前面有分析过Camera的实现，现在来看看MediaRecorder的实现，这里我不会太去关注它的分层结构，我更关注它的逻辑！

APP层/path/to/aosp/frameworks/base/media/java/android/media/MediaRecorder.java
JNI层/path/to/aosp/frameworks/base/media/jni/android_media_MediaRecorder.cpp
调用NATIVE层的MediaRecorder(这里是BnMediaRecorderClient)
header /path/to/aosp/frameworks/av/include/media/mediarecorder.h
implementation/path/to/aosp/frameworks/av/media/libmedia/mediarecorder.cpp

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
MediaRecorder::MediaRecorder() : mSurfaceMediaSource(NULL)
{
    ALOGV("constructor");
    constsp<IMediaPlayerService>& service(getMediaPlayerService());
    if(service != NULL) {
        mMediaRecorder = service->createMediaRecorder(getpid());
    }
    if(mMediaRecorder != NULL) {
        mCurrentState = MEDIA_RECORDER_IDLE;
    }
    doCleanUp();
}

getMediaPlayerService()这个方法位于/path/to/aosp/frameworks/av/include/media/IMediaDeathNotifier.h

获取到MediaPlayerService(这个是BpMediaPlayerService)之后
调用IMediaPlayerService当中的

?
1
2
3
4
5
6
7
8
9
sp<IMediaRecorder> MediaPlayerService::createMediaRecorder(pid_t pid)
{
    sp<MediaRecorderClient> recorder = newMediaRecorderClient(this, pid);
    wp<MediaRecorderClient> w = recorder;
    Mutex::Autolock lock(mLock);
    mMediaRecorderClients.add(w);
    ALOGV("Create new media recorder client from pid %d", pid);
    returnrecorder;
}

创建MediaRecorderClient(这里是BnMediaRecorder)

但是通过binder拿到的是BpMediaRecorder
因为有如下的interface_cast过程

?
1
2
3
4
5
6
7
8
virtual sp<IMediaRecorder> createMediaRecorder(pid_t pid)
{
    Parcel data, reply;
    data.writeInterfaceToken(IMediaPlayerService::getInterfaceDescriptor());
    data.writeInt32(pid);
    remote()->transact(CREATE_MEDIA_RECORDER, data, &reply);
    returninterface_cast<IMediaRecorder>(reply.readStrongBinder());
}

而MediaRecorderClient当中又会创建StagefrightRecorder(MediaRecorderBase)，它位于
/path/to/aosp/frameworks/av/media/libmediaplayerservice/StagefrightRecorder.cpp

目前我们可以认为在APP/JNI/NATIVE这边是在一个进程当中，在MediaPlayerService当中的MediaRecorderClient/StagefrightRecorder是在另外一个进程当中，他们之间通过binder通信，而且Bp和Bn我们也都有拿到，后面我们将不再仔细区分Bp和Bn。

客户端这边
BnMediaRecorderClient
BpMediaRecorder
BpMediaPlayerService

服务端这边
BpMediaRecorderClient(如果需要通知客户端的话，它可以获得这个Bp)
BnMediaRecorder
BnMediaPlayerService

这有张图(点过去看原始大图)
Android MediaRecorder Diagram

我们以开始录影为例子，比如start()

在这里就兵分两路，一个CameraSource，一个MPEG4Writer(sp mWriter)
这两个class都位于/path/to/aosp/frameworks/av/media/libstagefright/当中

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
status_t StagefrightRecorder::startMPEG4Recording() {
    int32_t totalBitRate;
    status_t err = setupMPEG4Recording(
            mOutputFd, mVideoWidth, mVideoHeight,
            mVideoBitRate, &totalBitRate, &mWriter);
    if(err != OK) {
        returnerr;
    }
    int64_t startTimeUs = systemTime() / 1000;
    sp<MetaData> meta = newMetaData;
    setupMPEG4MetaData(startTimeUs, totalBitRate, &meta);
    err = mWriter->start(meta.get());
    if(err != OK) {
        returnerr;
    }
    returnOK;
}

