用Unity写了一个类似汤姆猫自动检测说话功能

简单的把脚本加到摄像机上就行

下面是我写的脚本，有个问题 因为是自动调用检测的 调用录音unity调用有延时 会出现一些延时小问题，可以参考我的代码 改改 做个通过按钮点击录音结束播放录音还是能实现的

using System.Collections.Generic;
using UnityEngine;
using System.Collections;

public class Tom2 : MonoBehaviour
{
    //public float power = 1;
    //public float min = 0;
    // public AudioClip audioClip = null;//AudioClip音频剪辑
    //public AudioSource audioSource;//AudioSource 音频源
    public bool audioclipWorking = false;
    public string audioIsPlaying;

    public bool microphoneWorking = false;
    private List<float> samplesList; //播放数据存储

    public float samplesMaxValue;
    public int ClipLength;

    private void Start()
    {
        samplesList = new List<float>(); //不固定
        startRecord();
    }

    private void Update()
    {
        audioIsPlaying = audio.isPlaying.ToString();
        Screen.sleepTimeout = SleepTimeout.NeverSleep;
        if (microphoneWorking)
        {
            if (!Microphone.IsRecording(null))
            {
                microphoneWorking = false;
                JudgeRecord();
            }
        }
        if (audioclipWorking)
        {
            if (audio.isPlaying == false)
            {
                samplesList.Clear();
                startRecord();
                audioclipWorking = false;
            }
        }
    }

    private void startRecord()
    {
        audio.clip = Microphone.Start(null, false, 1, 44100);
        while (Microphone.GetPosition(null) <= 0) ;
        microphoneWorking = true;
    }

    public void JudgeRecord()
    {
        float max = 0;
        float[] tempSamples; //临时数据存储
        tempSamples = new float[audio.clip.samples * audio.clip.channels];
        audio.clip.GetData(tempSamples, 0);
        foreach (float s in tempSamples)
        {
            float m = Mathf.Abs(s);
            if (max < m)
            {
                max = m; //刚刚录制音频数据的取最大值
            }
        }
        samplesMaxValue = max;
        if (max > 0.1)
        {
            //判断有人说话
            startRecord();
            foreach (float e in tempSamples)
            {
                samplesList.Add(e);//保存数据
            }
            //再录下一秒
        }
        else
        {

            //无人说话
            if (audio.clip != null && samplesList != null)
            {
                if (samplesList.Count > 1000)
                {
                    AudioClip myClip = AudioClip.Create("tom", samplesList.Count, 1, 44100, false, false);
                    myClip.SetData(samplesList.ToArray(), 0);
                    audio.clip = myClip;
                    audio.loop = false;
                    audio.pitch = 1.2f;
                    audio.Play();
                    audioclipWorking = true;
                }
                else
                {
                    startRecord();
                }
            }
            else
            {
                startRecord();
            }
        }
    }

    private void OnGUI()
    {
        GUI.Label(new Rect(60, 30 * 1, Screen.width, 20), "samplesMaxValue:" + samplesMaxValue);
        GUI.Label(new Rect(60, 30 * 2, Screen.width, 20), "ClipLength:" + ClipLength);
        GUI.Label(new Rect(60, 30 * 3, Screen.width, 20), "audioIsPlaying:" + audioIsPlaying);
        GUI.Label(new Rect(60, 30 * 4, Screen.width, 20), "audioclipWorking:" + audioclipWorking);
        GUI.Label(new Rect(60, 30 * 5, Screen.width, 20), "microphoneWorking:" + microphoneWorking);
        GUI.Label(new Rect(60, 30 * 6, Screen.width, 20), "audio.clip.samples:" + audio.clip.samples);
        GUI.Label(new Rect(60, 30 * 7, Screen.width, 20), "audio.clip.length:" + audio.clip.length);
        GUI.Label(new Rect(60, 30 * 8, Screen.width, 20), "audio.clip.channels:" + audio.clip.channels);
        GUI.Label(new Rect(60, 30 * 8, Screen.width, 20), "Microphone.GetPosition:" + Microphone.GetPosition(null));
    }
}