How to implement voice recognition in .NET?
I've got a bunch of .mp3s with computer generated voice of relatively high quality. In each .mp3 the voice says some random text and then a sequence of 6-7 digits in English. I'd like to extract those digits as text.
Is it possible to implement such a thing in .NET?
You can use Microsoft Speech to text Engines.. Its easy to use
Some samples here
You can use Microsoft SAPI SDK:
http://www.microsoft.com/downloads/details.aspx?FamilyID=5e86ec97-40a7-453f-b0ee-6583171b4530&DisplayLang=en
Check out the C# examples in SAPI SDK.
Here's a link to the System.Speech.Recognition namespace that may help you out.
When working with Speech Recognition the best thing is to work in Windows Vista or Windows 7 cause they have Microsoft SAPI (Speech API) 5.3 (and will soon have 5.4) and that gives a big diference in performance comparing with SAPI 5.1 from Windows XP.
Now, here you have a video and source code (it's in c++ but you can pass it to another language) of an application that almost do what you want.
To recognize words in a better way, they make a grammar file with the words that they want to recognize. You can see that video and you will see that will be quite easy to make what you want.
Good luck!