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
status_t StagefrightRecorder::setupMPEG4Recording(
        intoutputFd,
        int32_t videoWidth, int32_t videoHeight,
        int32_t videoBitRate,
        int32_t *totalBitRate,
        sp<MediaWriter> *mediaWriter) {
    mediaWriter->clear();
    *totalBitRate = 0;
    status_t err = OK;
    sp<MediaWriter> writer = newMPEG4Writer(outputFd);
    if(mVideoSource < VIDEO_SOURCE_LIST_END) {
        sp<MediaSource> mediaSource;
        err = setupMediaSource(&mediaSource); // very important
        if(err != OK) {
            returnerr;
        }
        sp<MediaSource> encoder;
        err = setupVideoEncoder(mediaSource, videoBitRate, &encoder); // very important
        if(err != OK) {
            returnerr;
        }
        writer->addSource(encoder);
        *totalBitRate += videoBitRate;
    }
    // Audio source is added at the end if it exists.
    // This help make sure that the "recoding" sound is suppressed for
    // camcorder applications in the recorded files.
    if(!mCaptureTimeLapse && (mAudioSource != AUDIO_SOURCE_CNT)) {
        err = setupAudioEncoder(writer); // very important
        if(err != OK) returnerr;
        *totalBitRate += mAudioBitRate;
    }
    ...
    writer->setListener(mListener);
    *mediaWriter = writer;
    returnOK;
}

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
// Set up the appropriate MediaSource depending on the chosen option
status_t StagefrightRecorder::setupMediaSource(
                      sp<MediaSource> *mediaSource) {
    if(mVideoSource == VIDEO_SOURCE_DEFAULT
            || mVideoSource == VIDEO_SOURCE_CAMERA) {
        sp<CameraSource> cameraSource;
        status_t err = setupCameraSource(&cameraSource);
        if(err != OK) {
            returnerr;
        }
        *mediaSource = cameraSource;
    }elseif (mVideoSource == VIDEO_SOURCE_GRALLOC_BUFFER) {
        // If using GRAlloc buffers, setup surfacemediasource.
        // Later a handle to that will be passed
        // to the client side when queried
        status_t err = setupSurfaceMediaSource();
        if(err != OK) {
            returnerr;
        }
        *mediaSource = mSurfaceMediaSource;
    }else{
        returnINVALID_OPERATION;
    }
    returnOK;
}

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
status_t StagefrightRecorder::setupCameraSource(
        sp<CameraSource> *cameraSource) {
    status_t err = OK;
    if((err = checkVideoEncoderCapabilities()) != OK) {
        returnerr;
    }
    Size videoSize;
    videoSize.width = mVideoWidth;
    videoSize.height = mVideoHeight;
    if(mCaptureTimeLapse) {
        if(mTimeBetweenTimeLapseFrameCaptureUs < 0) {
            ALOGE("Invalid mTimeBetweenTimeLapseFrameCaptureUs value: %lld",
                mTimeBetweenTimeLapseFrameCaptureUs);
            returnBAD_VALUE;
        }
        mCameraSourceTimeLapse = CameraSourceTimeLapse::CreateFromCamera(
                mCamera, mCameraProxy, mCameraId,
                videoSize, mFrameRate, mPreviewSurface,
                mTimeBetweenTimeLapseFrameCaptureUs);
        *cameraSource = mCameraSourceTimeLapse;
    }else{
        *cameraSource = CameraSource::CreateFromCamera(
                mCamera, mCameraProxy, mCameraId, videoSize, mFrameRate,
                mPreviewSurface,true/*storeMetaDataInVideoBuffers*/);
    }
    mCamera.clear();
    mCameraProxy.clear();
    if(*cameraSource == NULL) {
        returnUNKNOWN_ERROR;
    }
    if((*cameraSource)->initCheck() != OK) {
        (*cameraSource).clear();
        *cameraSource = NULL;
        returnNO_INIT;
    }
    // When frame rate is not set, the actual frame rate will be set to
    // the current frame rate being used.
    if(mFrameRate == -1) {
        int32_t frameRate = 0;
        CHECK ((*cameraSource)->getFormat()->findInt32(
                    kKeyFrameRate, &frameRate));
        ALOGI("Frame rate is not explicitly set. Use the current frame "
             "rate (%d fps)", frameRate);
        mFrameRate = frameRate;
    }
    CHECK(mFrameRate != -1);
    mIsMetaDataStoredInVideoBuffers =
        (*cameraSource)->isMetaDataStoredInVideoBuffers();
    returnOK;
}

