我已经研究过要在iphone中播放与频率有关的哔声。我给的分贝。
我提到的链接：
http://www.politepix.com/2010/06/18/decibel-metering-from-an-iphone-audio-unit/
http://atastypixel.com/blog/using-remoteio-audio-unit/
http://www.politepix.com/2010/06/18/decibel-metering-from-an-iphone-audio-unit/
How to play a sound of paticular frequency and framework not found AudioUnit question
我也使用Flite在我的应用程序中进行文本到语音。
我可以知道，是否可以在iphone中播放与频率相关的哔声。使用flite分贝。
我知道他们正在根据输入创建一个音频文件（仅与音高，方差，速度和给定的字符串相关）并在创建后通过它播放Audioplayer。
但他们没有自定义方法来设置频率和频率。分贝!!!!
所以任何人都可以在iPhone上为我提供一个好方法。
对此问题的任何帮助表示赞赏。
由于
答案 0 :(得分：0)
此类允许您以给定频率和给定幅度发出蜂鸣声。 它使用来自 AudioToolbox.framework 的AudioQueues。它只是一个草图，许多事情应该被改进，但创建信号的机制是有效的。
如果您看到@interface。
#import <AudioToolbox/AudioToolbox.h>
#define TONE_SAMPLERATE 44100.
@interface Tone : NSObject {
AudioQueueRef queue;
AudioQueueBufferRef buffer;
BOOL rebuildBuffer;
}
@property (nonatomic, assign) NSUInteger frequency;
@property (nonatomic, assign) CGFloat dB;
- (void)play;
- (void)pause;
@end
@implementation Tone
@synthesize dB=_dB,frequency=_frequency;
void handleBuffer(void *inUserData,
AudioQueueRef inAQ,
AudioQueueBufferRef inBuffer);
#pragma mark - Initialization and deallocation -
- (id)init
{
if ((self=[super init])) {
_dB=0.;
_frequency=440;
rebuildBuffer=YES;
// TO DO: handle AudioQueueXYZ's failures!!
// create a descriptor containing a LPCM, mono, float format
AudioStreamBasicDescription desc;
desc.mSampleRate=TONE_SAMPLERATE;
desc.mFormatID=kAudioFormatLinearPCM;
desc.mFormatFlags=kLinearPCMFormatFlagIsFloat;
desc.mBytesPerPacket=sizeof(float);
desc.mFramesPerPacket=1;
desc.mBytesPerFrame=sizeof(float);
desc.mChannelsPerFrame=1;
desc.mBitsPerChannel=8*sizeof(float);
// create a new queue
AudioQueueNewOutput(&desc,
&handleBuffer,
self,
CFRunLoopGetCurrent(),
kCFRunLoopCommonModes,
0,
&queue);
// and its buffer, ready to hold 1" of data
AudioQueueAllocateBuffer(queue,
sizeof(float)*TONE_SAMPLERATE,
&buffer);
// create the buffer and enqueue it
handleBuffer(self, queue, buffer);
}
return self;
}
- (void)dealloc
{
AudioQueueStop(queue, YES);
AudioQueueFreeBuffer(queue, buffer);
AudioQueueDispose(queue, YES);
[super dealloc];
}
#pragma mark - Main function -
void handleBuffer(void *inUserData,
AudioQueueRef inAQ,
AudioQueueBufferRef inBuffer) {
// this function takes care of building the buffer and enqueuing it.
// cast inUserData type to Tone
Tone *tone=(Tone *)inUserData;
// check if the buffer must be rebuilt
if (tone->rebuildBuffer) {
// precompute some useful qtys
float *data=inBuffer->mAudioData;
NSUInteger max=inBuffer->mAudioDataBytesCapacity/sizeof(float);
// multiplying the argument by 2pi changes the period of the cosine
// function to 1s (instead of 2pi). then we must divide by the sample
// rate to get TONE_SAMPLERATE samples in one period.
CGFloat unit=2.*M_PI/TONE_SAMPLERATE;
// this is the amplitude converted from dB to a linear scale
CGFloat amplitude=pow(10., tone.dB*.05);
// loop and simply set data[i] to the value of cos(...)
for (NSUInteger i=0; i<max; ++i)
data[i]=(float)(amplitude*cos(unit*(CGFloat)(tone.frequency*i)));
// inform the queue that we have filled the buffer
inBuffer->mAudioDataByteSize=sizeof(float)*max;
// and set flag
tone->rebuildBuffer=NO;
}
// reenqueue the buffer
AudioQueueEnqueueBuffer(inAQ,
inBuffer,
0,
NULL);
/* TO DO: the transition between two adjacent buffers (the same one actually)
generates a "tick", even if the adjacent buffers represent a continuous signal.
maybe using two buffers instead of one would fix it.
*/
}
#pragma - Properties and methods -
- (void)play
{
// generate an AudioTimeStamp with "0" simply!
// (copied from FillOutAudioTimeStampWithSampleTime)
AudioTimeStamp time;
time.mSampleTime=0.;
time.mRateScalar=0.;
time.mWordClockTime=0.;
memset(&time.mSMPTETime, 0, sizeof(SMPTETime));
time.mFlags = kAudioTimeStampSampleTimeValid;
// TO DO: maybe it could be useful to check AudioQueueStart's return value
AudioQueueStart(queue, &time);
}
- (void)pause
{
// TO DO: maybe it could be useful to check AudioQueuePause's return value
AudioQueuePause(queue);
}
- (void)setFrequency:(NSUInteger)frequency
{
if (_frequency!=frequency) {
_frequency=frequency;
// we need to update the buffer (as soon as it stops playing)
rebuildBuffer=YES;
}
}
- (void)setDB:(CGFloat)dB
{
if (dB!=_dB) {
_dB=dB;
// we need to update the buffer (as soon as it stops playing)
rebuildBuffer=YES;
}
}
@end
该类产生以给定整数频率振荡的cos波形（幅度* cos（2pi *频率* t））;整个工作由void handleBuffer(...)完成，使用具有线性PCM，mono，float @ 44.1kHz格式的AudioQueue。要更改信号形状，您只需更改该行即可。例如，以下代码将生成方波：
float x = fmodf(unit*(CGFloat)(tone.frequency*i), 2 * M_PI);
data[i] = amplitude * (x > M_PI ? -1.0 : 1.0);
对于浮点频率，您应该考虑在一秒钟的音频数据中没有必要的整数个振荡，因此所表示的信号在两个缓冲区之间的连接处是不连续的，并产生一个奇怪的'蜱'。例如，您可以设置较少的样本，以便结点处于信号周期的末尾。