快速检测说话答案

【问题标题】：Detecting Talking with swift快速检测说话
【发布时间】：2016-07-04 20:13:52
【问题描述】：

我有一个应用需要检测用户何时与该应用交谈。我不想拥有任何第三方 API。我使用了this turtorial，但我必须将其转换为 swift 2。这是我正在使用的代码：

import UIKit
import AVFoundation
import CoreAudio

class ViewController: UIViewController, AVAudioPlayerDelegate, AVAudioRecorderDelegate {

var audioRecorder: AVAudioRecorder?

var timer = NSTimer()

@IBAction func start(sender: AnyObject) {
    if audioRecorder?.recording == false {
        audioRecorder?.record()
        timer = NSTimer.scheduledTimerWithTimeInterval(1, target: self, selector: #selector(ViewController.checkForAudio), userInfo: nil, repeats: true)
    }
}


@IBAction func stop(sender: AnyObject) {
    if audioRecorder?.recording == true {
        audioRecorder?.stop()
        timer.invalidate()
    }
}


func checkForAudio() {
    audioRecorder?.updateMeters()

    print("Average: \(audioRecorder?.averagePowerForChannel(0)) Peak: \(audioRecorder?.peakPowerForChannel(0))")
}

override func viewDidLoad() {
    super.viewDidLoad()
    let fileMgr = NSFileManager.defaultManager()
    let dirPaths = fileMgr.URLsForDirectory(.DocumentationDirectory, inDomains: .UserDomainMask)
    let soundFileURL = dirPaths[0].URLByAppendingPathComponent("sound.caf")

    let recordSettings = [AVSampleRateKey: 44100.0, AVNumberOfChannelsKey: 2, AVEncoderBitRateKey: 12800, AVLinearPCMBitDepthKey: 16, AVEncoderAudioQualityKey: AVAudioQuality.Max.rawValue]

    let audioSession = AVAudioSession.sharedInstance()

    do {
        try audioSession.setCategory(AVAudioSessionCategoryPlayAndRecord)

        } catch let error as NSError {
            print("audioSession error: \(error.localizedDescription)")
        }

    do {
        try audioRecorder = AVAudioRecorder(URL: soundFileURL, settings: recordSettings as! [String : AnyObject])
        audioRecorder?.prepareToRecord()
        audioRecorder?.meteringEnabled = true
    } catch let error as NSError {
        print("audioSession error: \(error.localizedDescription)")
    }

}

override func didReceiveMemoryWarning() {
    super.didReceiveMemoryWarning()
    // Dispose of any resources that can be recreated.
}

class func getDocumentsDirectory() -> String {
    let paths = NSSearchPathForDirectoriesInDomains(.DocumentDirectory, .UserDomainMask, true)
    let documentsDirectory = paths[0]
    return documentsDirectory
}

}

然后在日志中它只是打印这个，我发出响亮的声音和故事，但它仍然没有改变。

Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)
Average: Optional(-120.0) Peak: Optional(-120.0)

【问题讨论】：

如果您只能针对 iOS 10，您可以考虑研究 Apple 的新语音识别 API； developer.apple.com/videos/play/wwdc2016/509
问题是它需要互联网连接。

标签： ios swift swift2 avaudiorecorder avaudiosession

【解决方案1】：

我试过你的代码，你的文件网址在某种程度上是错误的，我将文件网址更改为临时路径：

let pathStr = NSTemporaryDirectory().stringByAppendingString("sound.caf")
    let soundFileURL = NSURL(fileURLWithPath: pathStr)

它可以工作

Average: Optional(-27.2326279) Peak: Optional(-13.5865393)
Average: Optional(-33.1891823) Peak: Optional(-23.3389206)
Average: Optional(-49.0175018) Peak: Optional(-22.505867)
Average: Optional(-38.9896851) Peak: Optional(-18.3879299)
Average: Optional(-33.0929604) Peak: Optional(-20.2775478)

【讨论】：