Geek Answers Handbook

How to make raw speech in a text converter?

I have severe deaf deafness from a very young age, but, fortunately, I can speak like an ordinary person. Verbal communication has always been difficult for me because of my weakened speech recognition abilities even when reading lips. I went to school and college just by reading boards, Powerpoint slides, books, and the Internet. I am very good at my current software development work, but lately I feel that I have to make some efforts to improve the situation.

Subtitles are my lifeguard in this country to understand films / shows on TV, and I only enjoyed this for the last 7 years (now I'm 31).

I really feel the need for the ability to see subtitles in real life when I talk with someone, even strangers. I want to develop unprepared speech in a text converter, and as a start, he does not even need to specify exact words for me, only the melodies on the syllables / phonetics will also be in order.

I have been looking for this for a while, but most of the results are text or semi-successful speech recognition attempts to give voice commands to the computer. I would really like to receive some guidance on how to start this project. In particular, I need steps such as, for example, how to process audio files and what processing should I do to get approximate phonetics as quickly as possible.

+6
algorithm speech-to-text phonetics
source share
3 answers

You might want to see the CMU Sphinx project , which makes speech in text in real time. They have demos to try.

+3
source share

Look at the DSP guide , it's more about low-level materials, but methods like Fourier transforms and filtering are important for audio processing. Even if you are not starting from scratch, it may be useful to evaluate principles and applications.

Nevertheless, I am sure that starting from scratch, you can create something that can distinguish the main set of sounds with the work of several days ...

+1
source share

Here are some more questions that may give you ideas:

  • Automatically convert WMA / MP3 audio?
  • How to convert text to speech?

And take a look at SIL Linguistics Computing .

Good luck.

+1
source share

More articles:


All Articles