Get excitation from voiced and unvoiced signal

I have two signals, one whisper data and one normal voiced data of the same sentence.

I'm trying to build a voice conversion, but just want to start off with some simple signal manipulations.

I wanted to get the excitation of several vowels first to apply the excitation to the whispered data; basically to get a Electrolarynx kinda robotic sound. Therefore, I manually splitted the signal into smaller wav-files with only vowels. So now I have, e.g. an 'i' signal in whispered data and normal voiced data.

How can I get the excitation now? These are some buzz words I heard concerning this topic:

• calculating the fft of both signals
• inverse fft

I'm pretty clueless here, so any help/tip is appreciated.