之前的項目中要用到把用戶輸入的文字轉換為語音文件再播放出來，當時也是一頭霧水沒搞明白怎么弄，查了好多資料有的是使用在線合成語音文件，這個就有局限性了，不能離線合成，所以就放棄了。后來發現訊飛是有離線語音合成包了，但是官網沒有C#調用的離線包，windows平台只有C++寫的包，這個就比較坑了。后來就想到用C#調用C++dll的方式看能不能實現，參考了訊飛官網和論壇中大神們寫的代碼，最終還是實現了C#文字轉換語音的功能。現在把代碼貼出來，小伙們們可以直接使用，不過是要去訊飛官網購買離線包的。
public class iFlyTTS
{
/// <summary>
/// 引入TTSDll函數的類
/// </summary>
private class TTSDll
{
#region TTS dll import
[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)]
public static extern int MSPLogin(string one, string two, string configs);
[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)]
public static extern void MSPLogout();
[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)]
public static extern IntPtr QTTSSessionBegin(string _params, ref int errorCode);
[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)]
public static extern int QTTSTextPut(string sessionID, string textString, uint textLen, string _params);
[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)]
public static extern IntPtr QTTSAudioGet(string sessionID, ref int audioLen, ref SynthStatus synthStatus, ref int errorCode);
[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)]
public static extern IntPtr QTTSAudioInfo(string sessionID);
[DllImport("msc.dll", CallingConvention = CallingConvention.StdCall)]
public static extern int QTTSSessionEnd(string sessionID, string hints);
[DllImport("msc.dll", CallingConvention = CallingConvention.Cdecl)]
public static extern int QTTSGetParam(string sessionID, string paramName, string paramValue, ref uint valueLen);
#endregion
}
private string sessionID;
public iFlyTTS(string configs)
{
int ret = TTSDll.MSPLogin(null, null, configs);
if (ret != 0) throw new Exception("初始化TTS引擎錯誤，錯誤代碼：" + ret);
}
public void MultiSpeek(string SpeekText, string outWaveFlie = null)
{
MemoryStream mStream = new MemoryStream();
try
{
speek(SpeekText, ref mStream);
WAVE_Header header = getWave_Header((int)mStream.Length - 44); //創建wav文件頭
byte[] headerByte = StructToBytes(header); //把文件頭結構轉化為字節數組 //寫入文件頭
mStream.Position = 0; //定位到文件頭
mStream.Write(headerByte, 0, headerByte.Length); //寫入文件頭
if (outWaveFlie != null)
{
FileStream ofs = new FileStream(outWaveFlie, FileMode.Create);
mStream.WriteTo(ofs);
ofs.Close();
ofs = null;
}
}
catch (Exception ex)
{
}
finally
{
TTSDll.MSPLogout();
mStream.Close();
mStream = null;
}
}
/// <summary>
/// 把文本轉換成聲音，寫入指定的內存流
/// </summary>
/// <param name="SpeekText">要轉化成語音的文字</param>
/// <param name="mStream">合成結果輸出的音頻流</param>
private void speek(string SpeekText, ref MemoryStream mStream)
{
string szParams = "engine_type = local, voice_name = xiaoyan, text_encoding = GB2312, tts_res_path = fo|res\\tts\\xiaoyan.jet;fo|res\\tts\\common.jet, sample_rate = 8000,aue=speex-wb;7, speed = 50, volume = 50, pitch = 50, rdn = 2";
int ret = 0;
try
{
sessionID = Ptr2Str(TTSDll.QTTSSessionBegin(szParams, ref ret));
if (ret != 0) throw new Exception("初始化TTS引會話錯誤，錯誤代碼：" + ret);
ret = TTSDll.QTTSTextPut(sessionID, SpeekText, (uint)Encoding.Default.GetByteCount(SpeekText), string.Empty);
if (ret != 0) throw new Exception("向服務器發送數據，錯誤代碼：" + ret);
int audio_len = 0;
SynthStatus synth_status = SynthStatus.TTS_FLAG_STILL_HAVE_DATA;
