Brian Wilson AI Model Release (so-vits-svc 4.0)

I appreciate the little "Let's Put Our Hearts Together" demo: it definitely helps put into perspective what this model both is and isn't capable of doing. (For what it's worth, Mike, your own singing voice probably runs rounds around mine.)

And thus, the floodgates have opened. Not 100% sure I'll be catching the resulting wave myself, but either way, I'm looking forward to seeing what folks use this for. Remember, fellas, this might not be an Al A.I., but let's "keep it clean" all the same!

Can't get the BAT installer to do anything, everytime I click run it does something just disappears forever.

Aw crap, I missed the release. Will download it tomorrow though, thanks a lot, Mike!

I didnt not get how you can get the trained model

Yes!!! Can't wait to have a go with this model. Thanks so much again for uploading!!
P.S., Looking forward to the stereo isolations going up! I keep checking back in the MEGA folder lol

This is neat!

cool! but it sounds to dirty! when you infer it maybe the hifi-gan decoder is destroying the signal . i trained some models with only 5000 steps and they sound much clearer. i will try your model.

Thank you so much! Does so-vits github repo also work on a MAC, or is this Windows only?

For the life of me, I can't download it from MEGA. The site just keeps saying that I have to pay to extend something! Can't you upload the vocals somewhere else like Drive or any other platform?

This is amazing! I'd love to do all his songs from his solo album and newer all with his young voice. Especially "Let it Shine". Do you know if this is possible on a mac? Also, do you know if there is a Jeff Lynne model or how to get it or request it?

How do you create an ai model? I'm trying to do this with my own voice, would be awesome to have me singing covers

Dude I have a podcast called Hold Onto The Colours, I have had many well-known people on including Van Dyke Parks, Ron Carter, Mike Linett and more. I would love to have you on to discuss the Smile project and your Brian AI model. Cheers. Paul

The endless possibilities... Imagine any 70s Brian songs sang with Pet Sounds Brian! The mashups are gonna be insane!

I’ve been training a Gwen Stefani model and it sounds nothing like her. Well it’s sounds like vocaloid version of Gwen 😅

do you have a zip on this?

btw, the one you are using is 3.0 or 4.0?? i installed it like u do but when run svcg it said version 3.11

not sure how to download this on mac

took me forever to get it downloaded and working on my Mac but totally worth it. Thanks!

Gift from God 😧 because of you I have learned a lot about this technology. Thank you!

nice! how do you make your own model tho? could you give us tutorials on that?

do you know if there is a Neil Young model??

This is great however do we have necessary permissions from 'Brian Wilson' to use this?

amazing thanks mate .. are their any other models of other singers available to download anywhere?

Can someone do Tune X (Boundless Love) Jacob Borshard?? That would be so sweet.

Admire your ethical stipulations, but remember this is the internet: we can’t have nice things.

I took one of my own songs, stripped the vocal out to use as an input but followed your steps otherwise. I got a blank output with some little noises here or there. I also noticed I couldn't use the GPU and preset you chose as it wouldn't do anything. Are you on an M1?

I want to pay someone to train a custom model for me 🙏

Yo i have a full stydio with ssl 4k etc, lemme know how/ if i can help? Looking for somebody interested in bci and also ai w ssl g cpu etc. Hmu

Hi Mike!
where can i download the voice of liam gallagher AI model?

The voice we know from The Beach Boys wasn’t Brian’s but Carl…..anyway, there should be a model to compose and produce like Brian…..the best producer in history who produced the best album in history……

Hi I'm seeking somebody VERY skilled with these models. I want to have our friend sing some songs for her own funeral. Sadly, she just died at 34. She was a big Alan Parsons fan, anybody available for hire? 1st Song: alan parsons: Old and wise.

Open-minded. Hmu. Have many recordings of her speeking . Her funeral is in 1 month. Thanks!

autopredict activates automatically when i go infer??? what can i do about it i turn it off and then go infer and it turns on :(((

Promo SM 🌟

Thanks for your tutorial!
Some models that I've found have a pth file but no json file (they seem to work on Colab but not on local. Is there something I can do?
Also, igot (not so good) results with a Bowie model that I found, but the BWilson only gives clics & pops!

How can I download the vocals for my Mac computer?

You're take was fine as it was you ruined it with that autotune 😢

I am a 31 year old audio engineer (started at age 13) and I've been using reaper for over 10 years. but when you said MIXCRAFT my jaw dropped. I didn't even know that was still going... That was my first DAW. Anyways, I'm a huge BB fanatic and this is awesome. thanks mike

I've got to hear that on the original song.

I'm curious as to what you trained this on, I input some extremely basic 'oooh' vox and it spit out 'ahohhah' that aren't in any way usable. oh well, thanks for the proof of concept.

I downloaded it, but when I try to open it, it closes after half a second.