visionOS平台AVCaptureAudioDataOutput替代方案及AVAudioBuffer获取咨询

• 基于录制回调直接捕获
  既然已确认音频录制功能可用，可通过AVAudioRecorder结合自定义逻辑捕获AVAudioBuffer。先配置好音频会话和录制格式，再借助底层回调或引擎tap获取缓冲区：

  let session = AVAudioSession.sharedInstance()
  do {
      try session.setCategory(.record, mode: .default)
      try session.setActive(true)
      
      let settings: [String: Any] = [
          AVFormatIDKey: kAudioFormatLinearPCM,
          AVSampleRateKey: 44100.0,
          AVNumberOfChannelsKey: 1,
          AVLinearPCMBitDepthKey: 16,
          AVLinearPCMIsFloatKey: false,
          AVLinearPCMIsBigEndianKey: false
      ]
      
      let recorder = try AVAudioRecorder(url: temporaryFileURL, settings: settings)
      recorder.prepareToRecord()
      // 后续可结合AudioQueue或AVAudioEngine的tap逻辑获取buffer
  } catch {
      print("配置失败: \(error)")
  }
  

• 尝试AVAudioEngine输入节点Tap
  尽管visionOS初期对AVAudioEngine支持有限，但输入节点tap是获取实时音频缓冲区的常规方案，可尝试如下实现：

  let engine = AVAudioEngine()
  let inputNode = engine.inputNode
  let inputFormat = inputNode.inputFormat(forBus: 0)
  
  inputNode.installTap(onBus: 0, bufferSize: 1024, format: inputFormat) { (buffer: AVAudioPCMBuffer, time: AVAudioTime) in
      // 此处直接拿到AVAudioPCMBuffer，可直接处理
      print("捕获缓冲区帧数: \(buffer.frameLength)")
  }
  
  do {
      try engine.start()
  } catch {
      print("引擎启动失败: \(error)")
  }
  

  注意：必须在Info.plist中添加NSMicrophoneUsageDescription字段申请麦克风权限，否则会无响应或崩溃。

• 底层Core Audio API兜底
  如果高层API无法生效，可直接用Core Audio的AudioQueue录制并包装成AVAudioBuffer，兼容性更强：

  var audioFormat = AudioStreamBasicDescription()
  audioFormat.mSampleRate = 44100.0
  audioFormat.mFormatID = kAudioFormatLinearPCM
  audioFormat.mFormatFlags = kAudioFormatFlagIsSignedInteger | kAudioFormatFlagIsPacked
  audioFormat.mChannelsPerFrame = 1
  audioFormat.mBitsPerChannel = 16
  audioFormat.mBytesPerPacket = 2
  audioFormat.mFramesPerPacket = 1
  audioFormat.mBytesPerFrame = 2
  
  func setupAudioQueue() {
      var queue: AudioQueueRef?
      AudioQueueNewInput(&audioFormat, audioQueueCallback, nil, nil, nil, 0, &queue)
      AudioQueueStart(queue!, nil)
  }
  
  func audioQueueCallback(queue: AudioQueueRef, buffer: AudioQueueBufferRef, startTime: UnsafePointer<AudioTimeStamp>, numPackets: UInt32, packetDescs: UnsafePointer<AudioStreamPacketDescription>?) {
      // 将AudioQueueBuffer数据拷贝到AVAudioPCMBuffer
      guard let pcmFormat = AVAudioFormat(streamDescription: &audioFormat) else { return }
      let pcmBuffer = AVAudioPCMBuffer(pcmFormat: pcmFormat, frameCapacity: AVAudioFrameCount(numPackets))
      pcmBuffer?.frameLength = AVAudioFrameCount(numPackets)
      memcpy(pcmBuffer?.int16ChannelData?[0], buffer.pointee.mAudioData, Int(buffer.pointee.mAudioDataBytesCapacity))
      
      // 处理pcmBuffer
      AudioQueueEnqueueBuffer(queue, buffer, 0, nil)
  }
  

• 检查会话配置与权限
  所有音频操作的前提是正确配置会话并获取权限：

  let session = AVAudioSession.sharedInstance()
  do {
      try session.setCategory(.playAndRecord, options: .mixWithOthers)
      try session.setActive(true)
      let hasPermission = await session.requestRecordPermission()
      guard hasPermission else {
          print("未获取麦克风权限")
          return
      }
  } catch {
      print("会话配置失败: \(error)")
  }
  

内容的提问来源于stack exchange，提问作者Pietro Messineo