Would you consider the player generators (QNG, BLG, PM, etc.) A. I. ?

[BRIGGS]

I feel like I have an army of droids working for me, and my brain keeps finding ways to better utilize them.

[DaveyG]

For me they are not AI. They don't learn or adapt. They just work to a set of pre-programmed rules.

[BRIGGS]

For me they are not AI. They don't learn or adapt. They just work to a set of pre-programmed rules.

I guess that makes sense, since I'm the one doing the learning and adapting.

[motuscott]

Robot armies make the musics to which the yoot want to twerk.
I got no dog in this race

[Enlightenspeed]

There's different definitions of AI, but for all Rack Extensions it's impossible to do AI as a rule.

[BRIGGS]

There's different definitions of AI, but for all Rack Extensions it's impossible to do AI as a rule.

Good, because the last thing I need, is Skynet taking over my projects.

[BRIGGS]

Robot armies make the musics to which the yoot want to twerk.
I got no dog in this race

Robots should stick to making other robots twerk.

[motuscott]

It's raining memes

[BRIGGS]

It's raining memes

[avasopht]

There's different definitions of AI, but for all Rack Extensions it's impossible to do AI as a rule.

There's nothing stopping the use of reinforcement learning and neural networks in Rack Extensions (other than there probably is no real need).

There may be issues with saving (I've not checked out the SDK since 2016 ... so maybe things have changed by now ).

You will have to roll out your own library and figure out how to make it work with the RE limitations (I wrote a few classes to make allocation in real-time code fairly painless).

The real problem is making it useful.

Now that being said ... QMUL audio professor, Marcus Pearce wrote a paper (A Probabilistic Model of Meter Perception: Simulating Enculturation) that gives you one of the only practical theoretical models based on human perception of rhythm that could be used to generate rhythms on a per-genre basis.

He has written code to demonstrate the theory (not included in the paper), but the model is included.

Worth checking out some of the sources he cites as there's an interesting model (strikingly similar to what is used in Propellerhead's Bassline generator) ... plus lots of findings on the perception of melody that could help you out with generation.

[BRIGGS]

There's nothing stopping the use of reinforcement learning and neural networks in Rack Extensions

kittysized.jpg (53.85 KiB) Viewed 1454 times

[aeox]

The AI that has become increasingly popular is the "neural network" AI. Where you give it parameters and input and using trained models, it creates an output based on the models you trained it on. The player devices and such are more like procedural generation or some other term that is similar

[avasopht]

The AI that has become increasingly popular is the "neural network" AI. Where you give it parameters and input and using trained models, it creates an output based on the models you trained it on. The player devices and such are more like procedural generation or some other term that is similar

Sorry my bad, I was talking about a Rack Extensions (even though the thread is clearly asking about players ‍♂️).

I've no idea what players allow.

[dvdrtldg]

But what is AI exactly, in the music software world? It's one of those terms that get bandied around very loosely

Take something like Sonible's Frei:raum EQ. You play your audio into it, it "listens" to the audio for a few seconds, and then you dial in a frequency curve based on what the plugin thinks is best suited to what it's "hearing". Is that AI?

[homemadenoise]

Good, because the last thing I need, is Skynet taking over my projects.

I kind of wish Skynet would finish a few of tracks for me or at least do some mixing and mastering.

Seriously though A. I. has to biggest buzz term floating around the tech world now. It must help that sweet venture capital pour in.

But there is no "I" in them. They just follow rules. But do not make up new rules or learn. The "A" being to true is debatable. But don't really care. I really like the players for ideas and Chord Generator is a time saver.

[jam-s]

For me they are not AI. They don't learn or adapt. They just work to a set of pre-programmed rules.

They can fall under some definition of AI, but the subject is a bit complex and from your statement you'd most likely only count general AI or super AI as "AI".

[BRIGGS]

I kind of wish Skynet would finish a few of tracks for me or at least do some mixing and mastering.

Seriously though A. I. has to biggest buzz term floating around the tech world now. It must help that sweet venture capital pour in.

But there is no "I" in them. They just follow rules. But do not make up new rules or learn. The "A" being to true is debatable. But don't really care. I really like the players for ideas and Chord Generator is a time saver.

I agree.

The "I", is just pie in the sky.

