I\'m performing some text-to-speech and I\'d like to specify some special pronunciations in a lexicon file. I have ran MSDN\'s AddLexicon example verbatim, and it speaks the
You can use System.Speech.Synthesis.SpeechSynthesizer.SpeakSsml() instead of a lexicon.
This code changes pronunciation of "blue" to "yellow" and "dog" to "fish".
SpeechSynthesizer synth = new SpeechSynthesizer();
string text = "This is a blue dog";
Dictionary<string, string> phonemeDictionary = new Dictionary<string, string> { { "blue", "jelow" }, { "dog", "fyʃ" } };
foreach (var element in phonemeDictionary)
{
text = text.Replace(element.Key, "<phoneme ph=\"" + element.Value + "\">" + element.Key + "</phoneme>");
}
text = "<speak version=\"1.0\" xmlns=\"http://www.w3.org/2001/10/synthesis\" xml:lang=\"en-US\">" + text + "</speak>";
synth.SpeakSsml(text);
After a lot of research and pitfalls I can assure you that your assumption is just plain wrong.
For some reason System.Speech.Synthesis.SpeechSynthesizer.AddLexicon() adds the lexicon to an internal list, but doesn't use it at all.
Seems like nobody tried using it before and this bug went unnoticed.
Microsoft.Speech.Synthesis.SpeechSynthesizer.AddLexicon() (which belongs to the Microsoft Speech SDK) on the other hand works as expected (it passes the lexicon on to the COM object which interprets it as advertised).
Please refer to this guide on how to install the SDK: http://msdn.microsoft.com/en-us/library/hh362873%28v=office.14%29.aspx
Notes:
I've been looking into this a little recently on Windows 10.
There are two things I've discovered with System.Speech.Synthesis.
Any Voice you use, must be matched against the language in the Lexicon file. Inside the lexicon you have the language:
<lexicon version="1.0"
xmlns="http://www.w3.org/2005/01/pronunciation-lexicon"
alphabet="x-microsoft-ups" xml:lang="en-US">
I find that I can name my Lexicon as "blue.en-US.pls" and make a copy with "blue.en-GB.pls". Inside it will have xml:lang="en-GB"
In the code you'd use:
string langFile = Path.Combine(_appPath, $"blue.{synth.Voice.Culture.IetfLanguageTag}.pls");
synth.AddLexicon(new Uri(langFile), "application/pls+xml");
The other thing I discovered is, it doesn't work with "Microsoft Zira Desktop - English (United States)" at all. I don't know why. This appears to be the default voice on Windows 10.
Access and change your default voice here: %windir%\system32\speech\SpeechUX\SAPI.cpl
Otherwise you should be able to set it via code:
var voices = synth.GetInstalledVoices();
// US: David, Zira. UK: Hazel.
var voice = voices.First(v => v.VoiceInfo.Name.Contains("David"));
synth.SelectVoice(voice.VoiceInfo.Name);
I have David (United States) and Hazel (United Kingdom), and it works fine with either of those.
This appears to be directly related to whether the voice token in the registry has the SpLexicon key value. The Microsoft Zira Desktop voice does not have this registry value.
While Microsoft David Desktop voice has the following:
Computer\HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Speech\Voices\Tokens\TTS_MS_EN-US_DAVID_11.0\Attributes\SpLexicon = {0655E396-25D0-11D3-9C26-00C04F8EF87C}