AI and Mixing: Will Artificial Intelligence Replace Engineers?
Love Music More · hosted by Luke F. Walton (Scoobert Doobert) · Solo episode
Listen
Topics discussed
Host note
"Every mix shouldn't be the same; every vocal chain shouldn't be programmed by Dave Pensado." That's the line I kept returning to after NAMM 2023, where everyone had bold takes about AI replacing mix engineers. The gap between what top engineers said and what ML researchers said was striking, and it comes down to who's looking at the problem from the top down versus the bottom up.
I weigh in on the "top 5% stays human, 95% gets automated" claim, what Waves' repeated subscription pivots actually signal about where the money is, and how Splice is approaching AI differently. Routine mixing getting automated is real, but routine mixing was never the interesting part.
The question I care most about is where creative decisions live, and why outsourcing those to a model trained on consensus is a fundamentally different move than outsourcing gain staging.
Selected moments
- Introduction and AI Buzz at NAMM 0:50 I discuss the buzz around AI at NAMM 2023 and its implications for audio engineering.
- Contrasting Views on AI and Engineering 1:32 I share thoughts on the different perspectives between top engineers and machine learning scientists.
- Waves and Monetization Issues 2:21 I critique Waves' move towards subscription models for their plugins.
- AI's Potential to Automate Mixing 4:32 I discuss how AI could change traditional mixing processes.
- Human vs AI in Mixing 6:00 I reflect on the statement that most mixing jobs might be replaced by AI.
- Creativity vs Automation 8:35 I emphasize the importance of keeping creativity at the center of mixing.
- Final Thoughts on AI in Music 16:30 I wrap up my thoughts on how AI can assist rather than replace creativity.
Selected excerpts
Artificial intelligence and what it's going to do to audio engineering. So everybody had these really bold takes.
If you're at the top sometimes it's harder to see what all the people at the bottom are actually doing.
Waves makes, you know, virtual instruments, virtual reverbs, virtual versions of analog gear... But the funny thing about them is that they keep realizing that they aren't making enough money.
I'm going to let the AI look through this audio and go, 'Oh, this kind of sounds like a female vocalist.'
The top 5% will stay human; the rest, those 95%, will be AI versions of himself.
Mixing is highly creative. Allow me to still make my decisions, my creative decisions.
Every mix shouldn't be the same; every vocal chain shouldn't be programmed by Dave Pensado.
FAQ
Will AI replace audio engineers?
I discuss the implications of AI on audio engineering jobs and argue that while automation can assist workflows, creativity should remain a central focus.
How is Waves adapting to AI in audio mixing?
I critique Waves for shifting to a subscription model for plugins and discuss their innovations in light of AI technology.
What are the benefits and drawbacks of using AI in music production?
I see potential for AI to ease certain tasks but warn against fully replacing human creativity in the mixing process.
Curated notes only — no public transcript. Listen on the links above.