iphone - 如何在 iOS 中录制语音时以编程方式生成音频波形？

Question

如何在 iOS 中录制语音时以编程方式生成音频波形？

我正在 iOS 中处理语音调制音频...一切正常...只需要一些最简单的方法来生成检测噪声的音频波形...

请不要向我推荐...speakhere 和 auriotouch...的代码教程...我需要原生应用开发人员的一些最佳建议。

我已经录制了音频，并在录制后播放了它。我已经创建了波形并附上了屏幕截图。但它必须在 View 中绘制为正在进行的录音

-(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;
}

接下来是一个接受 AVURLAsset 并返回 PNG 数据的方法

- (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获得实时图形，请使用 RemoteIO Audio Unit，这是大多数原生 iOS 应用程序开发人员用于低延迟音频的工具，以及 Metal 或 Open GL 用于绘制波形，这将为您提供最高帧率。为此，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:在 iOS 8 及更高版本中，Metal API 可能能够以比 OpenGL 更高的性能呈现图形可视化。

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