Take a normal voice sample

Pitch it downwards masses, whilst keeping the length of the sample the same

Throw some processing at it (filter VST/PSP Nitro)