Convert ANY sentence to text with windows speech
Currently I have a C# program that takes speech using the MS Windows Speech Recognition (SR) System, converts it to text, sends the question as text to Worlfram Alpha, and then reads the text-based answer back using text-to-speech (TTS).
Unfortunately the computer can only recognize a list of predefined sentences and phrases that are defined in grammars that are input into the SR. I need the program to recognize any sentence without having to define an enormous grammar file containing every word in the English language. How can I have the program recognize any English phrase without having to define a grammar file that contains the rules for the complete English language? Is this possible?
There are two basic types of speech recognition systems today; one type supports the use of grammars that list the rules for expected recognition results, and the other uses statistical modeling (ex: Hidden Markov Model) to get recognition results based on probability of matching a set of data that was used to "train" the system. What you are describing as "a list of predefined sentences and phrases" are the grammars. I am pretty certain that MS Windows Speech Recognition System only supports a grammar-based solution. What you are describing as your desired solution is a statistical modeling system for SR. You may want to take a look at the Sphinx project. It is open source and it supports statistical modeling. Beware that setting up a statistical modeling system is not easy and takes a lot of time to get it tuned correctly.
Check this out - SpeechSynthetizer Class and also this related stack overflow question How to write an application that reads text and produces speech
For Speech to Text look at this codeproject article http://www.codeproject.com/Articles/380027/Csharp-Speech-to-Text
链接地址: http://www.djcxy.com/p/34362.html下一篇: 用Windows语音将任意句子转换为文本