Question

enter image description here

如何在iOS中录制语音时以编程方式生成音频波形？

在iOS上处理语音调制音频...一切正常......只需要一些最简单的方法就可以在检测噪声上生成音频波形...

请不要向我推荐... speakhere和auriotouch的代码教程......我需要本机应用程序开发人员的一些最好的建议。

我录制了音频，录制完成后播放。我创建了波形和附加截图。但它必须在视图中绘制为正在录制的录音

-(UIImage *) audioImageGraph:(SInt16 *) samples
                normalizeMax:(SInt16) normalizeMax
                 sampleCount:(NSInteger) sampleCount
                channelCount:(NSInteger) channelCount
                 imageHeight:(float) imageHeight {

    CGSize imageSize = CGSizeMake(sampleCount, imageHeight);
    UIGraphicsBeginImageContext(imageSize);
    CGContextRef context = UIGraphicsGetCurrentContext();

    CGContextSetFillColorWithColor(context, [UIColor blackColor].CGColor);
    CGContextSetAlpha(context,1.0);
    CGRect rect;
    rect.size = imageSize;
    rect.origin.x = 0;
    rect.origin.y = 0;

    CGColorRef leftcolor = [[UIColor whiteColor] CGColor];
    CGColorRef rightcolor = [[UIColor redColor] CGColor];

    CGContextFillRect(context, rect);

    CGContextSetLineWidth(context, 1.0);

    float halfGraphHeight = (imageHeight / 2) / (float) channelCount ;
    float centerLeft = halfGraphHeight;
    float centerRight = (halfGraphHeight*3) ;
    float sampleAdjustmentFactor = (imageHeight/ (float) channelCount) / (float) normalizeMax;

    for (NSInteger intSample = 0 ; intSample < sampleCount ; intSample ++ ) {
        SInt16 left = *samples++;
        float pixels = (float) left;
        pixels *= sampleAdjustmentFactor;
        CGContextMoveToPoint(context, intSample, centerLeft-pixels);
        CGContextAddLineToPoint(context, intSample, centerLeft+pixels);
        CGContextSetStrokeColorWithColor(context, leftcolor);
        CGContextStrokePath(context);

        if (channelCount==2) {
            SInt16 right = *samples++;
            float pixels = (float) right;
            pixels *= sampleAdjustmentFactor;
            CGContextMoveToPoint(context, intSample, centerRight - pixels);
            CGContextAddLineToPoint(context, intSample, centerRight + pixels);
            CGContextSetStrokeColorWithColor(context, rightcolor);
            CGContextStrokePath(context);
        }
    }

    // Create new image
    UIImage *newImage = UIGraphicsGetImageFromCurrentImageContext();

    // Tidy up
    UIGraphicsEndImageContext();

    return newImage;
}

接下来是一个采用AVURLAsset的方法，并返回PNG数据

- (NSData *) renderPNGAudioPictogramForAssett:(AVURLAsset *)songAsset {

    NSError * error = nil;


    AVAssetReader * reader = [[AVAssetReader alloc] initWithAsset:songAsset error:&error];

    AVAssetTrack * songTrack = [songAsset.tracks objectAtIndex:0];

    NSDictionary* outputSettingsDict = [[NSDictionary alloc] initWithObjectsAndKeys:

                                        [NSNumber numberWithInt:kAudioFormatLinearPCM],AVFormatIDKey,
                                        //     [NSNumber numberWithInt:44100.0],AVSampleRateKey, /*Not Supported*/
                                        //     [NSNumber numberWithInt: 2],AVNumberOfChannelsKey,    /*Not Supported*/

                                        [NSNumber numberWithInt:16],AVLinearPCMBitDepthKey,
                                        [NSNumber numberWithBool:NO],AVLinearPCMIsBigEndianKey,
                                        [NSNumber numberWithBool:NO],AVLinearPCMIsFloatKey,
                                        [NSNumber numberWithBool:NO],AVLinearPCMIsNonInterleaved,

