Dec-ification continues at its snail pace. I have successful clones that are quickly produced and are impressive in their own right, but they don't have the ranting, raving, throaty baritone quality of the radio DJ's narration - they're much too sober sounding, as the models have been trained on normal, sane conversations.
At the moment it just reminds me of Patrick Tomlinson.
Francification.wav
Ver archivo adjunto 8749591
So I'm running an arduous finetuning process to try and capture that Francis schizo-zeal.
Ver archivo adjunto 8749469
I've decided it's probably prudent
not to allude to the specs or process I went through to reach this point. I don't want to risk making actual harassment any easier by posting cliffnotes.
Really I'm just doing this to learn how to synth voices for my media projects. It's a fun edge case.