I've been coding for about 10 years now and it just got to my mind that I would like to learn about how I can create speech recognition software. I am mostly interested in doing this from ground 0 not because there are no other libs I can use, but it's more of an interest to me.
So starting this topic to ask for help and directions what steps I should take in order to grasp this new and magnificent feature that's becoming more and more used nowadays.
My preference would be coding in C++ (more because of OOP that's why not C) and I don't really like Java.
I know what I'm getting involved in but I am thrilled and willing to go all the way to achieve my goal.
Speech recognition is a pretty complex area. Even professional software doesn't have it 100% right yet.
I'm not sure how it's done myself. Any info I could give you on the subject would be speculation, or stuff I dug up on the internet just now. Mainly, the wikipedia page on it:
As for a starting point, I don't think it would be wise to start out with speech recognition right away. Rather, the reverse (text->audio) exercises many of the same concepts and would be much easier. I would start with making a robot voice that reads some text first as sort of like an introduction to the topics. Once you have that working, then look into audio->text.
Starting with small steps is good. I would like to go first at creating text->audio, but again I don't really know where to start. Any suggestions ? I have read the article about the Speech Recognition Algorithms before coming here to post. It seems quite interesting that's why I want to go on with it.
So any clues where I should start with the text->audio ? Or what steps should I take to create a robot voice?
My experience in audio was from writing retro-style chiptune synthesizers. This subject is totally different and I'd be starting from square 1 on it... so you probably know as much as I do =/