SUE-NO

Well, Suno’s getting sued from just about every angle. For those of us who use it, it’s certainly getting interesting.

Today I read about a singer who’s suing them for the method in which they trained their models, after discerning that his actual voice and what he feels is his unique style was used without his consent.

So I’ve been thinking about this.

It doesn’t change anything, but I can only think of one time when we asked for a type of voice that was more than just an extremely vague prompt (male baritone on a country bed with a specified sub-genre, that sort of thing)—it was when we were writing about the Elvis impersonator I used to work for, and I don’t know what prompts Chad used for the vocalist in that case but…I don’t believe that to be Elvis Presley’s actual voice. (Is it? I’m not an Elvis connoisseur.) It sounds more like an impersonator to me, not quite as warm as Elvis himself.

We didn’t even get the sound we wanted until we added “a hubba hubba hey” to the lyrics. And what’s funny about that is where it came from: we were once seated on a flight with a guy who wouldn’t stop singing that at me once he found out I was a musician. We always thought it sounded like he was imitating Elvis, so I looked it up.

My search turned up zilch for Elvis Presley ever singing that phrase. But that’s how we got that style.

Other than that, I honestly couldn’t identify any of the “vocalists” we’ve used. We used one that sounded something like the absolutely incredible Jonah Nilsson style-wise, and that was a complete accident; Metal Head Chad doesn’t listen to Dirty Loops—that’s my thing, and I never asked for it.

If we want soul, we prompt for soul. If we want jazz, we prompt for jazz. And also, in our case, we don’t present as artists. We’re making demos of our own material to showcase it to actual musicians, and that’s our entire intent. I know there are a lot of people that use Suno to try to make money posing as performing artists, and I can absolutely see why that’s so aggravating for popular musicians.

That is something I would never do. I’m not ever going to be anything but up front about what we do, and why and how we do it.

Now, Chad does all the prompting (the only thing I ever do is ask if it’s possible to prompt a certain way…trying to adjust the number of bars, tweak the intensity, manipulate the phrasing, change the style or architecture of the song, change the voice, etc.) so he may be able to shed some light on what he’s putting together across the table, but my understanding is that Suno restricts the use of any artist or band’s name in their prompts.

But the other thing I’ve been pondering is: I don’t really sound like anyone, but I will tell you I used to work with a great singer who sounds almost indistinguishable from the lead singer of the band Styx. He’s so good at singing these tunes, I keep telling him if Styx (they’re still together, but without Dennis DeYoung) ever hears him and is between vocalists, there’s no way he doesn’t get that gig. And I’m not kidding.

So who did he train on? Well…Dennis DeYoung, for one.

I worked with another great singer who sounds almost indistinguishable from James Ingram. Same thing.

Everybody trained on somebody. Almost anyone who’s ever played in a cover band or a tribute band was trying to sound like someone else. We’ve all copied styles. Every one of us. And we’ve all gotten paid to do it.

And it never bothered anybody that I know.

So Suno uses someone’s voice and style as a model, but restricts people from actually re-creating their work (I’m told it will throw a copyright warning if a user actually tries to plagiarize somebody). If my Dennis DeYoung-sounding friend writes a song that sounds like something Styx might have done and sings it himself, what’s the material difference between that and a machine doing it?

The debate is pretty interesting. And I can see both sides. I just wonder what the differences between humans copying and machines copying really are.

Thea