AI Live Mixing?

[Smallmouth_Bass]

So, as many of us playing smaller gigs and bars know, you often have to bring your own equipment and run the PA yourself. My main band does and I am usually at the helm, for better or worse. The problem is, you cannot effectively mix from the stage and especially while you are playing.

All this talk of artificial intelligence (AI) these days and it's potential capabilities has me thinking, could the future of mixing be AI where there's a mic in the front (or somewhere in the room, maybe multiple points) and it doing all the adjustments automatically? How great would that be?

 

[Lammchop93]

I don't see why it can't be done. By everyone's perception of a "perfect mix" is different. So the AI would have to be trained according to a specific sound the band is going for.

 

[Smallmouth_Bass]

I don't see why it can't be done. By everyone's perception of a "perfect mix" is different. So the AI would have to be trained according to a specific sound the band is going for.

"AI, mix in the style of ______."

 

[Smallmouth_Bass]

Or with different types of "presets".

 

[Lammchop93]

That would make sense! The designer would have to train it for all of those presets, but that can't be that difficult.

I could see it being useful, and the technology is out there to do it. Now just need someone smarter than me to actually do it

 

[Smallmouth_Bass]

That would make sense! The designer would have to train it for all of those presets, but that can't be that difficult.

I could see it being useful, and the technology is out there to do it. Now just need someone smarter than me to actually do it

Now if they could add a feature to make the acoustic drums quieter! Might have to rig the drum throne to give shocks if it goes over a certain volume.

 

[Bassaga]

That will be nice. And it'll be a nice compromise between AI enhancement and real live human music that I think most people could accept.

 

[Discount Bassy]

Presonus AI Mixer, now featuring patented Predictive Suck Technology (PST!), automatically senses and evaluates talent deficiencies in real time, adjusting sliders to minimize problems in less than 10mS. For an additional fee, players can enhance their individual channels to increase volume over AI suggested levels and add realistic applause with their Enhanced Gain Option (EGO) monthly service plan.

 

[Marko 1]

You'd have to get rid of the amps on stage...

and have e-drums...

at least for most bands.

 

[MultiScaleMale]

It probably isn't that difficult to implement. Auto-ducking for the kick, prominence for the vocals, proper slotting, auto tempo match for the delay, instant feedback elimination. But will they do it? No need for a sound person. It would be like Fender making a bass that plays itself. No need for a bass player.

 

[Smallmouth_Bass]

You'd have to get rid of the amps on stage...

and have e-drums...

at least for most bands.

You'd have to give it a little info and then it would make adjustments in consequence. But yeah, acoustic drums would likely be the biggest challenge.

 

[Smallmouth_Bass]

It probably isn't that difficult to implement. Auto-ducking for the kick, prominence for the vocals, proper slotting, auto tempo match for the delay, instant feedback elimination. But will they do it? No need for a sound person. It would be like Fender making a bass that plays itself. No need for a bass player.

Maybe it could tap into the guitarist's multi-effects and clean things up and take out all the bass hogging frequencies!

 

[john_g]

Maybe it could send in Arnold Schwarzenegger to take out the drummer if he messes up

 

[chadds]

"AI?"
"Get a better singer and fire the drummer."

 

[callofcthulhu]

It probably isn't that difficult to implement. Auto-ducking for the kick, prominence for the vocals, proper slotting, auto tempo match for the delay, instant feedback elimination. But will they do it? No need for a sound person. It would be like Fender making a bass that plays itself. No need for a bass player.

Fender would absolutely sell a bass that could play itself if they thought they could sell more of them in a year than they could P basses.

The target market for an AI mixer wouldn't be sound guys, it would be people making music who want to sound good but for whatever reasons aren't able to operate a mixing board. I'm willing to bet there are a lot more of those people out there in the world than there are sound guys.

Besides which we'd probably need to see "AI assisted" mixing boards emerge onto the scene for a while before the tech could advance to a truly automated mode, and such an "assist" feature might very well be useful even to a well seasoned pro.

 

[logdrum]

Audient EV0 8 already implemented this. I used it for the drums and the room mic sound. It was more an automatic gain control not an AI mix per se. I slaved the EVO8 with my my Antelope Audio Zen Tour via ADAT. It was a live recording of a band, where I was also participating as a horn player. Worked great.

 

[REV]

It's bad enough having a sound guy in control of how the band sounds let alone a computer making those decisions.

 

[Carl Hillman]

Are they going to program it to give you attitude...like a real sound guy?

 

[Nashrakh]

We can get inaudible mixes with way too much kick drum without the need to twiddle knobs? Sign me up!

 

[Deleted member 271037]

This particular use of AI is probably VERY far down the list of priorities at companies with the necessary level of competence for such an advanced use of AI. There are LOTS of variables involved in live mixing. Unexpected things that may happen, different mixes are needed for different genres of music, what do do when someone pulls out an unusual ethnic instrument, human preferences in sound, different rooms, the available gear/mics, etc.etc. I don't see how this can work flawlessly without developing a whole infrastructure around it, where the AI can seamlessly access the mixer/effects and be able to "listen" to mics at various locations in the room. Maybe even an ampless setup where the AI can access individual virtual "amps".

The AI would also need to be able to respond to "more bass in the monitor" (but NOT if the lyrics of the song contain "more bass in the monitor").

Anything other than a generic "one mix fits all" solution for a standard vocals/guitar/bass/keys/drums setup on smaller gigs would require a VERY advanced level of AI. Although I guess even a simple "one mix fits all" solution could be an improvement over the bass player doing the mixing on-stage...