/// <summary>
/// 转换语音
/// </summary>
/// <param name="text">合成语语音所用文本</param>
/// <param name="filename">"输出音频文件名"</param>
/// <param name="voicename">声音人物选择</param>
/// <param name="voicespeed">语速</param>
public void ProcessVoice(string text, string filename, VoiceName voicename = VoiceName.xiaoyan, int voicespeed = 5)
{
string param = string.Format(" vcn={0}, spd ={1}, vol = 50, bgs=0, aue=speex-wb, smk = 3", voicename, voicespeed);
string login_param = " appid = 58d376fc,work_dir = . ";
Begin_ProcessVoice(text, filename, param, login_param);
}
/// <summary>
/// Generates SSML.
/// </summary>
/// <param name="locale">The locale.</param>
/// <param name="gender">The gender.</param>
/// <param name="name">The voice name.</param>
/// <param name="text">The text input.</param>
private string GenerateSsml(string locale, string gender, VoiceName voicename, string text, int pitchdelta)
{
string voice = ConvertVoiceNametoString(voicename);
XNamespace xmlns = "http://www.w3.org/2001/10/synthesis";
var ssmlDoc = new XDocument(
new XElement(xmlns + "speak",
new XAttribute("version", "1.0"),
new XAttribute(XNamespace.Xml + "lang", locale), // was locked to "en-US"
new XElement("voice",
new XAttribute(XNamespace.Xml + "lang", locale),
new XAttribute(XNamespace.Xml + "gender", gender),
new XAttribute("name", voice),
new XElement("prosody",
new XAttribute("pitch", pitchdelta.ToString() + "Hz"),
text))));
//var ssmlDoc = new XDocument(
// new XElement(xmlns + "speak",
// new XAttribute("version", "1.0"),
// new XAttribute(XNamespace.Xml + "lang", locale), // was locked to "en-US"
// new XElement("voice",
// new XAttribute(XNamespace.Xml + "lang", locale),
// new XAttribute(XNamespace.Xml + "gender", gender),
// new XAttribute("name", voice),
// text)));
return(ssmlDoc.ToString());
}
public virtual Stream TextToSpeech(string text, SpeechLocaleOptions locale, VoiceName voiceName, GenderOptions voiceType, AudioOutputFormatOptions outputFormat)
{
return(PolicyService.ExecuteRetryAndCapture400Errors(
"SpeechService.TextToSpeech",
ApiKeys.SpeechRetryInSeconds,
() =>
{
var result = SpeechRepository.TextToSpeech(text, locale, voiceName, voiceType, outputFormat);
return result;
},
null));
}
public virtual Stream TextToSpeech(string text, SpeechLocaleOptions locale, VoiceName voiceName, GenderOptions voiceType, AudioOutputFormatOptions outputFormat)
{
var response = RepositoryClient.GetAudioStream(
ApiKeys.SpeechEndpoint,
text,
locale,
voiceName,
voiceType,
outputFormat,
GetSpeechToken());
return(response);
}
/// <summary>
/// Generates SSML.
/// </summary>
/// <param name="locale">The locale.</param>
/// <param name="gender">The gender.</param>
/// <param name="name">The voice name.</param>
/// <param name="text">The text input.</param>
private string GenerateSsml(string locale, string gender, VoiceName voicename, string text)
{
string voice = ConvertVoiceNametoString(voicename);
var ssmlDoc = new XDocument(
new XElement("speak",
new XAttribute("version", "1.0"),
new XAttribute(XNamespace.Xml + "lang", locale), // was locked to "en-US"
new XElement("voice",
new XAttribute(XNamespace.Xml + "lang", locale),
new XAttribute(XNamespace.Xml + "gender", gender),
new XAttribute("name", voice),
text)));
return(ssmlDoc.ToString());
}
public string GetVoiceLocale(VoiceName voicename)
{
return(ConvertVoiceNametoString(voicename).Substring(46, 5));
}
/// <summary>
/// Converts a specific VoioceName enum option into its string counterpart as expected
/// by the API when building the SSML string that is sent to Cognitive Services.
/// Make sure that each option in the enum is included in the switch below.
