Whitehot Magazine
"The Best Art In The World"
Bossa Nova AI generated image.
By ANTHONY HADEN-GUEST January 26, 2025
The project kicked off a few weeks ago while I was staying in Virginia with friends, Elyn Wollensky and her husband, Mark Frieser. In the 90s Frieser had been the singer with a swing band. “We made swing versions of popular hits” he says. He has since created a comfy workspace for himself in the music world, collaborating with movie makers, producers of TV, games, apps and branding ventures, indeed whoever requires a musical element in what they’re doing. In 2012 he launched Sync Summit, a series of meets for such folk at which he makes connections, oversees the music and sorts out the legal quiddities. In 2023 he made a commercial for a Japanese record label, Nippon. It was for music that reflected the inner monologue of a toddler and it won him a Clio, the adworld’s equivalent of an Oscar.
AI entered our conversation, as it tends to nowadays, and I learned that he has been using AI in his work. Which triggered an ask. I write and publish rimes. Here’s a recent one.
Audio link:
https://suno.com/song/0b670394-150b-494f-96b3-37bf01882fb3
Lyrics:
I understand a brand-new virus
Is on the loose in our dimension
The toll it’s taking will soon rise higher
Or so I heard a science dude mention
Not that we’re paying too much attention
Rumtumtum, Tiddlypom,
Let’s all go out and have us some fun
Not sit around waiting for the bomb
Okay, we’ve rather been feeling the heat
Some whine it’ll get a lot hotter than this
So it’s hello. climate? Say goodbye, mate?
Squoosh those liars, truth deniers, starting fires
Don’t let them spoil our meet and greet
Get out of the way of our dancing feet
A cool party is something we just mustn’t miss
It’s dumb just to sit on your bum at home
Glooming, and waiting for the bomb
AI’s already everywhere
It’s in the Earth, the Sea, the Air
It’s taken over the brain, the eye, the ear
It’s gotten into our art, our song, our writing
So it knows how we make stuff. Will it ever know why?
No problem, says Generative AI.
We’ll be on top of that before much time goes by
Now AI’s muscling up to do our fighting
So things might be getting quite exciting
Perhaps you can guess where I’m coming from?
Rumtumtum,Tiddlypom!
We’re not just waiting for the bomb
Could Frieser give this the AI treatment?
Yes, he could. And would. We got down to work in his studio. Music, he told me, can be created by AI by use of three different systems. Studio, Suno and Mureka. He began with Mureka, a system of Chinese origin, his fingers jitterbugging on the pad. “What we are doing now is entering the lyrics of your song,” he explained.
The results were speedy, startling, the wordage being presented over and over in wholly different sonic clothing. It was like watching your portrait being painted in rapid succession by Lucian Freud, Francis Bacon, Frida Kahlo, Richard Hambleton. “Do you want to go back to classical? Or folk? Bluegrass?” Frieser asked after one such bout of listening. “Swing didn’t really work. Did it? What would you like to do? Let’s try this one. What we have here is Disco.”
Mark Frieser at work. Courtesy of the author.
A young female voice belted out I understand a brand-new virus. A finger twitch from Freiser, her voice deepened. “That’s good! A bit more like Chic” he observed.
But a lot of smooshes followed. “Forget Mureka,” Frieser said. “It didn’t do a great job. Didn’t like Mureka. Let’s look at the ones on the playlist in my library. I have a list of the ones that we like.”
He played a remastered version of the Barbershop one. Male voices, really slow. He smooshed it. “No. I don’t like it.” Two versions follow “No, that’s too Afrobeat”.
But along came a goodie. “Amazing”. There was a double trumpet sound. “Omigod! It’s incredible what it did. Let’s extend this one. That is so good that it’s sick. Let’s do a longer version of it. Extend …extend …extend”
Another goodie. What was that? I asked
“That was Dancehall. Now I’m on Studio. When I asked Studio for Bluegrass it did give me something.” Another fingertwitch launched I understand a brand-new virus “It did a pretty good job, right? So I’m going to put a bit of harp on that one.”
We moved on. “I like this a lot. This is the Hank Williams version.” A longer listen. “When we go through these I can extend the length. But right now I want to try to create a bossa nova song out of it. Just bear with me. It may not work. Okay. Let’s see what happens here.”
A female Waiting for the bomb is followed by another breathier female rendition. “Okay. The first one was more traditional. Now I asked it to add a violin. This one is okay. Definitely. I like the first one the best but I like this one too.”
Here’s one with a snappy sound. What’s this?
“It’s Bazzacarda A kind of Brazilian music,” Frieser said. He smooshed it. “No. Forget that. It sucks.” Pause. ”Now here. Let’s see what happens. It’s Colombian salsa. As opposed to Cuban.” Jaunty female voices. “Salsa could work” he decided.
“What about – there’s one more I wanted to do – what about English ska? As opposed to Jamaican ska. Let’s see what it does.” A medley of gruff male voices are instantly smooshed. “It’s definitely not ska.” he says. “Okay, one more. Yacht Rock in the style of Michael McDonald. But that’s a copyright thing right there. If it lets me get away with this I’ll be so happy. But it probably won’t.”
“It” , of course, being AI. Thus the Michael McDonald version got a pass but a Marlene Dietrich got a no no. How about Noel Coward? I suggested, referencing the Brit playwright/performer of nearly a century ago, an acerbically witty songwriter with an unforgettable voice.
“Okay, we’ll do that,” Frieser said. “But let’s see if it does Yacht Rock.” It did, and Waiting for the Bomb got jaunty. “Ohmigod! Omigod!” said Frieser “This SO sounds like Kenny Loggins.”
Next up was a strangulated voice, a guy at death’s door. What was that, I wondered?
“I put in Noel Coward” Frieser said. “So the point is that it’s not perfect is it?” Clearly not. Up came two heavy metal versions which were swiftly smooshed. “I put in Noel Coward again.” Frieser said. “Will it do Noel Coward or some weird metal shit. I don’t know where it’s getting that stuff from. Because those last two were disasters. Just straight up metal.”
Another Coward was created.
“That’s not working,” Frieser said. “Let’s try Sea Shanty.”
This went well. Another goodie with female voices followed. “This one sounds like Abba,” Frieser said joyfully. Another female voice. “Wow! That’s bullshit. Let’s get rid of that one”.
You decide pretty quick, I noted
“This is what I do, “ Frieser said. “ I listen to hundreds of songs a week. I know what I’m listening to within fifteen seconds. It’s my business to know that.”
Many, many hours of such work passed over two days. “We primarily used Suno. It has the best overall tech,” said Mark Frieser, using the “we” as a kindness. “We created 78 versions of the song, using your lyrics.”
He whittled the versions way, way down. Here you will be able to hear The Yacht Rock Version - in the style of Kenny Loggins or Michael McDonald, The Sea Shanty Version, which is kind of like Black47; The Dancehall Version - Jamaican accent and delivery like Bounty Killer or Shabba Ranks; The Dixieland/1930s Version The Disco Version, which was supposed to sound like Nile Rodger’s band Chic; Bossanova, which is in the style of 1960’s Bossanova and the Pirate version. Mark Frieser and AI must be amongst the effective musical combos since Gilbert and Sullivan. WM
Anthony Haden-Guest (born 2 February 1937) is a British writer, reporter, cartoonist, art critic, poet, and socialite who lives in New York City and London. He is a frequent contributor to major magazines and has had several books published including TRUE COLORS: The Real Life of the Art World and The Last Party, Studio 54, Disco and the Culture of the Night.
view all articles from this author