参考文档:
1:github上的一个开源代码:点击打开链接,因为存储音频需要自己写文件头部,所以这里借鉴了它的文件存储功能。
2:一个很有用的帖子代码:点击打开链接,虽然有点小瑕疵,但是下面有解答。
3:一个说明类的博客:点击打开链接
这里加一些说明:windows下多线程尽量用_beginthreadex而不是CreateThread,网上原因说的很详细(这里需要说明一点,使用_beginthreadex时runtime library必须是/MT或/MTd。但是很可惜如果你用的是cocos2d-x引擎,默认的是/MD,因此还是乖乖的用pthread吧,否则会有编译错误。选择runtime library的步骤为,项目-->属性-->C/C++-->代码生成-->运行库)。其次,在关闭录音时,一定不要对waveIn设备进行缓冲区操作,要不然容易生成死锁。示例代码如下:
创建终止拷贝线程的代码:(上文博客里面提示,不能在录音的回调函数里将数据写到文件,因此需要创建线程,单独拷贝)
bool stopRecording(const char *fileName, int nameSize){
char *fileNameTrue = (char *)malloc(sizeof(char)* (nameSize+1));
strcpy_s(fileNameTrue,nameSize+1, fileName);
//终止录音
unsigned threadID;
hAudio_Detect_Thread = (HANDLE)_beginthreadex(NULL, 0, &Audio_Detect, fileNameTrue, 0, &threadID);
WaitForSingleObject(hAudio_Detect_Thread, INFINITE);
CloseHandle(hAudio_Detect_Thread);
return true;
}
void processSamples()
{
if(bRecordingStarted ){
waveInUnprepareHeader(hWaveIn, WaveHeaders[iBuffIdx], sizeof(WAVEHDR));
iBuffIdx ++;
if (iBuffIdx == BUFFERCOUNT)
{
iBuffIdx = 0;
}
WaveHeaders[iBuffIdxT]->dwBufferLength = WaveBufSize;
WaveHeaders[iBuffIdxT]->dwFlags = 0;
WaveHeaders[iBuffIdxT]->dwLoops = 0;
waveInPrepareHeader(hWaveIn, WaveHeaders[iBuffIdxT], sizeof(WAVEHDR));
waveInAddBuffer(hWaveIn, WaveHeaders[iBuffIdxT], sizeof(WAVEHDR));
<span style="white-space:pre"> </span>iBuffidxT ++;
if (iBuffIdxT == BUFFERCOUNT)
{
iBuffIdxT = 0;
}
}
}
void writeData(char *name){
//停止录音
printf("进入写的函数里面了, the name is %s\n", name);
bRecordingStarted = false;
waveInReset(hWaveIn);
int stopcode = waveInStop(hWaveIn);
printf("the stopcode is %d\n", stopcode);
COutputFile *file = NULL;
file = new CWaveOutputFile;
if (!file->Create(name, m_waveformin))
{
printf("初始化就错误了\n");
}
printf("正在写数据\n");
file->WriteData(m_AudioDataPool.lpdata, m_AudioDataPool.dwLength);
printf("the length of m_AudioDataPool.length is %d", m_AudioDataPool.dwLength);
file->Flush();
//释放资源
free(name);
free(m_AudioDataPool.lpdata);
m_AudioDataPool.lpdata = NULL;
m_AudioDataPool.dwLength = 0;
}