/// </summary>
/// <param name="voicename"></param>
/// <returns></returns>
public string ConvertVoiceNametoString(VoiceName voicename)
{
switch (voicename)
{
case VoiceName.enAUCatherine:
return("Microsoft Server Speech Text to Speech Voice (en-AU, Catherine)");
case VoiceName.enAUHayleyRUS:
return("Microsoft Server Speech Text to Speech Voice (en-AU, HayleyRUS)");
case VoiceName.enCALinda:
return("Microsoft Server Speech Text to Speech Voice (en-CA, Linda)");
case VoiceName.enCAHeatherRUS:
return("Microsoft Server Speech Text to Speech Voice (en-CA, HeatherRUS)");
case VoiceName.enGBSusanApollo:
return("Microsoft Server Speech Text to Speech Voice (en-GB, Susan, Apollo)");
case VoiceName.enGBHazelRUS:
return("Microsoft Server Speech Text to Speech Voice (en-GB, HazelRUS)");
case VoiceName.enGBGeorgeApollo:
return("Microsoft Server Speech Text to Speech Voice (en-GB, George, Apollo)");
case VoiceName.enIESean:
return("Microsoft Server Speech Text to Speech Voice (en-IE, Sean)");
case VoiceName.enINHeeraApollo:
return("Microsoft Server Speech Text to Speech Voice (en-IN, Heera, Apollo)");
case VoiceName.enINPriyaRUS:
return("Microsoft Server Speech Text to Speech Voice (en-IN, PriyaRUS)");
case VoiceName.enINRaviApollo:
return("Microsoft Server Speech Text to Speech Voice (en-IN, Ravi, Apollo)");
case VoiceName.enUSZiraRUS:
return("Microsoft Server Speech Text to Speech Voice (en-US, ZiraRUS)");
case VoiceName.enUSJessaRUS:
return("Microsoft Server Speech Text to Speech Voice (en-US, JessaRUS)");
case VoiceName.enUSBenjaminRUS:
return("Microsoft Server Speech Text to Speech Voice (en-US, BenjaminRUS)");
case VoiceName.deATMichael:
return("Microsoft Server Speech Text to Speech Voice (de-AT, Michael)");
case VoiceName.deCHKarsten:
return("Microsoft Server Speech Text to Speech Voice (de-CH, Karsten)");
case VoiceName.deDEHedda:
return("Microsoft Server Speech Text to Speech Voice (de-DE, Hedda)");
case VoiceName.deDEHeddaRUS:
return("Microsoft Server Speech Text to Speech Voice (de-DE, HeddaRUS)");
case VoiceName.deDEStefanApollo:
return("Microsoft Server Speech Text to Speech Voice (de-DE, Stefan, Apollo)");
case VoiceName.esESHelenaRUS:
return("Microsoft Server Speech Text to Speech Voice (es-ES, HelenaRUS)");
case VoiceName.esESLauraApollo:
return("Microsoft Server Speech Text to Speech Voice (es-ES, Laura, Apollo)");
case VoiceName.esESPabloApollo:
return("Microsoft Server Speech Text to Speech Voice (es-ES, Pablo, Apollo)");
case VoiceName.esMXHildaRUS:
return("Microsoft Server Speech Text to Speech Voice (es-MX, HildaRUS)");
case VoiceName.esMXRaulApollo:
return("Microsoft Server Speech Text to Speech Voice (es-MX, Raul, Apollo)");
case VoiceName.frCACaroline:
return("Microsoft Server Speech Text to Speech Voice (fr-CA, Caroline)");
case VoiceName.frCAHarmonieRUS:
return("Microsoft Server Speech Text to Speech Voice (fr-CA, HarmonieRUS)");
case VoiceName.frCHGuillaume:
return("Microsoft Server Speech Text to Speech Voice (fr-CH, Guillaume)");
case VoiceName.frFRJulieApollo:
return("Microsoft Server Speech Text to Speech Voice (fr-FR, Julie, Apollo)");
case VoiceName.frFRHortenseRUS:
return("Microsoft Server Speech Text to Speech Voice (fr-FR, HortenseRUS)");
default:
return("Microsoft Server Speech Text to Speech Voice (en-US, JessaRUS)");
}
}
public virtual async Task <Stream> GetAudioStreamAsync(string url, string text, SpeechLocaleOptions locale, VoiceName voiceName, GenderOptions voiceType, AudioOutputFormatOptions outputFormat, string token)