                                        nil];


    AVAssetReaderTrackOutput* output = [[AVAssetReaderTrackOutput alloc] initWithTrack:songTrack outputSettings:outputSettingsDict];

    [reader addOutput:output];
    [output release];

    UInt32 sampleRate,channelCount;

    NSArray* formatDesc = songTrack.formatDescriptions;
    for(unsigned int i = 0; i < [formatDesc count]; ++i) {
        CMAudioFormatDescriptionRef item = (CMAudioFormatDescriptionRef)[formatDesc objectAtIndex:i];
        const AudioStreamBasicDescription* fmtDesc = CMAudioFormatDescriptionGetStreamBasicDescription (item);
        if(fmtDesc ) {

            sampleRate = fmtDesc->mSampleRate;
            channelCount = fmtDesc->mChannelsPerFrame;

            //    NSLog(@"channels:%u, bytes/packet: %u, sampleRate %f",fmtDesc->mChannelsPerFrame, fmtDesc->mBytesPerPacket,fmtDesc->mSampleRate);
        }
    }


    UInt32 bytesPerSample = 2 * channelCount;
    SInt16 normalizeMax = 0;

    NSMutableData * fullSongData = [[NSMutableData alloc] init];
    [reader startReading];


    UInt64 totalBytes = 0;


    SInt64 totalLeft = 0;
    SInt64 totalRight = 0;
    NSInteger sampleTally = 0;

    NSInteger samplesPerPixel = sampleRate / 50;


    while (reader.status == AVAssetReaderStatusReading){

        AVAssetReaderTrackOutput * trackOutput = (AVAssetReaderTrackOutput *)[reader.outputs objectAtIndex:0];
        CMSampleBufferRef sampleBufferRef = [trackOutput copyNextSampleBuffer];

        if (sampleBufferRef){
            CMBlockBufferRef blockBufferRef = CMSampleBufferGetDataBuffer(sampleBufferRef);

            size_t length = CMBlockBufferGetDataLength(blockBufferRef);
            totalBytes += length;


            NSAutoreleasePool *wader = [[NSAutoreleasePool alloc] init];

            NSMutableData * data = [NSMutableData dataWithLength:length];
            CMBlockBufferCopyDataBytes(blockBufferRef, 0, length, data.mutableBytes);


            SInt16 * samples = (SInt16 *) data.mutableBytes;
            int sampleCount = length / bytesPerSample;
            for (int i = 0; i < sampleCount ; i ++) {

                SInt16 left = *samples++;

                totalLeft  += left;



                SInt16 right;
                if (channelCount==2) {
                    right = *samples++;

                    totalRight += right;
                }

                sampleTally++;

                if (sampleTally > samplesPerPixel) {

                    left  = totalLeft / sampleTally;

                    SInt16 fix = abs(left);
                    if (fix > normalizeMax) {
                        normalizeMax = fix;
                    }


                    [fullSongData appendBytes:&left length:sizeof(left)];

                    if (channelCount==2) {
                        right = totalRight / sampleTally;


                        SInt16 fix = abs(right);
                        if (fix > normalizeMax) {
                            normalizeMax = fix;
                        }


                        [fullSongData appendBytes:&right length:sizeof(right)];
                    }

                    totalLeft   = 0;
                    totalRight  = 0;
                    sampleTally = 0;

                }
            }



            [wader drain];


            CMSampleBufferInvalidate(sampleBufferRef);

            CFRelease(sampleBufferRef);
        }
    }


    NSData * finalData = nil;

    if (reader.status == AVAssetReaderStatusFailed || reader.status == AVAssetReaderStatusUnknown){
        // Something went wrong. return nil

        return nil;
    }

    if (reader.status == AVAssetReaderStatusCompleted){

        NSLog(@"rendering output graphics using normalizeMax %d",normalizeMax);

