用AVAssetReader绘制波形

Question

我使用assetUrl (代码名为audioUrl)从iPod库中读取歌曲，我可以用多种方式播放它，我可以剪切它，我可以用它进行一些处理，但是……我真的不明白我该怎么用这个CMSampleBufferRef来获取绘制波形的数据！我需要关于峰值的信息，我如何才能以这种(也许是另一种)方式获得它？

    AVAssetTrack * songTrack = [audioUrl.tracks objectAtIndex:0];
    AVAssetReaderTrackOutput * output = [[AVAssetReaderTrackOutput alloc] initWithTrack:songTrack outputSettings:nil];
    [reader addOutput:output];
    [output release];

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

    while (reader.status == AVAssetReaderStatusReading){

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

        CMSampleBufferRef sampleBufferRef = [trackOutput copyNextSampleBuffer];

        if (sampleBufferRef){/* what I gonna do with this? */}

请帮帮我!

Answer 1

您应该能够从sampleBuffRef获取音频缓冲区，然后迭代这些值以构建波形：

CMBlockBufferRef buffer = CMSampleBufferGetDataBuffer( sampleBufferRef );
CMItemCount numSamplesInBuffer = CMSampleBufferGetNumSamples(sampleBufferRef);
AudioBufferList audioBufferList;
CMSampleBufferGetAudioBufferListWithRetainedBlockBuffer(
                                                            sampleBufferRef,
                                                            NULL,
                                                            &audioBufferList,
                                                            sizeof(audioBufferList),
                                                            NULL,
                                                            NULL,
                                                                  kCMSampleBufferFlag_AudioBufferList_Assure16ByteAlignment,
                                                            &buffer
                                                            );

// this copies your audio out to a temp buffer but you should be able to iterate through this buffer instead
SInt32* readBuffer = (SInt32 *)malloc(numSamplesInBuffer * sizeof(SInt32));
memcpy( readBuffer, audioBufferList.mBuffers[0].mData, numSamplesInBuffer*sizeof(SInt32));

Answer 2

另一种使用Swift 5并使用AVAudioFile的方法：

///Gets the audio file from an URL, downsaples and draws into the sound layer.
func drawSoundWave(fromURL url:URL, fromPosition:Int64, totalSeconds:UInt32, samplesSecond:CGFloat) throws{

    print("\(logClassName) Drawing sound from \(url)")

    do{
        waveViewInfo.samplesSeconds = samplesSecond

        //Get audio file and format from URL
        let audioFile = try AVAudioFile(forReading: url)

        waveViewInfo.format = audioFile.processingFormat
        audioFile.framePosition = fromPosition * Int64(waveViewInfo.format.sampleRate)

        //Getting the buffer
        let frameCapacity:UInt32 = totalSeconds * UInt32(waveViewInfo.format.sampleRate)

        guard let audioPCMBuffer = AVAudioPCMBuffer(pcmFormat: waveViewInfo.format, frameCapacity: frameCapacity) else{ throw AppError("Unable to get the AVAudioPCMBuffer") }
        try audioFile.read(into: audioPCMBuffer, frameCount: frameCapacity)
        let audioPCMBufferFloatValues:[Float] = Array(UnsafeBufferPointer(start: audioPCMBuffer.floatChannelData?.pointee,
                                                                          count: Int(audioPCMBuffer.frameLength)))

        waveViewInfo.points = []
        waveViewInfo.maxValue = 0
        for index in stride(from: 0, to: audioPCMBufferFloatValues.count, by: Int(audioFile.fileFormat.sampleRate) / Int(waveViewInfo.samplesSeconds)){

            let aSample = CGFloat(audioPCMBufferFloatValues[index])
            waveViewInfo.points.append(aSample)
            let fix = abs(aSample)
            if fix > waveViewInfo.maxValue{
                waveViewInfo.maxValue = fix
            }

        }

        print("\(logClassName) Finished the points - Count = \(waveViewInfo.points.count) / Max = \(waveViewInfo.maxValue)")

        populateSoundImageView(with: waveViewInfo)

    }
    catch{

        throw error

    }

}

///Converts the sound wave in to a UIImage
func populateSoundImageView(with waveViewInfo:WaveViewInfo){

    let imageSize:CGSize = CGSize(width: CGFloat(waveViewInfo.points.count),//CGFloat(waveViewInfo.points.count) * waveViewInfo.sampleSpace,
                                  height: frame.height)
    let drawingRect = CGRect(origin: .zero, size: imageSize)

    UIGraphicsBeginImageContextWithOptions(imageSize, false, 0)
    defer {
        UIGraphicsEndImageContext()
    }
    print("\(logClassName) Converting sound view in rect \(drawingRect)")

    guard let context:CGContext = UIGraphicsGetCurrentContext() else{ return }

    context.setFillColor(waveViewInfo.backgroundColor.cgColor)
    context.setAlpha(1.0)
    context.fill(drawingRect)
    context.setLineWidth(1.0)
    //        context.setLineWidth(waveViewInfo.lineWidth)

    let sampleAdjustFactor = imageSize.height / waveViewInfo.maxValue
    for pointIndex in waveViewInfo.points.indices{

        let pixel = waveViewInfo.points[pointIndex] * sampleAdjustFactor

        context.move(to: CGPoint(x: CGFloat(pointIndex), y: middleY - pixel))
        context.addLine(to: CGPoint(x: CGFloat(pointIndex), y: middleY + pixel))

        context.setStrokeColor(waveViewInfo.strokeColor.cgColor)
        context.strokePath()

