Question

I have to implement a iphone application which will record user’s voice as you starts speaking, and change the pitch of the recorded sound and play it back. I am able to record the audio on the detection of the sound by the help of AVAudiorecorder, and using Dirac library I have changed the pitch of recorded sound. The problem with this approach is that the output sound is noisy enough. I got the response for using SoundEngine but I didn’t get the way to implement it. Can anyone please explain me about any other way to implement this?

my code//
        -(void)initialSetup
    { 
        count=0; 
        silenceTime=0;

    //[[AVAudioSession sharedInstance] setCategory:AVAudioSessionCategoryPlayAndRecord error: nil];
    recordSetting = [[NSMutableDictionary alloc] init];
    [recordSetting setValue:[NSNumber numberWithInt:kAudioFormatAppleLossless] forKey:AVFormatIDKey];
    [recordSetting setValue:[NSNumber numberWithFloat:44100.0] forKey:AVSampleRateKey]; 
    [recordSetting setValue:[NSNumber numberWithInt: 2] forKey:AVNumberOfChannelsKey];
    recordedTmpFile = [NSURL fileURLWithPath:[NSTemporaryDirectory() stringByAppendingPathComponent:[NSString stringWithFormat: @"%.0f.%@",[NSDate timeIntervalSinceReferenceDate]*1000.0,                  @"caf"]]];
    recorder = [[ AVAudioRecorder alloc] initWithURL:recordedTmpFile settings:recordSetting error:&error];
    //recorder = [[ AVAudioRecorder alloc] init];
    [recorder setDelegate:self];
    [recorder updateMeters];
    [recorder prepareToRecord];
    //[[AVAudioSession sharedInstance] setCategory: AVAudioSessionCategoryPlayAndRecord error: nil];
    //In Order To Move Sound To The Speaker
    //UInt32 audioRouteOverride = kAudioSessionOverrideAudioRoute_Speaker;
    //AudioSessionSetProperty (kAudioSessionProperty_OverrideAudioRoute,sizeof(audioRouteOverride),&audioRouteOverride);

    NSArray *dirPaths;
    NSString *docsDir;
    dirPaths = NSSearchPathForDirectoriesInDomains(NSDocumentDirectory, NSUserDomainMask, YES);
    docsDir = [dirPaths objectAtIndex:0];
    NSString *soundFilePath = [docsDir stringByAppendingPathComponent:@"audio.caf"];
    recordedTmpFile1 =  [NSURL fileURLWithPath:soundFilePath];

    recordSetting1 =   [[NSMutableDictionary alloc] init];
    recordSetting1 =   [NSDictionary dictionaryWithObjectsAndKeys:
                        [NSNumber numberWithInt:AVAudioQualityMin],AVEncoderAudioQualityKey,
                        //[NSNumber numberWithInt:kAudioFormatAppleIMA4],AVFormatIDKey,
                        [NSNumber numberWithInt:16], 
                        AVEncoderBitRateKey,
                        [NSNumber numberWithInt: 2], 
                        AVNumberOfChannelsKey,
                        [NSNumber numberWithFloat:44100.0], 
                        AVSampleRateKey,nil];
    recorder1 = [[AVAudioRecorder alloc] initWithURL:recordedTmpFile1 settings:recordSetting1 error:&error];
    [recorder1 prepareToRecord];
    [recorder1 setDelegate:self];
    if(recorder) 
    {
        recorder.meteringEnabled = YES;
        [recorder record];
        double val=[recorder peakPowerForChannel:0];
        NSLog(@"The Very First Value Of The Recorder Is=%f",val);
        levelTimer = [NSTimer scheduledTimerWithTimeInterval:0.4 target: self selector: @selector(levelTimerCallback:) userInfo: nil repeats: YES];
    }
    else
    {
        NSLog(@"error in initilising of the recorder=%@",[error localizedDescription]);
    }
}


-(void)levelTimerCallback:(NSTimer *)timer                      
{ 
    [recorder updateMeters];
    const double ALPHA = 0.05;
    //NOISE FILERATION ALGORITHMS
    double peakPowerForChannel = pow(10,(0.05 *[recorder peakPowerForChannel:0]));
    double audioMonitorResults1 = ALPHA * peakPowerForChannel + (1.0 - ALPHA) * audioMonitorResults1;
    double audioMonitorResults;
    audioMonitorResults= [recorder peakPowerForChannel:0];
    NSLog(@"This time only  frequency is==>%f",audioMonitorResults1);
    //if (audioMonitorResults1 >0.020)
    if(audioMonitorResults1 > .05)//the value of audioMonitorResults may be equal to -10 for device
    {

        [recorder1 updateMeters];
        recorder1.meteringEnabled=YES;
        //recorder.meteringEnabled=YES;
        [recorder1 record];
        NSLog(@"SOUND IS DETECTED");
        NSLog(@"%f",[recorder1 peakPowerForChannel:0]);
        NSLog(@"Recording is going on");
        count=1;
        silenceTime=0;

    }
    else
    {

        NSLog(@"NO SOUND IS DETECTED");
        silenceTime=silenceTime+0.3;
        if(count==1 && silenceTime>1)
        { 

            [levelTimer invalidate];
            [recorder1 stop];

        }

    }

}
- (void)audioRecorderDidFinishRecording:(AVAudioRecorder *)recorder successfully:(BOOL)flag
{
    NSLog(@"Recorder stop delegate is processing");
    if(flag)
    {
        NSLog(@"Player has finished successfully");
        [self playRecording];
    }
    else
    { 
        NSLog(@"problem in recording.......not recorded");
    }
}

Answer 1

1. How to detect the sound?
   I solved my first problem with the help of a tutorial..here is the link for this,
   link: http://mobileorchard.com/tutorial-detecting-when-a-user-blows-into-the-mic/
   Here we can easily understand the recording of the sound on detection of some noise.

2. How to change the pitch of the recorded sound and play back.
   Second problem I faced in changing the pitch of the sound. As we can only record the sound with the help of AVAudioRecorder we can’t alter the pitch by that.
   For this purpose I used an external library DIRAC.
   Here is the link for the Dirac library.
   This comes up with some sample project (for mobile apps as well as desktop apps)about the implementation of the dirac library to the application.

Another way I found to solve this issue by implementing Cocoas2D, Cocos Denshion with Dirac. Because the above proess was not working well with my application.
Here is the link for implementing this (a sample project about the changing the pitch and play the recorded sound).

I found another link related to sound recording.