Hey,

What engine would be the best to start with ie Sphinx or Julius. I've seen a presentation where Travis Rudd did it using Dragon Naturally Speaking (DNS) and some duct tape ie Python etc.

https://www.youtube.com/watch?v=8SkdfdXWYaI

Is something similar possible using open source tech or would it involve a ton of work. I've seen a few open source versions using various engines but I've not seen any particular to writing code using emacs/vim.

Note: I don't mind doing a good amount of programming or training to get something working for me ie my own personal acoustic model, recording training data etc. I'm highly motivated to solve this problem for me. I'm very comfortable in C/Java/Perl/Python etc.

Note: If the choice is still to use DNS for this then I'm fine with this also. I'd just like a pointer in the general direction I should take before I do several months worth of work on it. 

--

H

--- (Edited on 10/6/2014 9:04 pm [GMT-0500] by Visitor) ---