Azaran on 24/3/2021 at 00:46
Here's an idea. Using audio deepfake technology to clone Garrett's voice.
There's a few paid apps out there, but (
https://github.com/CorentinJ/Real-Time-Voice-Cloning) this is supposed to be a free one where you can upload voice clips and clone a voice based on just a few seconds of audio. Imagine the possibilities....
SlyFoxx on 24/3/2021 at 02:54
Well that's me out of a job.
bbb on 24/3/2021 at 03:24
Cool idea but I am more of a traditionalist. A friend does a lot of home recording and takes his voice and uses a device that uses samples to change the voice. He did a recording and converted the two voices to Barry White and Tinkerbell. Neat effect but not the same as a real voice.
I would be interested to see how it works but if I need a Garrett voice, I am calling you Sly! Don't retire yet. You actually did a Garret voice for one of my missions and it worked really well.
bbb
A.Stahl on 24/3/2021 at 04:58
So try it.
There was some modern mission where I notice how clean and easy to understand the voices were (I almost don't know English so I can't just parse any mumbling). Later, in the FM's thread, the author told that voices were machine-generated.
Dr.Sahnebacke on 24/3/2021 at 17:41
I would also prefer a really actor with all his melody in his voice and timbre.
BazookaBear on 28/3/2021 at 15:37
I tried setting up the code/program your referring to but I failed miserably. But I did find this-https://www.youtube.com/watch?v=I-tpsOSy5l4, Which is the same program just a lot more user friendly(but maybe limited?). I attempted to clone Garrett's voice using some voice lines (even combining some) and every result was unusable. I think there are a few reasons why this idea won't work(or at least right out of the box). First, The software/programs are usually created using audio books (People talking, Not acting). Second, Even if you were able to find a more "audio book/conversational" recording of Stephen Russell, The result would be very "Robotic/Audio book-like". In other words, The ideal result would sound like Stephen Russell ordering a coffee and not Garrett.
bjack on 28/3/2021 at 23:42
SlyFoxx, over the years of playing FMs, I have come to hear your voice as the real Garrett. You do a great impression, yet are your own talent. I’m so used to hearing your voice that the original seems almost fake.
As a VO person myself, I dislike these clone software tactics. It kills nuance. I was a sap a few years ago and agreed to be part of a “charitable” VO gift giving mission. I was supposed to read a few lines to “help” disabled people communicate... However, the recipients of my work took my voice and cloned it for their own purposes, w/o compensation, yet they would charge their clients. F’ed up.
Synth speech is always obvious. It always sucks. Yes, it is getting better, but it still sucks.
downwinder on 29/3/2021 at 03:40
i am curious what it would sound like,is anyone going to make a sample audio of garretts voice ?
request = have garrett say a small story about dewdrop as he never said that word then we can see if the program is any good
BazookaBear on 29/3/2021 at 15:39
Quote Posted by downwinder
i am curious what it would sound like,is anyone going to make a sample audio of garretts voice ?
request = have garrett say a small story about dewdrop as he never said that word then we can see if the program is any good
And then have SlyFoxx record the same story and which ever one brings us to tears the fastest wins.