Half-Way Progress
This is a plot of the average group delay of each of our audio files after the denoising filtering has been applied. It uses MATLAB's Filter Visualization Tool to show exactly at how many samples and at what normalized frequency the file has been delayed by. This was an important piece to our denoising filter because if we did not find the appropriate delay, we would not have been able to fix our file to be able to use it with our moving average filter (coming soon!).
This is a plot of the de-noising result of our program for the "Noisybark.wav" file. In green is the unfiltered signal, and the blue is the filtered signal once the unwanted noise was removed from the file. We accomplished this by creating a filter that first read in the original .wav file, converted it to the frequency domain, and used MATLAB's Filter Designer app to create the appropriate filter. A delay was then created when the filter was applied, and we added zero padding to remove the delay.
This is a plot of the de-noising result of our program for the "Dog.wav" file. In green is the unfiltered signal, and in blue is the filtered signal once the unwanted noise was removed from the file. We wanted to use this file as well as the previous one because we wanted to make sure our de-noising filter worked with multiple iterations of barks and we can see from the plots that it did. We accomplished this by creating a filter that first read in the original .wav file, converted it to the frequency domain, and used MATLAB's Filter Designer app to create the appropriate filter. A delay was then created when the filter was applied, and we added zero padding to remove the delay.
Half-Way Updates
So far, we have collected samples of various animal noises (including dog, cat, and bird noises). We found a link to extract the animal sounds from their original clips and then worked through de-noising several dog samples in MATLAB. We have also been conducting research on creating a Moving Average Filter using MATLAB, as well as attempted to create a Pitch Recognition file, but have not yet been able to finalize it and utilize it for our sound clips. We are hoping that it will take in a longer sound file and only detect the moments in time which contain a specified frequency range and will give us better fidelity on when to cut off the signal.
What We Learned
We discovered the Filter Visualization Tool and Filter Design app to be especially helpful in constructing our final signal. They became extremely useful as we moved on to moving average filtering. Also, we implemented de-noising to our signals so that when we ran our program it was able to better match signals and help create our final product.
We discovered the Filter Visualization Tool and Filter Design app to be especially helpful in constructing our final signal. They became extremely useful as we moved on to moving average filtering. Also, we implemented de-noising to our signals so that when we ran our program it was able to better match signals and help create our final product.