The sister site has free software called Replay allowing voices from a database of famous artists to be used in our music.
This is a track that uses "Freddie Mercury" and while it may not be exactly the famed Queen singer, neither am I.

 

I've added another using an Elton John model. It's passable if you close your eyes.

 

very interestink! I vill try it soon.

 

Professionally crafted beautiful songs.
But if I had to assign a percentage on voice similarity I would assign 40% to Freddie Mercury and 60% to Elton John, and believe me I have a steel memory for sounds and voices.

 

I appreciate the compliment about the songs. I began my career as a songwriter and I place the crafting of songs above all. 47 years now and I still write a song a week. I'm not getting better, but I'm not getting worse.

 

Based on the description, I was expecting something completely different in terms of overall quality. This is quite impressive! Not referring to the similarity to the original voices (don't care about that too much). It definitely makes me think about how this could be used as an "effect" (to smooth, alter timbre, etc..). Definitely tempted to try.

Did you try any of those paid services for comparison?

Again, nice work!

 

May i ask where you can downoad Elton's model?

 

That sounds great!

 

this Ai crap is the end of human music. no talent needed.

 

does anyone have help on how to get the trainer installed locally?

 

Here's a song with Lana Del Rey. She annoys me, so I'm not sure if I'm happy with this. But it suited the song.

I'm sorry if I'm not much help with how to use this. My method is best suited for me. I have my own songs with clean vocal stems, so I'm not using the app to it's full potential and as such I'm not sure of how to use it to isolate vocals....which I think it can do, but I'm not sure.

 

On the far left is a column labelled "Categories" go to "musician" and in the column called "Filter Models" keep scrolling down. All available are listed alphabetically.

 

....and finally, this is supposedly Lady Gaga.... i say finally because the novelty is beginning to wear off.

 

I think your music style lacks something important. Very very important... Do you know what it is?

Listen to this music and compare yours with it and tell me what it is...

 

Youthful optimism? A sense of joy? I lost those back in the early 80's.

 

haha funny as hell !!!! you think these singers of today are talented there just processed to the hell bro.

get most of these talented people you talk and talk of in a live venue singing live and not mimeing to there studio backing track they sound absolutely shitful live anyway

you may as well of said make yourself sound the same as the no talents who think they are talented

as this is a much cheaper and timeless option to fix there as you say talented shit voices haha

 

Last edited: Nov 12, 2023

I've been experimenting with Replay for a couple of days, but I haven't been able to come up with anything interesting. I believe AI still has a long way to go to become what they say it should be.

 

Now go away, or I shall taunt you a second time-a

 

haha funny as hell

 

If I listened to these with no information about how they were done, I would probably guess they were done by a human singer. Just not the singers the models are supposed to be of. For my own purposes, I think I would prefer that anyway. I do not see much point in adding a "featuring" of a famous artist just to drum up some more clicks and listens; if it isn't them. If someone wants a track to ride on someone's famous coat tails; a new version/remix with real parts would seem like it would get more mileage. I am more interested in this kind of technology to replace average vocalists-for-hire from the freelancer sites, or to even just reduce the amount of auditioning time it takes using commercially available sample packs. I'd probably be trying to make them sound less accurate, not more.

That Miley Cyrus vocal has a lot of "human imperfections" to it, smokey changes to lower register, raspiness; that is going to take time for the technology to progress more. Most of the plugin developers are still telling us how "analog" they sound, and we are only trying to mimic a compressor or something.