Empeyot
Newbie level 1
Hi!
Being new myself to DSP and after doing some preliminary googling, I'd like to kindly ask the more experienced members for an estimation of the feasability of the following project and brain-storming for terminology to start help me get started with further investigations.
Background:
A friend is physically severly handicapped and no longer able to control his computer with hands or similarly. Yet he has quite decent control over his voice. However, it's too fuzzy and distorted for of-the-market voice recognition plus due to his artificial breathing machine he is situated in a somewhat noisy environment.
Project description:
My idea is to set-up a real-time voice control that uses well-defined pitch and a better morse-like code (might sound like humming in the end).
Apart from band-width filtering, I thought of using two microfones with known positions relative to sound and noice sources in order to harness spacial information for further signal control.
Questions:
1. To those more experienced in DSP, does this _generally_ sound like a feasible project or is it wishful thinking (if so, what is the challenge I did not see?)?
2. In my preliminary web search I stumbled across a somewhat related problem (cocktail party problem), that seems to have given birth to a lot of algorithms.
Is there a name for the isolation/filtering of a sound signal with multiple microfones with known relative position to the source - how would you call it in proper DSP terminology, so as to uncover already existing algorithms?
Thanks for following me thus far, I hope I have been clear enough and am looking forward to every comments, questions and recommendation of related literature! :smile:
Being new myself to DSP and after doing some preliminary googling, I'd like to kindly ask the more experienced members for an estimation of the feasability of the following project and brain-storming for terminology to start help me get started with further investigations.
Background:
A friend is physically severly handicapped and no longer able to control his computer with hands or similarly. Yet he has quite decent control over his voice. However, it's too fuzzy and distorted for of-the-market voice recognition plus due to his artificial breathing machine he is situated in a somewhat noisy environment.
Project description:
My idea is to set-up a real-time voice control that uses well-defined pitch and a better morse-like code (might sound like humming in the end).
Apart from band-width filtering, I thought of using two microfones with known positions relative to sound and noice sources in order to harness spacial information for further signal control.
Questions:
1. To those more experienced in DSP, does this _generally_ sound like a feasible project or is it wishful thinking (if so, what is the challenge I did not see?)?
2. In my preliminary web search I stumbled across a somewhat related problem (cocktail party problem), that seems to have given birth to a lot of algorithms.
Is there a name for the isolation/filtering of a sound signal with multiple microfones with known relative position to the source - how would you call it in proper DSP terminology, so as to uncover already existing algorithms?
Thanks for following me thus far, I hope I have been clear enough and am looking forward to every comments, questions and recommendation of related literature! :smile: