newtonoid
Junior Member level 3
Hello, edaboard.
I created this "sound recognition doohickey" yesterday in simulink. It computes a correlation coefficient between the spectral estimation of audio being picked up by my laptop microphone and the spectral estimation of stored audio in attempt to "listen" for a specific melody. I currently have the algorithm sending a trigger to an arduino to flash an LED. It's not extremely computationally efficient or optimized, but I just threw it together for fun.
I'm thinking about improving the algorithm a bit and trying to embed it into a beaglebone or something. What do you guys think?
I created this "sound recognition doohickey" yesterday in simulink. It computes a correlation coefficient between the spectral estimation of audio being picked up by my laptop microphone and the spectral estimation of stored audio in attempt to "listen" for a specific melody. I currently have the algorithm sending a trigger to an arduino to flash an LED. It's not extremely computationally efficient or optimized, but I just threw it together for fun.
I'm thinking about improving the algorithm a bit and trying to embed it into a beaglebone or something. What do you guys think?
Last edited: