Voice recognition? How about song or whislte recognition?
Already have an account? Log in.
To make the experience fit your profile, pick a username and tell us what interests you.
Submitted to the Hackaday.io "Human Interface" competition.
Very much in a formative stage at this point, although it is pretty certain this will be implemented on a RaspberryPI 3 and will probably utilize one of the cloud-based deep learning systems.
Agreed that an FFT approach should be a good starting point.
https://www.youtube.com/watch?v=7SmG0SMAeqU
Sounds fun! There should be a way to implement this on RPI without the cloud. Whistling must produce some pure tones. FFT that audio and you don't need no stinking deep learning. Something like that could be written in a day or two using FFTW. https://en.wikipedia.org/wiki/FFTW Phone systems were accepting 2600 Hz from a Captain Crunch whistle like 60 years ago :) https://en.wikipedia.org/wiki/2600_hertz
Charlie Lindahl
Johanna Shi
Greg Kennedy
Lutetium
Valentin
This project is pretty far down my list at this point, but I may revive it for next year's competition. Thanks for commenting.