如何从AudioBufferList获取样本

Question

我从wav文件(其采样率为44100 is，长时间为2秒)中获得AudioBufferList。但我拿不到44100*2=88200样本。实际上，我得到了一个包含512 nNumberBuffers的AudiobufferList。如何从AudioBufferList获取样本？

Answer 1

下面是我从wav文件中获取样本的方法

-(float *) GetSoundFile:(NSString *)fileName
{
NSBundle *bundle = [NSBundle mainBundle];
NSString *path = [bundle pathForResource:[fileName stringByDeletingPathExtension]
                                  ofType:[fileName pathExtension]];

NSURL *audioURL = [NSURL fileURLWithPath:path];
if (!audioURL)
{
    NSLog(@"file: %@ not found.", fileName);
}

OSStatus                        err = 0;
                                theFileLengthInFrames = 0;  //this is global
AudioStreamBasicDescription     theFileFormat;
UInt32                          thePropertySize = sizeof(theFileFormat);
ExtAudioFileRef                 extRef = NULL;
void*                           theData = NULL;
AudioStreamBasicDescription     theOutputFormat;

// Open a file with ExtAudioFileOpen()
err = ExtAudioFileOpenURL((__bridge CFURLRef)(audioURL), &extRef);

// Get the audio data format
err = ExtAudioFileGetProperty(extRef, kExtAudioFileProperty_FileDataFormat, &thePropertySize, &theFileFormat);

theOutputFormat.mSampleRate = samplingRate = theFileFormat.mSampleRate;
theOutputFormat.mChannelsPerFrame = numChannels = 1;

theOutputFormat.mFormatID = kAudioFormatLinearPCM;
theOutputFormat.mBytesPerFrame = sizeof(Float32) * theOutputFormat.mChannelsPerFrame ;
theOutputFormat.mFramesPerPacket = theFileFormat.mFramesPerPacket;
theOutputFormat.mBytesPerPacket = theOutputFormat.mFramesPerPacket * theOutputFormat.mBytesPerFrame;
theOutputFormat.mBitsPerChannel = sizeof(Float32) * 8 ;
theOutputFormat.mFormatFlags = 9 | 12;

//set the output property
err = ExtAudioFileSetProperty(extRef, kExtAudioFileProperty_ClientDataFormat, sizeof(AudioStreamBasicDescription), &theOutputFormat);

// Here I Get the total frame count and write it into gloabal variable
thePropertySize = sizeof(theFileLengthInFrames);
err = ExtAudioFileGetProperty(extRef, kExtAudioFileProperty_FileLengthFrames, &thePropertySize, &theFileLengthInFrames);

UInt32 dataSize = (UInt32)(theFileLengthInFrames * theOutputFormat.mBytesPerFrame);
theData = malloc(dataSize);

AudioBufferList theDataBuffer;
if (theData)
{
    theDataBuffer.mNumberBuffers = 1;
    theDataBuffer.mBuffers[0].mDataByteSize = dataSize;
    theDataBuffer.mBuffers[0].mNumberChannels = theOutputFormat.mChannelsPerFrame;
    theDataBuffer.mBuffers[0].mData = theData;

    // Read the data into an AudioBufferList
    err = ExtAudioFileRead(extRef, (UInt32*)&theFileLengthInFrames, &theDataBuffer);
}

return (float*)theDataBuffer.mBuffers[0].mData;  
//here is the data that has thefilelengthInframes amount frames 

}

Answer 2

AudioBufferList不应设置为固定大小，因为它是一个临时缓冲区，依赖于硬件。大小是不可预测的，您只能设置一个更好的大小，而且不允许在两次读取调用之间保持相同的大小。要获得88200个采样(或您喜欢的)，您必须使用AudioBufferList中的采样，按时间递增地填充新缓冲区。我建议使用这个循环缓冲区https://github.com/michaeltyson/TPCircularBuffer。希望这对你有所帮助

Answer 3

据我所知，AudioBufferList必须由您配置以接收数据，然后必须由某个读取函数(例如ExtAudioFileRead())填充。因此，您应该通过分配所需的缓冲区(通常是1或2)来准备它，将nNumberBuffers设置为该数字，并将音频数据读取给它们。AudioBufferList只是存储该缓冲区，它们将包含帧值。