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
status_t StagefrightRecorder::setupVideoEncoder(
        sp<MediaSource> cameraSource,
        int32_t videoBitRate,
        sp<MediaSource> *source) {
    source->clear();
    sp<MetaData> enc_meta = newMetaData;
    enc_meta->setInt32(kKeyBitRate, videoBitRate);
    enc_meta->setInt32(kKeyFrameRate, mFrameRate);
    switch(mVideoEncoder) {
        caseVIDEO_ENCODER_H263:
            enc_meta->setCString(kKeyMIMEType, MEDIA_MIMETYPE_VIDEO_H263);
            break;
        caseVIDEO_ENCODER_MPEG_4_SP:
            enc_meta->setCString(kKeyMIMEType, MEDIA_MIMETYPE_VIDEO_MPEG4);
            break;
        caseVIDEO_ENCODER_H264:
            enc_meta->setCString(kKeyMIMEType, MEDIA_MIMETYPE_VIDEO_AVC);
            break;
        default:
            CHECK(!"Should not be here, unsupported video encoding.");
            break;
    }
    sp<MetaData> meta = cameraSource->getFormat();
    int32_t width, height, stride, sliceHeight, colorFormat;
    CHECK(meta->findInt32(kKeyWidth, &width));
    CHECK(meta->findInt32(kKeyHeight, &height));
    CHECK(meta->findInt32(kKeyStride, &stride));
    CHECK(meta->findInt32(kKeySliceHeight, &sliceHeight));
    CHECK(meta->findInt32(kKeyColorFormat, &colorFormat));
    enc_meta->setInt32(kKeyWidth, width);
    enc_meta->setInt32(kKeyHeight, height);
    enc_meta->setInt32(kKeyIFramesInterval, mIFramesIntervalSec);
    enc_meta->setInt32(kKeyStride, stride);
    enc_meta->setInt32(kKeySliceHeight, sliceHeight);
    enc_meta->setInt32(kKeyColorFormat, colorFormat);
    if(mVideoTimeScale > 0) {
        enc_meta->setInt32(kKeyTimeScale, mVideoTimeScale);
    }
    if(mVideoEncoderProfile != -1) {
        enc_meta->setInt32(kKeyVideoProfile, mVideoEncoderProfile);
    }
    if(mVideoEncoderLevel != -1) {
        enc_meta->setInt32(kKeyVideoLevel, mVideoEncoderLevel);
    }
    OMXClient client;
    CHECK_EQ(client.connect(), (status_t)OK);
    uint32_t encoder_flags = 0;
    if(mIsMetaDataStoredInVideoBuffers) {
        encoder_flags |= OMXCodec::kStoreMetaDataInVideoBuffers;
    }
    // Do not wait for all the input buffers to become available.
    // This give timelapse video recording faster response in
    // receiving output from video encoder component.
    if(mCaptureTimeLapse) {
        encoder_flags |= OMXCodec::kOnlySubmitOneInputBufferAtOneTime;
    }
    sp<MediaSource> encoder = OMXCodec::Create(
            client.interface(), enc_meta,
            true/* createEncoder */, cameraSource,
            NULL, encoder_flags);
    if(encoder == NULL) {
        ALOGW("Failed to create the encoder");
        // When the encoder fails to be created, we need
        // release the camera source due to the camera's lock
        // and unlock mechanism.
        cameraSource->stop();
        returnUNKNOWN_ERROR;
    }
    *source = encoder;
    returnOK;
}

这里和OMXCodec关联起来
有一个叫media_codecs.xml的配置文件来表明设备支持哪些codec

我们录制MPEG 4的时候还会有声音，所以后面还有个setupAudioEncoder，具体的方法就不展开了，总之就是把声音也作为一个Track加入到MPEG4Writer当中去。
这里插个题外话，Google说把setupAudioEncoder放到后面是为了避免开始录影的那一个提示声音也被录制进去，但是实际发现它这样做还是会有bug，在一些设备上还是会把那声录制进去，这个遇到的都是靠APP自己来播放声音来绕过这个问题的。

另外MPEG4Writer当中有个
start(MetaData*)
启动两个方法
a) startWriterThread

启动一个thread去写

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
voidMPEG4Writer::threadFunc() {
    ALOGV("threadFunc");
    prctl(PR_SET_NAME, (unsigned long)"MPEG4Writer",0,0,0);
    Mutex::Autolock autoLock(mLock);
    while(!mDone) {
        Chunk chunk;
        bool chunkFound = false;
        while(!mDone && !(chunkFound = findChunkToWrite(&chunk))) {
            mChunkReadyCondition.wait(mLock);
        }
        // Actual write without holding the lock in order to
        // reduce the blocking time for media track threads.
        if(chunkFound) {
            mLock.unlock();
            writeChunkToFile(&chunk);
            mLock.lock();
        }
    }
    writeAllChunks();
}

b) startTracks

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
status_t MPEG4Writer::startTracks(MetaData *params) {
    for(List<Track *>::iterator it = mTracks.begin();
         it != mTracks.end(); ++it) {
        status_t err = (*it)->start(params);
        if(err != OK) {
            for(List<Track *>::iterator it2 = mTracks.begin();
                 it2 != it; ++it2) {
                (*it2)->stop();
            }
            returnerr;
        }
    }
    returnOK;
}