    }

     //        for pointIndex in waveViewInfo.points.indices{
    //
    //            let pixel = waveViewInfo.points[pointIndex] * sampleAdjustFactor
    //
    //            context.move(to: CGPoint(x: CGFloat(pointIndex) * waveViewInfo.sampleSpace, y: middleY - pixel))
    //            context.addLine(to: CGPoint(x: CGFloat(pointIndex) * waveViewInfo.sampleSpace, y: middleY + pixel))
    //
    //            context.setStrokeColor(waveViewInfo.strokeColor.cgColor)
    //            context.strokePath()
    //
    //        }

    //        var xIncrement:CGFloat = 0
    //        for point in waveViewInfo.points{
    //
    //            let normalizedPoint = point * sampleAdjustFactor
    //
    //            context.move(to: CGPoint(x: xIncrement, y: middleY - normalizedPoint))
    //            context.addLine(to: CGPoint(x: xIncrement, y: middleX + normalizedPoint))
    //            context.setStrokeColor(waveViewInfo.strokeColor.cgColor)
    //            context.strokePath()
    //
    //            xIncrement += waveViewInfo.sampleSpace
    //
    //        }

    guard let soundWaveImage = UIGraphicsGetImageFromCurrentImageContext() else{ return }

    soundWaveImageView.image = soundWaveImage
    //        //In case of handling sample space in for
    //        updateWidthConstraintValue(soundWaveImage.size.width)
    updateWidthConstraintValue(soundWaveImage.size.width * waveViewInfo.sampleSpace)

}

哪里

class WaveViewInfo {

    var format:AVAudioFormat!
    var samplesSeconds:CGFloat = 50
    var lineWidth:CGFloat = 0.20
    var sampleSpace:CGFloat = 0.20

    var strokeColor:UIColor = .red
    var backgroundColor:UIColor = .clear

    var maxValue:CGFloat = 0
    var points:[CGFloat] = [CGFloat]()

}

目前只打印一个声波，但它可以扩展。好处是你可以按部分打印音轨。

Answer 3

根据上面的答案进行一点重构(使用AVAudioFile)

import AVFoundation
import CoreGraphics
import Foundation
import UIKit

class WaveGenerator {
    private func readBuffer(_ audioUrl: URL) -> UnsafeBufferPointer<Float> {
        let file = try! AVAudioFile(forReading: audioUrl)

        let audioFormat = file.processingFormat
        let audioFrameCount = UInt32(file.length)
        guard let buffer = AVAudioPCMBuffer(pcmFormat: audioFormat, frameCapacity: audioFrameCount)
        else { return UnsafeBufferPointer<Float>(_empty: ()) }
        do {
            try file.read(into: buffer)
        } catch {
            print(error)
        }

//        let floatArray = Array(UnsafeBufferPointer(start: buffer.floatChannelData![0], count: Int(buffer.frameLength)))
        let floatArray = UnsafeBufferPointer(start: buffer.floatChannelData![0], count: Int(buffer.frameLength))

        return floatArray
    }

    private func generateWaveImage(
        _ samples: UnsafeBufferPointer<Float>,
        _ imageSize: CGSize,
        _ strokeColor: UIColor,
        _ backgroundColor: UIColor
    ) -> UIImage? {
        let drawingRect = CGRect(origin: .zero, size: imageSize)

        UIGraphicsBeginImageContextWithOptions(imageSize, false, 0)

        let middleY = imageSize.height / 2

        guard let context: CGContext = UIGraphicsGetCurrentContext() else { return nil }

        context.setFillColor(backgroundColor.cgColor)
        context.setAlpha(1.0)
        context.fill(drawingRect)
        context.setLineWidth(0.25)

        let max: CGFloat = CGFloat(samples.max() ?? 0)
        let heightNormalizationFactor = imageSize.height / max / 2
        let widthNormalizationFactor = imageSize.width / CGFloat(samples.count)
        for index in 0 ..< samples.count {
            let pixel = CGFloat(samples[index]) * heightNormalizationFactor

            let x = CGFloat(index) * widthNormalizationFactor

            context.move(to: CGPoint(x: x, y: middleY - pixel))
            context.addLine(to: CGPoint(x: x, y: middleY + pixel))

            context.setStrokeColor(strokeColor.cgColor)
            context.strokePath()
        }
        guard let soundWaveImage = UIGraphicsGetImageFromCurrentImageContext() else { return nil }

        UIGraphicsEndImageContext()
        return soundWaveImage
    }

    func generateWaveImage(from audioUrl: URL, in imageSize: CGSize) -> UIImage? {
        let samples = readBuffer(audioUrl)
        let img = generateWaveImage(samples, imageSize, UIColor.blue, UIColor.white)
        return img
    }
}

使用

let url = Bundle.main.url(forResource: "TEST1.mp3", withExtension: "")!
let img = waveGenerator.generateWaveImage(from: url, in: CGSize(width: 600, height: 200))