GenshinTextToSpeech: Unity 音频合成插件

GenshinTextToSpeech: Unity 音频合成插件

GenshinTextToSpeech 是一款基于 Unity 的音频合成插件，支持使用原神角色的声音进行文本朗读。该插件使用 RESTful API 进行语音合成，并提供多种音频设置选项，例如声音角色、音频格式、长度、噪声等。

代码示例

using System;
using System.Collections;
using System.Xml.Linq;
using UnityEngine;
using UnityEngine.Networking;

public class GenshinTextToSpeech : TTS
{

    #region 参数定义
    /// <summary>
    /// Genshin 配置项
    /// </summary>
    [SerializeField] private GenshinSettings m_GenshinSettings;
    /// <summary>
    /// 朗读的角色
    /// </summary>
    [Header("朗读声音设置")]
    public string speaker = '胡桃';
    /// <summary>
    /// 音频格式
    /// </summary>
    [Header("音频格式设置")]
    public string format = 'wav';
    /// <summary>
    /// 音频长度
    /// </summary>
    [Header("音频长度设置")]
    public float length = 1f;
    /// <summary>
    /// 噪声
    /// </summary>
    [Header("噪声设置")]
    public float noise = 0.5f;
    /// <summary>
    /// 噪声权重
    /// </summary>
    [Header("噪声权重设置")]
    public float noisew = 0.9f;
    /// <summary>
    /// 声调比例
    /// </summary>
    [Header("声调比例设置")]
    public float sdp_ratio = 0.2f;

    #endregion

    private void Awake()
    {
        // m_GenshinSettings = this.GetComponent<GenshinSettings>(); // 移除该行代码
        m_PostURL = string.Format("https://{0}/api", m_GenshinSettings.apiDomain);
    }

    /// <summary>
    /// 语音合成
    /// </summary>
    /// <param name="_msg"></param>
    /// <param name="_callback"></param>
    public override void Speak(string _msg, Action<AudioClip> _callback)
    {
        StartCoroutine(GetVoice(_msg, _callback));
    }

    /// <summary>
    /// 语音合成，返回合成文本
    /// </summary>
    /// <param name="_msg"></param>
    /// <param name="_callback"></param>
    public override void Speak(string _msg, Action<AudioClip, string> _callback)
    {
        StartCoroutine(GetVoice(_msg, _callback));
    }

    /// <summary>
    /// restful api 语音合成
    /// </summary>
    /// <param name="_msg"></param>
    /// <param name="_callback"></param>
    /// <returns></returns>
    private IEnumerator GetVoice(string _msg, Action<AudioClip> _callback)
    {
        stopwatch.Restart();
        // 发送报文
        string textToSpeechRequestBody = GenerateTextToSpeech(speaker, _msg, format, length, noise, noisew, sdp_ratio);

        using (UnityWebRequest speechRequest = new UnityWebRequest(m_PostURL, "POST"))
        {
            byte[] data = System.Text.Encoding.UTF8.GetBytes(textToSpeechRequestBody);
            speechRequest.uploadHandler = (UploadHandler)new UploadHandlerRaw(data);
            speechRequest.downloadHandler = (DownloadHandler)new DownloadHandlerAudioClip(speechRequest.uri, AudioType.WAV);

            yield return speechRequest.SendWebRequest();

            if (speechRequest.responseCode == 200)
            {
                AudioClip audioClip = DownloadHandlerAudioClip.GetContent(speechRequest);
                _callback(audioClip);
            }
            else
            {
                Debug.LogError("语音合成失败: " + speechRequest.error);
            }
        }

        stopwatch.Stop();
        Debug.Log("Genshin 语音合成耗时：" + stopwatch.Elapsed.TotalSeconds);
    }

    /// <summary>
    ///  restful api 语音合成，返回合成文本
    /// </summary>
    /// <param name="_msg"></param>
    /// <param name="_callback"></param>
    /// <returns></returns>
    private IEnumerator GetVoice(string _msg, Action<AudioClip, string> _callback)
    {
        stopwatch.Restart();
        // 发送报文
        string textToSpeechRequestBody = GenerateTextToSpeech(speaker, _msg, format, length, noise, noisew, sdp_ratio);

        using (UnityWebRequest speechRequest = new UnityWebRequest(m_PostURL, "POST"))
        {
            byte[] data = System.Text.Encoding.UTF8.GetBytes(textToSpeechRequestBody);
            speechRequest.uploadHandler = (UploadHandler)new UploadHandlerRaw(data);
            speechRequest.downloadHandler = (DownloadHandler)new DownloadHandlerAudioClip(speechRequest.uri, AudioType.WAV);

            yield return speechRequest.SendWebRequest();

            if (speechRequest.responseCode == 200)
            {
                AudioClip audioClip = DownloadHandlerAudioClip.GetContent(speechRequest);
                _callback(audioClip, _msg);
            }
            else
            {
                Debug.LogError("语音合成失败: " + speechRequest.error);
            }
        }

        stopwatch.Stop();
        Debug.Log("Genshin 语音合成耗时：" + stopwatch.Elapsed.TotalSeconds);
    }

    /// <summary>
    /// 报文格式转换
    /// </summary>
    /// <param name="speaker"></param>
    /// <param name="msg"></param>
    /// <param name="format"></param>
    /// <param name="length"></param>
    /// <param name="noise"></param>
    /// <param name="noisew"></param>
    /// <param name="sdp_ratio"></param>
    /// <returns></returns>
    public string GenerateTextToSpeech(string speaker, string msg, string format, float length, float noise, float noisew, float sdp_ratio)
    {
        string urlParams = string.Format("?speaker={0}&msg={1}&format={2}&length={3}&noise={4}&noisew={5}&sdp_ratio={6}", speaker, msg, format, length, noise, noisew, sdp_ratio);
        string url = m_PostURL + urlParams;

        return url;
    }
}

// GenshinSettings 类不需要修改，因为在调用 GenerateTextToSpeech 方法时，你已经将 m_GenshinSettings.apiDomain 传递给了 urlParams。

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

[CreateAssetMenu(fileName = "GenshinSettings", menuName = "ScriptableObjects/GenshinSettings")]
public class GenshinSettings : ScriptableObject
{
    public string apiDomain;
}

### 错误原因及解决方案

错误消息 “ArgumentException: GetComponent requires that the requested component 'GenshinSettings' derives from MonoBehaviour or Component or is an interface.”  表示你尝试使用 GetComponent 获取一个  GenshinSettings 组件，但  GenshinSettings  不是一个 MonoBehaviour 或 Component，而是一个 ScriptableObject。

解决方法：

1.  移除  `m_GenshinSettings = this.GetComponent<GenshinSettings>();`  代码行。
2.  在你的场景中创建一个  GenshinSettings  实例，并将其赋值给  `m_GenshinSettings`  变量。

这样，你就可以在不使用 GetComponent 的情况下使用  GenshinSettings  中的 apiDomain 属性。

### 总结

GenshinTextToSpeech 是一款功能强大的 Unity 音频合成插件，支持使用原神角色的声音进行文本朗读。通过修改代码并设置  GenshinSettings  ，你可以轻松地使用该插件合成各种音频，并在你的 Unity 项目中添加个性化的声音效果。