Swift中AudioKit 5的PitchTap无响应问题排查求助

问题背景 #

在Swift应用中部署AudioKit 5，曾在iPad真机上实现音高检测与生成功能，代码变更后PitchTap无法输出结果。已将代码精简至基础配置，确认音频会话与引擎配置正常，但PitchTap回调始终未触发。

核心问题分析 #

• 麦克风节点未接入音频流：PitchTap依赖音频数据流过目标节点才能捕获数据，当前代码中麦克风输入节点未连接到音频引擎输出，导致音频流未经过PitchTap监听的节点。
• PitchTap初始化时机错误：PitchTracker初始化时直接获取audioEngine.input，但此时音频引擎尚未完成启动，可能导致麦克风节点获取失败。
• 麦克风权限缺失：iOS平台下未申请麦克风权限，真机环境中麦克风会被禁用，无法采集音频数据。

修复方案 #

1. 添加麦克风权限配置 #

在Info.plist中添加权限描述，确保系统向用户申请麦克风权限：

<key>NSMicrophoneUsageDescription</key>
<string>需要访问麦克风进行音高检测</string>

在AudioEngineManager中添加权限请求逻辑：

func requestMicrophonePermission() {
    AVAudioSession.sharedInstance().requestRecordPermission { granted in
        DispatchQueue.main.async {
            granted ? print("麦克风权限已获取") : print("麦克风权限被拒绝")
        }
    }
}

并在configureAudioSession方法中调用该函数。

2. 确保麦克风节点接入音频流 #

修改PitchTracker的setupMicrophone方法，将麦克风节点连接到音频引擎输出，让音频流经过该节点：

private func setupMicrophone() {
    guard let input = audioEngine.input else {
        print("无法获取麦克风输入节点")
        return
    }
    self.mic = input
    // 连接麦克风到引擎输出，确保音频流通过节点
    audioEngine.output = input
}

3. 延迟PitchTap初始化时机 #

将PitchTap的初始化逻辑移到startTracking方法中，确保音频引擎已启动且麦克风节点可用：

func startTracking() {
    guard let mic = self.mic else {
        print("麦克风节点不可用，无法启动音高检测")
        return
    }
    self.pitchTap = PitchTap(mic) { [weak self] pitch, amp in
        print("PitchTap callback executed.")
        guard let self = self else { return }
        if amp[0] > self.minimumAmplitudeThreshold {
            self.detectedPitch = pitch[0]
            self.detectedAmplitude = amp[0]
            DispatchQueue.main.async {
                self.pitchDetectedHandler?(self.detectedPitch, self.detectedAmplitude)
            }
        }
    }
    self.pitchTap?.start()
    print("Pitch tracking started")
}

同时移除PitchTracker初始化方法中的setupPitchTap()调用。

4. 优化音频会话配置（可选） #

根据需求调整音频会话的选项，若不需要同时播放音频，可移除.defaultToSpeaker以避免路由冲突：

try audioSession.setCategory(.playAndRecord, mode: .measurement, options: [.allowBluetoothA2DP])

完整修复后的核心代码片段 #

修改后的PitchTracker类 #

import Foundation
import AudioKit
import SoundpipeAudioKit

class PitchTracker {
    static let shared = PitchTracker()

    private let audioEngine: AudioEngine = AudioEngineManager.shared.audioEngine
    var mic: AudioEngine.InputNode?
    
    private var oscillator: Oscillator
    private var pitchTap: PitchTap?
    
    let minimumAmplitudeThreshold: Float = 0.01
    var detectedPitch: Float = 0.0
    var detectedAmplitude: Float = 0.0
    
    var pitchDetectedHandler: ((Float, Float) -> Void)?

    init() {
        oscillator = Oscillator()
        setupMicrophone()
    }
    
    private func setupMicrophone() {
        guard let input = audioEngine.input else {
            print("无法获取麦克风输入节点")
            return
        }
        self.mic = input
        audioEngine.output = input
    }
    
    func startTracking() {
        guard let mic = self.mic else {
            print("麦克风节点不可用，无法启动音高检测")
            return
        }
        self.pitchTap = PitchTap(mic) { [weak self] pitch, amp in
            print("PitchTap callback executed.")
            guard let self = self else { return }
            if amp[0] > self.minimumAmplitudeThreshold {
                self.detectedPitch = pitch[0]
                self.detectedAmplitude = amp[0]
                DispatchQueue.main.async {
                    self.pitchDetectedHandler?(self.detectedPitch, self.detectedAmplitude)
                }
            }
        }
        self.pitchTap?.start()
        print("Pitch tracking started")
    }

    func stopTracking() {
        self.pitchTap?.stop()
        self.pitchTap = nil
        print("Pitch tracking stopped")
    }

    func startOscillator(frequency: AUValue = 440) {
        oscillator.frequency = frequency
        oscillator.amplitude = 0.5
        audioEngine.output = oscillator
        oscillator.start()
        print("Oscillator started at frequency: \(frequency)")
    }

    func stopOscillator() {
        oscillator.stop()
        if let mic = self.mic {
            audioEngine.output = mic
        }
        print("Oscillator stopped.")
    }
}

修改后的AudioEngineManager类 #

import AVFoundation
import AudioKit
import SoundpipeAudioKit

class AudioEngineManager {
    static let shared = AudioEngineManager()
    var audioEngine = AudioEngine()
    
    func configureAudioSession() {
        do {
            let audioSession = AVAudioSession.sharedInstance()
            try audioSession.setCategory(.playAndRecord, mode: .measurement, options: [.allowBluetoothA2DP])
            try audioSession.setActive(true)
            print("Audio session configured.")
            requestMicrophonePermission()
        } catch {
            print("Failed to configure audio session: \(error)")
        }
    }
    
    func requestMicrophonePermission() {
        AVAudioSession.sharedInstance().requestRecordPermission { granted in
            DispatchQueue.main.async {
                granted ? print("麦克风权限已获取") : print("麦克风权限被拒绝")
            }
        }
    }

    func startAudioEngine() {
        do {
            try audioEngine.start()
            print("Audio engine started successfully.")
        } catch {
            print("Failed to start audio engine: \(error)")
        }
    }
}

内容的提问来源于stack exchange，提问作者rtardif