in Projects, Programming, Technology

AI for Audio Sample Browsing

This is part of a series on Opportunities for AI in Music Production.

Loopcloud, for me, is currently the best tool for picking samples because it allows you to pitch them and play them together as you build your song idea. Unfortunately, the sample search is rudimentary. There are not enough dimensions to filter by and no automated concept of “other sounds that would sound good with this.”

Problem / I have 10GB of samples on my HDD. I want to browse them based on genre, instrument, duration, key, bpm, and warp & transpose them as I listen to them. I’d also browse a cloud view like XLN’s XO drum machine sample browser. Sononym is good, but far from great at this task.

Solution / Modern ML techniques should be able to categorize, transform, and even generate new samples like crazy. Give me a desktop app to do that on top of my sample library.

LoopCloud & Splice will get there slowly with cloud-based micropayment architecture – which is probably how it should be.

Sononym exists to solve the problem of having too many samples on your local hard drive. Categorization and browsing not as good as XO (which only handles drum sounds) yet, but the UI is modern (high quality) and hopefully will improve.

Update: Loopcloud has updated their drum sample player with a “recommend” feature to let you quickly try new drum sounds (or entire kits) that are similar to the one you’ve got.

Next: Text to Singing: AI to Generate Vocals

Write a Comment