }

    public AudioRecorder(int audioSource, int sampleRateInHz, int channelConfig, int audioFormat, int bufferSizeInBytes) {
        mAudioSource = audioSource;
        mSampleRateInHz = sampleRateInHz;
        mChannelConfig = channelConfig;
        mAudioFormat = audioFormat;
        mBufferSizeInBytes = bufferSizeInBytes;

        mAudioRecord = new AudioRecord(audioSource, sampleRateInHz, channelConfig, audioFormat, bufferSizeInBytes);
        mIsRecording = false;
        int minBufferSize = AudioRecord.getMinBufferSize(sampleRateInHz, channelConfig, AudioFormat.ENCODING_PCM_16BIT);

        mBuffer = new byte[Math.min(2048, minBufferSize)];
    }

    public void start() {
        if (mIsRecording) {
            return;
        }
        new Thread(new Runnable() {
            @Override
            public void run() {
                onStart();
            }
        }).start();
    }

    public void onStart() {
        if (mAudioRecord == null) {
            mAudioRecord = new android.media.AudioRecord(mAudioSource, mSampleRateInHz, mChannelConfig, mAudioFormat, mBufferSizeInBytes);
        }
        mAudioRecord.startRecording();
        mIsRecording = true;
        while (mIsRecording) {
            int len = mAudioRecord.read(mBuffer, 0, mBuffer.length);
            if (len > 0) {
                if (mCallback != null) {
                    mCallback.onAudioOutput(mBuffer);
                }
            }
        }
        mAudioRecord.stop();
        mAudioRecord.release();
        mAudioRecord = null;
    }

    public void stop() {
        mIsRecording = false;
    }
}

2. 音频编码

音频编码使用阻塞队列 BlockingQueue 来缓冲数据，编码成 AAC 格式。

public class AacEncoder {
    public static final int AAC_ENCODER = 2;
    private MediaCodec mAudioEncoder;
    private MediaFormat mMediaFormat;
    private BlockingQueue<byte[]> mDataQueue;
    private volatile boolean mIsEncoding;
    private Callback mCallback;

    private static final String AUDIO_MIME_TYPE = "audio/mp4a-latm";//就是 aac

    public void setCallback(Callback callback) {
        mCallback = callback;
    }

    public interface Callback {
        void outputMediaFormat(int type, MediaFormat mediaFormat);

        void onEncodeOutput(ByteBuffer byteBuffer, MediaCodec.BufferInfo bufferInfo);

        void onStop(int type);
    }

    public AacEncoder(int sampleRateInHz, int channelConfig, int bufferSizeInBytes) {
        try {
            mAudioEncoder = MediaCodec.createEncoderByType(AUDIO_MIME_TYPE);
            mMediaFormat = MediaFormat.createAudioFormat(AUDIO_MIME_TYPE, sampleRateInHz, channelConfig == AudioFormat.CHANNEL_OUT_MONO ? 1 : 2);
            mMediaFormat.setInteger(MediaFormat.KEY_AAC_PROFILE, MediaCodecInfo.CodecProfileLevel.AACObjectLC);
            mMediaFormat.setInteger(MediaFormat.KEY_CHANNEL_MASK, AudioFormat.CHANNEL_IN_STEREO);//CHANNEL_IN_STEREO 立体声
            int bitRate = sampleRateInHz * 16 * channelConfig == AudioFormat.CHANNEL_IN_MONO ? 1 : 2;
            mMediaFormat.setInteger(MediaFormat.KEY_BIT_RATE, bitRate);
            mMediaFormat.setInteger(MediaFormat.KEY_CHANNEL_COUNT, channelConfig == AudioFormat.CHANNEL_IN_MONO ? 1 : 2);
            mMediaFormat.setInteger(MediaFormat.KEY_SAMPLE_RATE, sampleRateInHz);
            mAudioEncoder.configure(mMediaFormat, null, null, MediaCodec.CONFIGURE_FLAG_ENCODE);
        } catch (IOException e) {

        }
        mDataQueue = new ArrayBlockingQueue<>(10);
        mIsEncoding = false;
    }

    public void start() {
        if (mIsEncoding) {
            return;
        }
        new Thread(new Runnable() {
            @Override
            public void run() {
                onStart();
            }
        }).start();
    }

    public void stop() {
        mIsEncoding = false;
    }

    private void onStart() {
        mIsEncoding = true;
        mAudioEncoder.start();
        byte[] pcmData;
        int inputIndex;
        ByteBuffer inputBuffer;
        ByteBuffer[] inputBuffers = mAudioEncoder.getInputBuffers();

        int outputIndex;
        ByteBuffer outputBuffer;
        ByteBuffer[] outputBuffers = mAudioEncoder.getOutputBuffers();

        MediaCodec.BufferInfo bufferInfo =