I have a file with some speech and some background noise. I used Grab Noise Sample from Selected Area and then Apply Spectral Subtraction Based on Noise Sample to reduce the noise. It does an excellent job of removing most of the noise, but it has a quirk I did not expect: when I chose "Apple Spectral Subtraction," it not only removes the noise, but also changes the length of the file. The file is about 80 seconds long and the length was reduced by 512 frames, or about 12 milliseconds (frame rate is 44100 frames per second), which is awfully small, but enough to cause me problems.
I'd like to hear the noise that was removed, by subtracting the waveform of the denoised file from the waveform of the original file, but this is only possible if there is a one-to-one correspondence between the frames in one file and the frames in the other. Can it be done?