Hi, Brian! Happy New Year. Could you describe why you and your team at Atomica chose AIMS?
About nine months ago we started using AIMS, comparing it to other music similarity search technology out there and trying to find one that really works. We got hooked on AIMS pretty quickly because we were finding real gold when it came to client briefs.
We compared AIMS to just about every other option on the market and found that other tools are less flexible, relying on tagging and lacking the option of segment-specific search.
AIMS is hands-down the best music similarity search tool we’ve used.
Sounds like AIMS was a perfect fit for Atomica!
Definitely. We create more reduced mixes than the industry average, and what really surprised us about AIMS was how it can find matching tracks we never would have thought to search for. We’d listen to the reference track and think “there’s nothing like this in our catalog, but let’s throw it into AIMS.” It would throw back six or seven reduced options, like just drums and bass, that really sounded like the seed track. AIMS finds tracks we wouldn’t think to search for because the album description tricked us into thinking of something too specific.
Other music similarity searches didn’t work in these situations. They mostly use tags of the primary title rather than an alternative mix. And the reduced mixes don’t fit the tag’s description. The pure sonic search you get with AIMS works much better than tagging because it finds altered mixes that aren’t even part of the genre but still fit.
That’s great to hear! So how does your team use AIMS on a daily basis?
We have three people on our team who use AIMS daily. Our model is such that 70% of our clients rely on us to do the music similarity searches, so a tool like this is fairly critical. AIMS has reduced our search time by 35 or 45%, plus it gives us tracks that we wouldn’t have found otherwise.
I know our music library very well. I look at a brief and immediately know four or five tracks that will match, and I know exactly where they are. But AIMS is still very helpful, especially the segmentation tool, which is critical for us.
Clients always send us a track and say we want something like this build or the overall vibe. With AIMS we can focus on just one particular aspect of the track, which is perfect. And I’m extremely pleased with how AIMS is handling vocals
You mentioned how well you know your catalog. Just out of curiosity, how do you see the relationship between Artificial Intelligence and human expertise in music similarity search?
The human mind is limited by how much it can remember. It also has its biases of what’s good and what’s not. Technology like AIMS doesn’t have that. It’s just facts. Biases and favorites are gone. Like any human, I have tracks and albums that I gravitate toward. AIMS has pushed me outside of that and helped me dig deeper into our catalog. I might not go back to music that’s six years old, but AIMS does. It helps us activate our whole catalog.
Years ago music similarity search technology didn’t work very well. Everyone said you just can’t beat the human element. Now we’re seeing the opposite side: there’s a part of the human element that you don’t want in a search. At Atomica we have 80,000 tracks, which is a lot, but it could also be a lot bigger. Who can keep up with that much music? The bigger your catalog, the more important technology like AIMS becomes.
If you were to recommend AIMS to others, what would you say?
AIMS is the first and only accurate music similarity search technology in the field. We’ve been testing tools like this for a long time and they’ve never worked. AIMS is the first one that’s accurate and it’s the best on market.
Want To Try AIMS?
It’s easy to set up a free demo with your own catalog and test AIMS in real-life situations. Just get in touch.