Zoe Wrenn, a musician, software engineer and founder of new music-making platform Tamber, is creating a future where musicians are able to make synthesizers sound like chocolate or snares that feel like the color blue. “If I were to describe this platform, I’d say I want it to make writing music feel magical. Inspiring. That’s my goal,” she says.
Tamber, which Wrenn has described previously as an “Adobe Creative Suite for music,” uses artificial intelligence to transform feelings, colors, sounds and other descriptive text into musical ideas, and Wrenn believes it could be the antidote to the rise of generative AI tools that are training on “stolen data” and dominating the market right now.
Related
“I’m seeing a lot of new AI music companies, and there are two questions I ask immediately: Where did their training data come from, and are they being sued by the musicians they’re claiming to help? If they don’t have good answers to that, I am not interested,” she says. “I believe there has to be another option.”
Tamber, which launches Monday (May 18), features a number of playful, hi-tech tools to jumpstart songwriting, including Gestures, which lets musicians shape the effects and tone music by recording via hand movements; Librarian, which allows users to scan their own sample library and transform them into something new; and City Packs, which offers sample libraries collected from cities around the world.
To make the service more approachable, Tamber is centered around Tamby, the company’s animated mascot, designed to work alongside a user and to help automate parameters, build chains, swap out instruments and more. Over time, Tamby remembers the user’s preferences and customizes itself to become an increasingly specialized production partner.
To date, Tamber has been backed, as part of a recent $5 million funding round, by Adobe Ventures, M13, Rackhouse Venture Capital and a number of undisclosed artist investors — but the toolkit wasn’t always intended to be a company.
Wrenn first developed Tamber during the pandemic, when tours were grounded, for her own personal use. But after “Hailey,” a song of hers created with Tamber, began to take off, Wrenn started sharing the tool with musicmaking friends and found it was a boon for burned-out creatives, prompting her to turn it from a personal project into a full fledged start-up.
“These days, there’s this overwhelming psychological pressure to keep up with the speed at which music is being discovered and created,” Wrenn says. “There was a turning point where all musicians were suddenly expected to create more music at an unnatural speed and to also become content creators. When I looked for tools to help me keep up with that, I found the tools were either blatantly robbing us or just weren’t particularly useful. I believe Tamber is the right solution.”
Tamber has a number of different tools to choose from. What is one tool you’re particularly excited about?
I’m really excited about our Librarian product. Basically, you can scan your entire computer and identify any piece of audio you have, and then index it with our models on top of it. So you can then search through your computer and be like, “I want a guitar that feels like blue,” or “tastes like chocolate,” etc. — but through your own samples. So they don’t have to be labeled. They can be labeled [with] a whole bunch of numbers or just nothing, and it will still be able to find them. And then you have this cool thing where you can press play and it pulls that sample up, and you can just kind of start stumbling through all of your samples. That was a need that a lot of musicians were asking for — to make their existing sample libraries and their computers much more intelligent, with our synesthetic data on top.
Do you see Tamber as a competitor of Splice, Suno or none of the above?
I think Tamber does a lot of things, so we feel we are defining a category as a sandbox for all kinds of music-making tools. I do think we are unique. In terms of comparing us to Splice — I know they’re trying to dive into some AI tools right now. I guess that would be somewhat similar. They’re doing a little bit more generative stuff, which we’re not really interested in touching right now.
With Suno, they’re doing the generative thing and now kind of backpedaling to be like, “Well, we’re all for artists, we’re just for artists,” because they realized artists are who they need. I have yet to see anything from them that I would want to use as an artist.
Tamber is powered by artificial intelligence, but you’re outspoken against companies like Suno. How do you see Tamber as leveraging this new tech in a positive way?
I mean, right off the bat, we don’t train on any third-party audio at all. A lot of the other music technology and AI companies have, as everyone knows, scraped a whole bunch of stuff and stolen the majority of their training data. We don’t do anything like that. We’re also not giving you a generative output — there’s no “press a button and a song spits out.” I’ve talked to a lot of musicians, and that’s not something people are interested in. We’re really not trying to flood the market with a whole bunch of slop. The main goal is to create something that helps people do things faster and better.
Related
One of the key components of Tamber is Tamby, a character, which kind of reminds me of what Clippy was like for Microsoft Word, that guides you through the service. How did you develop this character, and why do you think it works as a way to help users explore the product?
I want the user to be able to actually identify with something, instead of it being some kind of sterile thing with no face to it. I had just started reading this book called ‘Mascot,’ which identified some really amazing mascots in public media that are maybe a little less known. I was inspired from that and from a lot of kawaii emoticons as well. That is why it looks kind of early 2000s. I wanted to build something very cute, very futuristic and simple too.
There are a lot of AI music companies right now with different visions of the future. You have some pursuing AI-powered remixing for fans, high-tech DAWs, commercial music library creation or the “push a button and a song pops out” idea. How would you like to see the future play out for AI in music?
In terms of what I would want in the future, I want it to feel magical and not replicative, and I would want it to be things that artists actually want. I think it’s really insane that the first thing the space did was create a whole bunch of fully output songs built on stolen audio. It just feels like — what are we doing? It’s silly. And I do really think there’s going to be a future where the ones like Tamber succeed, where we’re not stealing from artists, where we genuinely care. Everyone here is an actual musician who has toured and done this thing.