        UIImage *test = [self audioImageGraph:(SInt16 *)
                         fullSongData.bytes
                                 normalizeMax:normalizeMax
                                  sampleCount:fullSongData.length / 4
                                 channelCount:2
                                  imageHeight:100];

        finalData = imageToData(test);
    }




    [fullSongData release];
    [reader release];

    return finalData;
}

我有

Answer 1

如果您想要从麦克风输入派生的实时图形，那么请使用RemoteIO音频单元，这是大多数原生iOS应用程序开发人员用于低延迟音频，而Metal或Open GL用于绘制波形，这将为您提供最高帧率。您将需要与问题中提供的代码完全不同的代码，因为AVAssetRecording，Core Graphic线条图和png渲染使用起来太慢了。

更新：使用iOS 8及更新版本，Metal API可以渲染图形可视化，其性能甚至超过OpenGL。

Uodate 2：以下是一些代码片段，用于使用音频单元录制实时音频，并使用Swift 3中的Metal绘制位图：https://gist.github.com/hotpaw2/f108a3c785c7287293d7e1e81390c20b

Answer 2

你应该看看EZAudio（https://github.com/syedhali/EZAudio），特别是EZRecorder和EZAudioPlot（或GPU加速的EZAudioPlotGL）。

还有一个示例项目正是您想要的，https://github.com/syedhali/EZAudio/tree/master/EZAudioExamples/iOS/EZAudioRecordExample

编辑：这是代码内联

/// In your interface

/**
 Use a OpenGL based plot to visualize the data coming in
 */
@property (nonatomic,weak) IBOutlet EZAudioPlotGL *audioPlot;
/**
 The microphone component
 */
@property (nonatomic,strong) EZMicrophone *microphone;
/**
 The recorder component
 */
@property (nonatomic,strong) EZRecorder *recorder;

...

/// In your implementation

// Create an instance of the microphone and tell it to use this view controller instance as the delegate
-(void)viewDidLoad {
    self.microphone = [EZMicrophone microphoneWithDelegate:self startsImmediately:YES];
}

// EZMicrophoneDelegate will provide these callbacks
-(void)microphone:(EZMicrophone *)microphone
 hasAudioReceived:(float **)buffer
   withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels {
  dispatch_async(dispatch_get_main_queue(),^{
    // Updates the audio plot with the waveform data
    [self.audioPlot updateBuffer:buffer[0] withBufferSize:bufferSize];
  });
}

-(void)microphone:(EZMicrophone *)microphone hasAudioStreamBasicDescription:(AudioStreamBasicDescription)audioStreamBasicDescription {
  // The AudioStreamBasicDescription of the microphone stream. This is useful when configuring the EZRecorder or telling another component what audio format type to expect.

  // We can initialize the recorder with this ASBD
  self.recorder = [EZRecorder recorderWithDestinationURL:[self testFilePathURL]
                                         andSourceFormat:audioStreamBasicDescription];

}

-(void)microphone:(EZMicrophone *)microphone
    hasBufferList:(AudioBufferList *)bufferList
   withBufferSize:(UInt32)bufferSize
withNumberOfChannels:(UInt32)numberOfChannels {

  // Getting audio data as a buffer list that can be directly fed into the EZRecorder. This is happening on the audio thread - any UI updating needs a GCD main queue block. This will keep appending data to the tail of the audio file.
  if( self.isRecording ){
    [self.recorder appendDataFromBufferList:bufferList
                             withBufferSize:bufferSize];
  }

}

Answer 3

我正在寻找同样的事情。（从录音机的数据中制作波形）。我找到了一些可能有用的库，值得检查代码以了解其背后的逻辑。

计算全部基于罪和数学公式。如果您查看代码，这很简单！

https://github.com/stefanceriu/SCSiriWaveformView

或

https://github.com/raffael/SISinusWaveView

这只是您可以在网上找到的几个例子。

如何在iOS中录制语音时以编程方式生成音频波形？

3 个答案: