The last post here (Speech (Audio) to TEXT Query.. tool) has reminded me to let everyone know that if you've only got a small amount of audio to recognise then you can do so easily and accurately at www.speechmatics.com.  The site allows you to upload very many audio formats and it'll email you back the results or if you have more audio there is a simple command line tool or API available.

By no means do I want to discourage people from building thier own ASR systems - I think that's a wonderful learning experiance and the world needs more people who can do this.   But if you only have an hour then we'll transcribe it for free and it'll only take a few minutes to get results.   And if you need more time because you are going to build an open source corpus or are working on an academic publication then let me know - the world needs more of these as well.

Tony

-- 

Dr Tony Robinson
Founder Cantab Research Ltd

--- (Edited on 8-December-2014 3:12 pm [GMT+0000] by TonyR) ---