然后调用每个Track的start方法

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
status_t MPEG4Writer::Track::start(MetaData *params) {
    ...
    initTrackingProgressStatus(params);
    ...
    status_t err = mSource->start(meta.get()); // 这里会去执行CameraSource(start)，这两个是相互关联的
    ...
    pthread_create(&mThread, &attr, ThreadWrapper, this);
    returnOK;
}
void*MPEG4Writer::Track::ThreadWrapper(void*me) {
    Track *track = static_cast<Track *>(me);
    status_t err = track->threadEntry();
    return(void*) err;
}

通过status_t MPEG4Writer::Track::threadEntry()
是新启动另外一个thread，它里面会通过一个循环来不断读取CameraSource(read)里面的数据，CameraSource里面的数据当然是从driver返回过来的(可以参见CameraSourceListener，CameraSource用一个叫做mFrameReceived的List专门存放从driver过来的数据，如果收到数据会调用mFrameAvailableCondition.signal，若还没有开始录影，这个时候收到的数据是被丢弃的，当然MediaWriter先启动的是CameraSource的start方法，再启动写Track)，然后写到文件当中。
注意：准确来说这里MPEG4Writer读取的是OMXCodec里的数据，因为数据先到CameraSource，codec负责编码之后，MPEG4Writer才负责写到文件当中！关于数据在CameraSource/OMXCodec/MPEG4Writer之间是怎么传递的，可以参见http://guoh.org/lifelog/2013/06/interaction-between-stagefright-and-codec/当中讲Buffer的传输过程。

回头再来看，Stagefright做了什么事情？我更觉得它只是一个粘合剂(glue)的用处，它工作在MediaPlayerService这一层，把MediaSource，MediaWriter，Codec以及上层的MediaRecorder绑定在一起，这应该就是它最大的作用，Google用它来替换Opencore也是符合其一贯的工程派作风(相比复杂的学术派而言，虽然Google很多东西也很复杂，但是它一般都是以尽量简单的方式来解决问题)。
让大家觉得有点不习惯的是，它把MediaRecorder放在MediaPlayerService当中，这两个看起来是对立的事情，或者某一天它们会改名字，或者是两者分开，不知道～～

当然这只是个简单的大体介绍，Codec相关的后面争取专门来分析一下！

有些细节的东西在这里没有列出，需要的话会把一些注意点列出来：

1. 时光流逝录影
CameraSource对应的就是CameraSourceTimeLapse

具体做法就是在
dataCallbackTimestamp
当中有skipCurrentFrame

当然它是用些变量来记录和计算
mTimeBetweenTimeLapseVideoFramesUs(1E6/videoFrameRate) // 两个frame之间的间隔时间
记录上一个frame的(mLastTimeLapseFrameRealTimestampUs) // 上一个frame发生的时间
然后通过frame rate计算出两个frame之间的相距离时间，中间的都透过releaseOneRecordingFrame来drop掉
也就是说driver返回的东西都不变，只是在SW这层我们自己来处理掉

关于Time-lapse相关的可以参阅
https://en.wikipedia.org/wiki/Time-lapse_photography

2. 录影当中需要用到Camera的话是通过ICameraRecordingProxy，即Camera当中的RecordingProxy(这是一个BnCameraRecordingProxy)
当透过binder，将ICameraRecordingProxy传到服务端进程之后，它就变成了Bp，如下：

?
1
2
3
4
5
6
7
8
9
caseSET_CAMERA: {
    ALOGV("SET_CAMERA");
    CHECK_INTERFACE(IMediaRecorder, data, reply);
    sp<ICamera> camera = interface_cast<ICamera>(data.readStrongBinder());
    sp<ICameraRecordingProxy> proxy =
        interface_cast<ICameraRecordingProxy>(data.readStrongBinder());
    reply->writeInt32(setCamera(camera, proxy));
    returnNO_ERROR;
}break;

在CameraSource当中会这样去使用

?
1
2
3
4
5
6
7
8
// We get the proxy from Camera, not ICamera. We need to get the proxy
// to the remote Camera owned by the application. Here mCamera is a
// local Camera object created by us. We cannot use the proxy from
// mCamera here.
mCamera = Camera::create(camera);
if(mCamera == 0)return-EBUSY;
mCameraRecordingProxy = proxy;
mCameraFlags |= FLAGS_HOT_CAMERA;

疑问点：

CameraSource当中这个
List<sp > mFramesBeingEncoded;
有什么用？
每编码完一个frame，CameraSource就会将其保存起来，Buffer被release的时候，会反过来release掉这些frame(s)，这种做法是为了效率么？为什么不编码完一个frame就将其release掉？
另外不得不再感叹下Google经常的delete this;行为，精妙，但是看起来反常！

原文地址： http://guoh.org/lifelog/2013/06/android-mediarecorder-architecture/

Android 4.4 MediaRecorder系统结构相关推荐

  1. android 读取音频音量,Android AudioRecord和MediaRecorder录音并实现了实时获取音量大小...