[avasopht]

But what is AI exactly, in the music software world? It's one of those terms that get bandied around very loosely

Take something like Sonible's Frei:raum EQ. You play your audio into it, it "listens" to the audio for a few seconds, and then you dial in a frequency curve based on what the plugin thinks is best suited to what it's "hearing". Is that AI?

Sony used it to clean up old movies from the 60s.

They could completely remove background noise and separate out sound sources for their remastering (so they were able to have a separate channel for dialogue, the crowd, the chariots, and completely eliminate the background noise).

These methods are the only reason we have the Beatles Docu-film.

I consider Neural Networks to be a marketing term, just like AI. All they are are a series of "linear" layers (which basically means they can be described as a single "matrix" operation (convolution can be represented as a matrix operation) sandwiched between layers of distortion (because if you remove the distortion, you effectively have one linear layer ... because they're linear).

You will notice that Deep learning (which is just Neural Networks with more than 2 linear layers) now dominates image processing. You used to need to learn a hell of a lot of DSP theory to do what you can learn to do much better with neural networks in a few hours.

They're completely replacing experts in things I can't say (NDA, etc.) but it's something that requires 10 years of higher education and active field experience in recognizing those images, but there is a severe and rising shortage of people with this expensive skillset (plus there are only a few positions for them so there are few incentives). They're insanely effective.

And whatever you see ML can do with images, it can do with audio (it's just that what we want to do with audio tends to be much more complex).

If you wanted ML to render a vocal verse from written lyrics, you've got rhythm, melody, tone, expression, and fitting in with all the other instruments.

But we are already seeing ML being used to separate audio. It's most likely the "secret sauce" behind Melodyne's DNA.

Google created Magenta Studio using a simple ML model that allows you to construct new instruments from two existing ones (using a slider to determine how much of the sound is, say, a harp and how much is a trumpet). You get a new acoustic instrument you can play that has characteristics from both, but you can hear is clearly not just blending the sounds together.

[arnigretar]

So if players aint A.I. does that mean they are not my friends anymore?

[crimsonwarlock]

So if players aint A.I. does that mean they are not my friends anymore?

Some people are friends with a doorpost, so I see no problem here

[selig]

But what is AI exactly, in the music software world? It's one of those terms that get bandied around very loosely

Take something like Sonible's Frei:raum EQ. You play your audio into it, it "listens" to the audio for a few seconds, and then you dial in a frequency curve based on what the plugin thinks is best suited to what it's "hearing". Is that AI?

Noise reduction systems have been doing this for decades, reading a sample of the noise and constructing a ‘filter’ to remove it. Same for plugins like VocAlign, where they ‘learn’ the input audio and then do ‘something’ in response.
So no, not AI, at least not how we refer to it today (or at best, a highly simplified/primitive version of it?).

[rorystorm]

I was watching Venus theorys latest video which is also about AI and the next thing that came up was Ryan's thing about the BLG update, and i was like, hmmm....

[avasopht]

Noise reduction systems have been doing this for decades, reading a sample of the noise and constructing a ‘filter’ to remove it. Same for plugins like VocAlign, where they ‘learn’ the input audio and then do ‘something’ in response.
So no, not AI, at least not how we refer to it today (or at best, a highly simplified/primitive version of it?).

But machine learning does far more than create a filter, ... it can separate dialogue from crowd noise, other people taking in the background, .. the train that passes by during filming.

I'm not sure if Sony has released the remasters of the movies they restored with machine learning, but no, nothing comes close to what ML can do.

[aeox]

But what is AI exactly, in the music software world? It's one of those terms that get bandied around very loosely

Take something like Sonible's Frei:raum EQ. You play your audio into it, it "listens" to the audio for a few seconds, and then you dial in a frequency curve based on what the plugin thinks is best suited to what it's "hearing". Is that AI?

neural network AI is where you can give input parameters such as

"create a 120 bpm track in the style of (insert your favorite artist here) mixed with (insert your other favorite artist here). "

Then, without the AI explicitly having any knowledge of such a combination of ideas, will try to interpret and give you an output. As long as the model has been trained to understand what each of those artists sounds like

With your EQ example, it's more like analysis and automation rather than modern neural net AI. I'm thinking it's probably matching frequency curves to the fletcher munson.

[homemadenoise]

So if players aint A.I. does that mean they are not my friends anymore?

QNG will always be my enemy.