AI-Powered Lighting Revealed - Is it Just a FAD, or does it work?

[@VladiVladiG]

I appreciate you putting up this video, however, during the rock song part of the video, I noticed that your pre-gain level was really low, your activity level was very low, and the kick drum meter was almost not registering any kicks. Because of that, your lightshow wasn't really at its peak and I think it didn't really show off what the Maestro can do. For example, if you look at your moving heads, they barely move, and when they do, it's because Maestro registered a kick, but for the most part they are not active at all. I think you should really redo that part of the video. I think the demonstration of it wasn't performed well. I have a Maestro, and when dialed in right, it's a nice unit. Thanks again for posting this video!

[@AlterationProduction]

Would love to see your demo of this thing

[@jaredkling2486]

Which is better sound switch or MaestroDMX?

[@RWTupper]

If a MIDI TRIGGERED background effect or cue is running (like a "breathing" effect) on a group of fixtures: What happens AFTER another cue is activated via midi that affects some of the lighting originally controlled by the underlying cue/effect? Does the original queue recover that DMX fixture (and remain in sync) OR does it start over OR does it terminate?

[@djtravismcguire]

I use a product called AI Light Show. It offers a lot more manual control & overrides on the fly and it is a physical controller.

[@ledhed5717]

AI light show is great for DJs. Maestro is a better option for live bands. Great hands free set it and forget it.

[@steinarjonsson_]

So essentially this AI powered lighting controller is a lighting designer in the same sense that the shuffle button is a DJ.

[@BenZuzelski]

It’s some crazy stuff! Great video, I have not gotten a chance to go to LDI yet but sounds like a ton of fun. Personally I hate sound active mode for everything because it tends to look very disorganized. Interesting how you can pre-program the AI. I feel like this would also be helpful for people who don’t know how to program or don’t have the money to hire a lighting designer.

[@mrhalloweenguy]

I am not sure this is the best AI solution. I have been watching Maestro DMX and AI light show for the last several months. Can you pickup the AI light show box and do a comparison?

[@OmegaDerrick]

Personally running ialightshow. And its awesome for my needs. And im using it in live band gigs with 4 chauvet fx hex and 6 chinese light bars. And they working on a new update currently for better static scene controls. Hopefully its out this month or next😎 i saw the meastro vids and i wasnt impressed much.

[@mrhalloweenguy]

⁠@@OmegaDerrickYeah, I have been watching both setups and I like having the physical buttons kind of like the ADJ WMX1 but with AI. Now I am just looking at the finer details like ease of use patching in the fixtures. All this should put a damper on Sound Switch’s future business. I don’t know too many mobile DJ’s that want to spend the time to pre-program their extensive song library in sound switch, plus you can ditch the monthly subscription.

[@DINU_R]

They suggest the pre gain level to be in always green

[@ManjinderBenning]

The audio input was too low in your demo and you didn't calibrate the kick drum sensitivity. This is an unfair representation of Maestro's capabilities. I would recommend you redo the demo once you have had a chance to dial in your audio input better.

[@braintrust12]

i don't think this maestro thing is "AI." for it to be AI, it would have to have a massive set of training data. in this case, time coded light shows synchronized to all sorts of different songs in all sorts of different genres of music. if we're just using other publicly released AI/LLM's as a model (midjourney, chatGPT, invideo etc.) it would suggest that maestro would've had to consume like, a trillion hours' worth of time coded light shows synchronized to music. i'm having a hard time imagining where such a set of training data could be found. i guess the obvious answer is concert footage? if we play devil's advocate and assume there's enough filmed concert footage out there (there's not) the concert footage would have to be static, and concert footage is never static... it cuts all the time between audience/crowd reaction shots, close ups of the band members etc. etc. If we play devil's advocate again, and assume there's enough static concert footage available to train on, the training data would have to be relatively similar depending upon factors in the music like BPM, genre, song structure etc. yet in all my years, I've never seen a scenario in which every LD would light every song the same way, or even a similar way. This isn't an issue for midjourney, where a picture of an elephant is a picture of an elephant is a picture of an elephant: if a human sees ONE picture of elephant, you pretty much get the idea. For midjourney to "get the idea," it needs to see a trillion pictures of elephants - which isn't an issue because trillions of pictures of elephants exist online, and all elephants more or less look similar. Maybe I'm just not understanding exactly WHAT maestro is? But as far as I can tell it's just another sound reactive algorithm that's parsing out frequencies and manipulating characteristics of the light based upon the activity in those frequencies. Basically just another "Lightjams" suite, but this time with a hardware box and marketed as being "AI" because AI is hot and trendy right now. Of course, I'm only speculating here. But I'll be honest - when i watched the demo of maestro doing its thing, it looked like pretty much every other sound reactive algorithm to me: random and arbitrary color fading and animations detached from the qualities of the music. At every point in the music, I feel like I would've made a very different choice. Of course, I have the benefit of being a human with creative control over my mind. But some of choices I would've made differently than maestro seem obvious to me. Don't get me wrong, i absolutely LOVE the concept of maestro! If it truly was some kind of AI solution to creating time-coded-quality light shows at an infinitely scalable rate, that would be a watershed moment in the community... and I'm here for it! But as of right now, I'm skeptical. I'm having a hard time understanding how this is different than Lightjams, and lightjams is going on like 10+ years old now, right?

[@andrewm1967]

need to add this AI to Xlights . will advance th AI really fast

[@JoshSmith-me7oe]

I saw this at LDI this year as well

[@rolisaenz]

Well, at 750 bux.... i better pass AI and still do it with QLC+ or Freestyler and a decommissioned laptop or mini PC

[@nikolaucznaum4312]

Nothing in this application to justify the cost!!...

[@DaBaum69]

If you're too lazy to do your own programming, AI is going to put you out of a job.... Wait for it. Start looking for a different gig, now.