    [实例简介] Android AudioRecord和MediaRecorder录音并实现了实时获取音量大小 看代码就知道了 [实例截图] [核心代码] MediaRecorer和AudioRecor ...

  2. Android MediaRecorder系统结构

    前面有分析过Camera的实现,现在来看看MediaRecorder的实现,这里我不会太去关注它的分层结构,我更关注它的逻辑! APP层 /path/to/aosp/frameworks/base/m ...

  3. Android 中使用MediaRecorder进行录像详解（视频录制）

    简单的视频录制功能. package com.video; import java.io.IOException; import android.app.Activity; import androi ...

  4. Android：使用MediaRecorder录制高清视频，TextureView预览

    用户权限获取 <uses-permission android:name="android.permission.CAMERA" /> <uses-permiss ...

  5. android之通过MediaRecorder进行手机录音

    直接给上个详细的解说代码: package cn.com.chenzheng_java.media; import java.io.IOException; import android.app.Ac ...

  6. Android 录音（MediaRecorder）与播放（MediaPlayer）

    通过MediaRecorder和MediaPlayer实现声音的录制和播放,代码比较简单,直接贴代码.xml文件里面只有四个按钮就不贴了. UI 代码: VoiceActivity.class pac ...

  7. Android视频录制-MediaRecorder流程

    MediaRecorder流程如下: 当使用CameraSource的时候MediaCodecSource会从CameraSource中取数据. 当使用Surface的时候不用CameraSource ...

  8. Android音频录制—MediaRecorder

    MediaRecorder (基于文件录音) 集成了录音,编码,压缩等,支持少量的音频格式文件. 优点:封装度很高,操作简单 缺点:无法实现实时处理音频,输出的音频格式少. 首先来看看 MediaRe ...

  9. Android实现边录边播(MediaRecorder和AudioRecord)

    Android可以通过MediaRecorder和AudioRecord这两个工具来实现录音,MediaRecorder直接把麦克风的数据存到文件,并且能够直接进行编码(如AMR,MP3等),而Aud ...

最新文章

  1. 【转】Linux思维导图
  2. python中的time库安装步骤-python中time库的使用
  3. Latex学习（脚注）
  4. shell编程中date用法(转)
  5. Java里的 for (;;) 与 while (true)，哪个更快？
  6. 【BZOJ2594】水管局长加强版，LCT+并查集+二分查找位置
  7. GCC帧指针的开启与关闭以及反汇编测试
  8. 计算机在剪辑方面的应用,多媒体技术对影视后期制作的应用
  9. win10安装wget
  10. 计算机软件自启动设置,设置开机启动项,详细教您电脑如何设置开机启动项
  11. pm9screw php,PHP使用Screw把源代码加密
  12. （一）前端html+css学习笔记
  13. 准备windows请勿关闭计算机要多久,windows10系统提示正在准备windows请勿关闭计算机怎么办...
  14. HTML/CSS-花样边框案例
  15. 企企通持续助力全球管道预制先行者「迈科管道」，二期项目逐步启动
  16. excel操作技巧，你知道吗？反正我不知道
  17. Python 操控手机 APP
  18. mint系统用wine打开exe文件
  19. UE4-SubSystem
  20. 云计算的应用有哪些？

热门文章

  1. JBPM学习(六)：详解流程图
  2. 关于Unity游戏开发方向找工作方面的一些个人看法
  3. web在线聊天系统。非ajax轮询
  4. ADO.NET Entity Framework
  5. [转]FFT倒序算法—雷德算法
  6. matlab中利用princomp实现PCA降维
  7. tsp遗传算法 c语言,【分享】遗传算法解决TSP问题的源程序
  8. linux卸载python3.6,当python3.6位于/usr/local/bin/python3.6时，如何在ubuntu上卸载它
  9. Python脚本-批量修改文件名
  10. Sdk Manager.exe 闪退问题的解决