【发布时间】:2011-08-06 15:58:46
【问题描述】:
我做了一个简单的应用程序,当超过一定的噪音水平时会发出警报。因此,我有一个 AudioQueue 可以记录声音并测量记录声音的级别(仅重要的代码部分如下所示):
#import "AudioRecorder.h"
#include <AudioToolbox/AudioToolbox.h>
#include <iostream>
using namespace std;
@implementation AudioRecorder
@synthesize sp; //custom object SoundPlayer
@synthesize bias; //a bias, if the soundlevel exeeds this bias something happens
AudioRecorder* ar;
//callback function to handle the audio data contained in the audio buffer.
//In my case it is called every 0.5 seconds
static void HandleInputBuffer (...) {
...
char* levelMeterData = new char[size];
AudioQueueGetProperty ( inAQ, kAudioQueueProperty_CurrentLevelMeter, levelMeterData, &size );
AudioQueueLevelMeterState* meterState = reinterpret_cast<AudioQueueLevelMeterState*>(levelMeterData);
cout << "mAveragePower = " << meterState->mAveragePower << endl;
cout << "mPeakPower = " << meterState->mPeakPower << endl;
if( meterState->mPeakPower > ar.bias )
[ar playAlarmSound];
}
...
//The constructor of the AudioRecorder class
-(id) init {
self = [super init];
if( self ) {
ar = self;
sp = [[SoundPlayer alloc] init];
}
return self;
}
-(void)playAlarmSound {
[sp playSound];
}
所以这里基本上发生的情况如下:
我在 iphone 屏幕上点击一个按钮,该按钮导致音频队列从 iphone 中的麦克风录制声音。当队列已满时,将调用回调函数“HandleInputBuffer”来处理音频缓冲区中的数据。在我的特定情况下,处理数据意味着我想测量声音强度。如果强度超过偏差,则调用“playAlarmSound”方法。
所以这个调用发生在主线程之外,即在一个额外的线程中。
对象“sp”(SoundPlayer)有以下实现:
//callback function. Gets called when the sound has finished playing
void soundDidFinishPlaying( SystemSoundID ssID, void *clientData ) {
NSLog(@"Finished playing system sound");
sp.soundIsPlaying = NO;
}
-(id)init {
self = [super init];
if(self) {
self.soundIsPlaying = NO;
srand(time(NULL));
soundFilenames = [[NSArray alloc] initWithObjects:@"Alarm Clock Bell.wav", @"Bark.wav", @"Cartoon Boing.wav", @"Chimpanzee Calls.wav", @"School Bell Ringing.wav", @"Sheep Bah.wav", @"Squeeze Toy.wav", @"Tape Rewinding.wav", nil];
[self copySoundsIfNeeded];
sp = self;
[self playSound]; //gets played without any problems
}
return self;
}
-(void)playSound {
if( !self.soundIsPlaying ) {
NSAutoreleasePool *pool = [[NSAutoreleasePool alloc] init];
int index = rand() % [soundFilenames count];
NSString* filePath = [ [self getBasePath] stringByAppendingPathComponent: [soundFilenames objectAtIndex:index] ];
NSFileManager *fileManager = [NSFileManager defaultManager];
if( [fileManager fileExistsAtPath:filePath] )
NSLog(@"File %@ exists", [soundFilenames objectAtIndex:index]);
else
NSLog(@"File %@ NOT exists", [soundFilenames objectAtIndex:index]);
CFURLRef url = CFURLCreateWithFileSystemPath ( 0, (CFStringRef) filePath, kCFURLPOSIXPathStyle, NO );
SystemSoundID outSystemSoundID;
AudioServicesCreateSystemSoundID( url, &outSystemSoundID );
AudioServicesAddSystemSoundCompletion ( outSystemSoundID, 0, 0, soundDidFinishPlaying, 0 );
self.soundIsPlaying = YES;
AudioServicesPlaySystemSound( outSystemSoundID );
[pool drain];
}
}
实际上代码工作正常,所以代码中应该没有错误,而且波形文件的格式是正确的。这就是发生的事情:
iPhone 模拟器: 一切正常。在创建对象 SoundPlayer 时,在构造函数中调用 [self playSound] 会使模拟器播放随机声音,从而证明该方法已正确实现。然后我开始录音,当超过某个声级时,从 AudioRecorder(在一个单独的线程中)调用该方法,并且声音也正在播放。所以这是期望的行为。
实际 iPhone 设备: 创建对象 SoundPlayer 时,会播放随机声音,因此可以正常工作。但是当我开始录音并且超过噪音水平时,Soundplayer 中的方法 playSound 被 AudioRecorder 调用,但没有播放声音。我怀疑这是因为这发生在一个单独的线程中。但我不知道如何修复。
我已经尝试通过使用来自默认通知中心的通知来修复它,但它也不起作用。如何管理声音的播放?
【问题讨论】:
标签: ios multithreading audio