Original upload date: Tue, 02 Jan 2018 00:00:00 GMT
Archive date: Sun, 07 Nov 2021 17:41:33 GMT
Sampled sound signals are stored to memory in a bulk mic-recording process. Each time-domain sound signal is automatically converted into the frequency-domain using Fourier Transform algorithm.
Then
...
, user is prompted to record a test signal for recognition (I'm "user" in the video above). Once the test signal is obtained, it will be compared with each previously sampled signal through a max-amplitude normalization (helps align the key harmonic feautures) and standard deviation measurement routine. The compared signal that scores the lowest on the deviation accumulator wins (kind of like golf)!
My programs are entirely my own content.
My programs are written in MATLAB and communicate with EXCEL files.
This specific program does use MATLAB's FFT (for speed), however I have written my own fully accurate FFT algorithm that could be interchanged into the program. Otherwise my programs rarely rely on any of MATLAB's built-in signal processing functions.
Any of my programs can be obtained for free from me through private email. Message me, privately. See all my projects at pitt.edu/~rdb37
