Can anybody explain some FORMANT processes to me? (Illustration provided)

[joehuergo]

I am not talking about FORMANT GET or SEE or GETSEE - those I understand.

My problem is trying to wrap my head around the exact differences between REPLACING, IMPOSING and VOCODING.

I have provided illustration for my experiment - through five different sounds. The 'source' sound, which is a strange drone produced by the BLUR process. The 'formant' sound, which is a clip of me speaking clearly. And then, the outputs for imposing, replacing, and vocoding the formant sound onto the source.

The formants were grabbed at the highest linear frequency-wise resolution.
The frequency domain is log scaled.

Source:


Formant:


Impose:


Replace:


Vocode:


Could anyone try explaining these to me? I'm probably just way overthinking it, but I'm quite obsessed with trying to develop a comprehensive, rigorous understanding of these processes.

[joehuergo]

went over the source code with a programmer friend. made a couple guesses but didn't really get anywhere. I don't really think I'll be able to get any definitive answers unless it was coming straight from the designers/programmers. oh well!