BMIM x ADE: How To Make The Best of AI

Buma Music In Motion was part of this year’s Amsterdam Dance Event with a fully stacked conference day at the Main Business Hub Felix Meritis. Hundreds of music professionals joined our talks and panels, where we dove into the latest developments in our field. Besides discussing future challenges and opportunities with experts, we invited many well-known composers, producers, and other music professionals to share their insights. In these features, we’ll share their knowledge with you.

Text by Meike Jentjes
Photos by Birgit Bijl

No fear of AI

To be honest, who hasn’t contemplated the idea of AI replacing their job in the not-so-foreseeable future? At Buma Music In Motion, we love to think about AI as an opportunity instead of a threat. Our panel, ‘How To Make the Best of AI’, follows up on our Buma Music In Motion Special panel, with composer Tomas Louter of Media.Monks. He answers some of the most pressing questions regarding this subject and paints a picture of how we can use AI to our best interest, without letting it scare us.

Digital-first

Louter explains that Media.Monks is a digital-first marketing, advertising, and technology services company working on AI since day one. The ideology: AI could be the new technology that wakes an entire industry up again, just like we’ve seen in the past with the internet, the smartphone, or even film in colour. That spans to music, too, so let’s dive into how AI affects music composers by means of the most advanced tools and three categories: music and sound generation, voiceover generation, and post-production.

Generating music and sound

‘Let’s start with prompt-based music. One of the easiest-to-use tools is Stable Audio, which is free up to twenty prompts per month and feels very unrestricted. The results may not be what you would use in commercials right away, but the surprise element of this AI-generated sound does help your creativity’, Tomas Louter explains. He goes on to say that computer-generated music has a totally different character than what people would make, which makes for good samples you can use. But, like with all AI, prompts are important. His biggest tip: Place commas between different parts so that the language part of the engine understands what you ask of it. See working with tools like Stable Audio, not as a one-stop shop; rather, use it to spark your imagination when you’re stuck in a rut. ‘Try using tools like Emergent Drums to generate musical elements like kick drums or melody samples to boost your catalogue and creativity.’

“Be aware that you’re still responsible as the tool user, instead of the company behind the tool.”

Tomas Louter

‘This is where it gets interesting for me. Synthplant has made a module on their existing synthesiser engine, to which you can give an input sample, and the synthesiser will actually make a sound much like the input. Imagine creating a sound like Hans Zimmer’s music: it’s all playable and possible. The whole theme of the synthesiser is DNA and evolution and the tool makes four strings that will make a sound that sounds just like what you would want.’ In terms of copyright: if you use this sample that sounds like another composer from Splice, you’ll be able to use it – but be aware that you’re still responsible as the tool user, instead of the company behind the tool.

Voiceover generation

How can we generate voiceovers that sound natural? How can we make translations that sound natural? And how can we make bad recordings sound like they are in a studio? As Tomas Louter explains, using voice samples can be a minefield in terms of copyright. ‘I can’t just use Drake’s voice in my productions without him getting very angry and coming after you for a lot of money. Having said that, artist Grimes has already said that if someone uses her voice and is successful with it, they get 50% royalties. Pretty much a deal! I work in media and advertising, so that doesn’t seem like a good idea for our company. But for use for yourself, it’s really fun.’

Louter is a fan of tools like emvoice, which has a text-to-speech form and a MIDI interface, with which you can project the letters of your sentences on a note and drag into Ableton afterwards. He explains: ‘A company like Eleven Labs does a great job with text-to-speech, especially with personalised experiences for customers. Media.Monks did a campaign with Tomorrowland where we used Eleven Labs, where users could create their own track with AI. That’s one of the coolest developments with AI, in my opinion.’

Post-production

‘GOYO is a great for mixing and mastering. It takes our recordings, breaks them down, and builds them back up to eliminate all the noise, reverb, and voice impurities. Say you have a recording that sounds like someone recorded somewhere in a bathroom; use GOYO to remove those background noises and clean up your sample.’