{
HttpClientHandler handler = new HttpClientHandler()
{
CookieContainer = new CookieContainer(), UseProxy = false
};
HttpClient client = new HttpClient(handler);
client.DefaultRequestHeaders.Clear();
var Headers = GetAudioHeaders(token, outputFormat);
foreach (var header in Headers)
{
client.DefaultRequestHeaders.TryAddWithoutValidation(header.Key, header.Value);
}
var request = new HttpRequestMessage(HttpMethod.Post, url)
{
Content = new StringContent(GetAudioContent(text, locale, voiceName, voiceType))
};
var response = await client.SendAsync(request, HttpCompletionOption.ResponseHeadersRead, CancellationToken.None);
if (!response.IsSuccessStatusCode)
{
throw new Exception($"({response.StatusCode}) {response.ReasonPhrase}");
}
return(await response.Content.ReadAsStreamAsync());
}
public virtual Stream GetAudioStream(string url, string text, SpeechLocaleOptions locale, VoiceName voiceName, GenderOptions voiceType, AudioOutputFormatOptions outputFormat, string token)
{
// Construct HTTP request to get the logo
HttpWebRequest httpRequest = (HttpWebRequest)
WebRequest.Create(url);
httpRequest.Method = WebRequestMethods.Http.Post;
var headers = GetAudioHeaders(token, outputFormat);
httpRequest.Headers.Add("X-Microsoft-OutputFormat", outputFormat.MemberAttrValue());
httpRequest.Headers.Add("Authorization", $"Bearer {token}");
httpRequest.ContentType = "application/ssml+xml";
httpRequest.UserAgent = "TTSClient";
byte[] reqData = Encoding.UTF8.GetBytes(GetAudioContent(text, locale, voiceName, voiceType));
httpRequest.ContentLength = (long)reqData.Length;
Stream requestStream = httpRequest.GetRequestStream();
requestStream.Write(reqData, 0, reqData.Length);
requestStream.Close();
// Get back the HTTP response for web server
HttpWebResponse httpResponse = (HttpWebResponse)httpRequest.GetResponse();
Stream httpResponseStream = httpResponse.GetResponseStream();
var newStream = new MemoryStream();
httpResponseStream.CopyTo(newStream);
httpResponse.Close();
newStream.Position = 0;
return(newStream);
}
public virtual string GetAudioContent(string text, SpeechLocaleOptions locale, VoiceName voiceName, GenderOptions voiceType)
{
var langValue = locale.MemberAttrValue();
var namespaceAttr = $"xmlns=\"http://www.w3.org/2001/10/synthesis\"";
var versionAttr = $"version=\"1.0\"";
var langAttr = $"xml:lang=\"{langValue}\"";
var genderAttr = $"xmlns:gender=\"{voiceType}\"";
var nameAttr = $"name=\"Microsoft Server Speech Text to Speech Voice {voiceName.MemberAttrValue()}\"";
var speakTag = $"<speak {versionAttr} {namespaceAttr} {langAttr}><voice {nameAttr}>{text}</voice></speak>";
return(speakTag);
}
public virtual bool TextToFile(string text, string filePath, SpeechLocaleOptions locale, VoiceName voiceName, GenderOptions voiceType, AudioOutputFormatOptions outputFormat)
{
if (File.Exists(filePath))
{
return(true);
}
//can only send 1000 characters at a time
//speech xml request = 178 characters
var textLimit = 822;
var textParts = (text.Length > textLimit)
? SplitToLength(text, textLimit)
: new List <string> {
text
};
using (var fileStream = new FileStream(filePath, FileMode.Append, FileAccess.Write, FileShare.Write))
{
foreach (var t in textParts)
{
using (var dataStream = TextToSpeech(t, locale, voiceName, voiceType, outputFormat))
{
if (dataStream != null)
{
dataStream.CopyTo(fileStream);
}
}
}
fileStream.Close();
}
return(true);
}