MemoryStream fs = mStream;
IntPtr audio_data;// Marshal.AllocHGlobal(1024);
Thread.Sleep(1000);
while (synth_status != SynthStatus.TTS_FLAG_DATA_END)
{
audio_data = TTSDll.QTTSAudioGet(sessionID, ref audio_len, ref synth_status, ref ret); // sdh 這個是會花費時間的
byte[] tmpArray = Encoding.Default.GetBytes(SpeekText);
if (ret != 0) break;
byte[] data = new byte[audio_len];
if (audio_len > 0) Marshal.Copy(audio_data, data, 0, audio_len);
fs.Write(data, 0, data.Length);
}
}
catch (Exception ex)
{
}
finally
{
ret = TTSDll.QTTSSessionEnd(sessionID, "");
if (ret != 0) throw new Exception("結束TTS會話錯誤，錯誤代碼：" + ret);
}
}
private struct WAVE_Header
{
public int RIFF_ID; //4 byte , 'RIFF'
public int File_Size; //4 byte , 文件長度
public int RIFF_Type; //4 byte , 'WAVE'
public int FMT_ID; //4 byte , 'fmt'
public int FMT_Size; //4 byte , 數值為16或18，18則最后又附加信息
public short FMT_Tag; //2 byte , 編碼方式，一般為0x0001
public ushort FMT_Channel; //2 byte , 聲道數目，1--單聲道；2--雙聲道
public int FMT_SamplesPerSec;//4 byte , 采樣頻率
public int AvgBytesPerSec; //4 byte , 每秒所需字節數,記錄每秒的數據量
public ushort BlockAlign; //2 byte , 數據塊對齊單位(每個采樣需要的字節數)
public ushort BitsPerSample; //2 byte , 每個采樣需要的bit數
public int DATA_ID; //4 byte , 'data'
public int DATA_Size; //4 byte ,
}
/// <summary>
/// 根據數據段的長度，生產文件頭
/// </summary>
/// <param name="data_len">音頻數據長度</param>
/// <returns>返回wav文件頭結構體</returns>
WAVE_Header getWave_Header(int data_len)
{
WAVE_Header wav_Header = new WAVE_Header();
wav_Header.RIFF_ID = 0x46464952; //字符RIFF
wav_Header.File_Size = data_len + 36;
wav_Header.RIFF_Type = 0x45564157; //字符WAVE
wav_Header.FMT_ID = 0x20746D66; //字符fmt
wav_Header.FMT_Size = 16;
wav_Header.FMT_Tag = 0x0001;
wav_Header.FMT_Channel = 1; //單聲道
wav_Header.FMT_SamplesPerSec = 8000; //采樣頻率
wav_Header.AvgBytesPerSec = 16000; //每秒所需字節數
wav_Header.BlockAlign = 2; //每個采樣1個字節
wav_Header.BitsPerSample = 16; //每個采樣8bit
wav_Header.DATA_ID = 0x61746164; //字符data
wav_Header.DATA_Size = data_len;
return wav_Header;
}
/// <summary>
/// 把結構體轉化為字節序列
/// </summary>
/// <param name="structure">被轉化的結構體</param>
/// <returns>返回字節序列</returns>
Byte[] StructToBytes(Object structure)
{
Int32 size = Marshal.SizeOf(structure);
IntPtr buffer = Marshal.AllocHGlobal(size);
try
{
Marshal.StructureToPtr(structure, buffer, false);
Byte[] bytes = new Byte[size];
Marshal.Copy(buffer, bytes, 0, size);
return bytes;
}
finally
{
Marshal.FreeHGlobal(buffer);
}
}
/// <summary>
/// 指針轉字符串
/// </summary>
/// <param name="p">指向非托管代碼字符串的指針</param>
/// <returns>返回指針指向的字符串</returns>
public static string Ptr2Str(IntPtr p)
{
List<byte> lb = new List<byte>();
while (Marshal.ReadByte(p) != 0)
{
lb.Add(Marshal.ReadByte(p));
p = p + 1;
}
byte[] bs = lb.ToArray();
return Encoding.Default.GetString(lb.ToArray());
}
}
使用的時候直接調用下面的方法就行，如果要改變音色或者播放速度的都可以配置的，這個小伙伴們自己研究下，我調試了好幾種感覺就現在的挺好。
private void 文字轉語音(String sourcewav) { try { //這個appid 是指在訊飛官網購買離線包的appid string login_params = "appid = ******, work_dir = ."; iFlyTTS tts = new iFlyTTS(login_params); //填寫上自己要保存的路徑 string strPath = System.IO.Path.Combine(***, sourcewav + ".wav"); tts.MultiSpeek(sourcewav.Trim(), strPath); } catch (Exception e) { } }