If you don't think visual perception can be quantified in any way, you should probably bail on this discussion immediately. Since that is what it is about. Specifically, how can one work with the new controls within Studio Artist V5.5 to to customize and fine turn your artwork.
By new controls, let's focus on two specific things.
1: Generative Strategies (and associated Generative Preference Options)
2: Visual Attribute Modulation
You can certainly think of a Generative Strategy as a recipe to follow to get from point A to point B.
Point A being a blank canvas.
Point B being a canvas with something interesting in it.
Now some people might immediately get all huffy and start start spurting spittle as they rant about how one will never achieve greatness following 'a recipe' to create art. Oh really.
Your generative strategy can be anything you want it to be. A particularly famous one is grabbing some cans of house paint, then pouring that liquid paint onto a canvas while you flick the paint brush within the falling stream of liquid paint, there-by melding hand-arm motor movement with the chaotic dynamics of liquid paint.
Is it genius or idiocy? That is not our place to judge. It certainly created a lot of attention for both the artwork and the artist at the time it was conceived.
And you are certainly free to grab that 'Generative Strategy' and run with it yourself. If you so choose to do so. There are endless ways one could emulate and or expand upon it within Studio Artist.
It is 'a recipe' one could follow to create new art. A recipe that incorporates elements of chaos, chance, randomness, emotional fury (mediating the hand-arm muscle movements that are a part of it). I think you are required to puff on a cigarette (preferably somewhat hung-over from excessive drinking the night before) while doing it if you really want to stay true to it's original form.
But that is art history.
We are going to discuss something new today.
We are going to talk about building generative strategies that incorporate perceptual modulation.
Before we deep dive into the specifics of what that exactly entails, let's take a look at an image.
What is going on in this image? Take a good look. What do you see?
It is the product of visual attribute modulation. Actually some fairly extreme visual attribute modulation (8 different instances of it in my quick run through of the vectorizer control panels).
Now what if we turn all of that visual attribute modulation off. What do we get.
Now this second image with no visual attribute modulation sure looks very different. Right?
What is going on in this second image? How is it different from the first one?
...
Let's change tack for a second.
How would one modify that second image to get to the first one?
Since we're talking about Studio Artist, how would one adjust the settings in Studio Artist to get from that second image to the first one?
Now historically, we would have said something along the lines of, oh that's easy, just drop into the Vectorizer Editor control panels, and dial up the visual attribute modulation you so desire, and you are done.
And that statement is certainly true (at least the last part). What i have notice over time is that the first part may be true for individuals like myself (and even then, oftentimes not necessarily easy but kind of tedious at times, especially for certain kinds of inter-dependent multi-parameter edits). But many people can't be bothered to even open the Editor up, or if they do their eyes glaze over, and they go back to the preset browser.
So part of what we're trying to do in Studio Artist V5.5 (and in the future in general) is to give people the ability to make these kinds of decisions at a much higher level. So rather than having to dive into the Editor control panels and start hand editing individual parameters to get to where you want to go, you can instead make some high level generative strategy decisions, then let Studio Artist do the work.
Now there is this whole other property associated with both the real world and the world of creating artwork. And that is the notion that relentlessly optimizing your objective is not necessarily the right way to reach your objective. In fact, it might be totally the wrong approach.
People who seriously study this phenomena (and yes, you can study it quite rigorously if you want to) talk about the need for systems to incorporate 'diversity'. So if you buy into this notion (and you should), then you ideally want your generative art strategy to incorporate elements of diversity creation into it as well.
...
So there's a quick into into some of the things we are trying to address in some of the new Studio Artist V5.5 features. I'll dive into some more specific examples of visual attribute modulation in a second post in this thread.
Replies
Of course that second image above (in the original top post) actually still has some 'visual attribute modulation' going on in it, since it's a Vectorizer Shape Technique effect. So the shapes are generated by the vectorizer analyzing some source image. It's a pretty abstract vectorizer shape effect, but it's still visually modulated in some way.
So if we totally turn off all visual attribute modulation, we get a white canvas. While that has a certain purity of form, it's ultimately pretty boring.
Here's a few examples of some experiments I ran last night. They are all generated using only the Studio Artist V5.5 Vectorizer. They all make pretty heavy usage of visual attribute modulation. I did no manual editing in the Editor. I did not use any pre-built presets. All I did was make some generative strategy decisions and press a button to start my generative strategy as a cyclical process, then pressed the Grab button when I saw something I liked enough to save.
Let's take a look at another image created with a slightly different generative strategy. I'm still only using the vectorizer. And I'm recursively processing what is inside of the canvas with the vectorizer.
So I'm not actually processing the source area at all with the vectorizer. It's using what is in the canvas as it's input, and then spitting it's output back onto the canvas. Over and over again.
However, some source area visual attributes are leaking into the internal vectorizer parameter modulation. So visual attribute modulation is actively at work in the generative process. Just not in the normal way, where you are using the source as the input to whatever Studio Artist algorithm you are running.
And I find this whole notion of what visual attribute modulation does to the perception of the modulated output imagery fascinating. it can seem to hint at some kind of representation hidden within it.
Or be much more abstract.
But the statistics of the real world are still being reproduced in some form in the resulting visual attribute modulated output images. That's why they look so very much different from totally non-visually-modulated output. Which is only going to reflect any 'random' modulation going on inside of the internal process that creates the imagery.
Has it symmetry generation for vectors?
You mean symmetry flips based on the indexing, or horz - vert region indexing? Or?
So adding symmetry as a modulator is something we've been thinking about.
I just ask myself the question, whether there is also the posibility to influence still the running generative process by user-interaction in some way (mouse follow or so). These latest samples are fascinating. thanks
So i thought about what you are asking for.
And keep in mind that you are asking to add interactive modulation from the pen to the Vectorizer. But since you can drop the Vectorizer into a DualMode Paint container, it does make sense in that situation to add the interactive modulation option.
Studio Artist V5.5 has this concept of one process being able to hijack the interior modulation of a second process and override that modulation with something else. The new DualMode Paint features use it extensively.
So that's the mechanism we can use to do what you want without mucking up the Vectorizer parameter space by adding interactive modulator options (since in normal vectorizer operations they make no sense).
Probably the first thing i will try with it is to interactively modulate the Vectorizer processing the paint canvas in a painting style DualMode Paint preset. You could tie pressure to region size, and pen tilt orientation to the region position offset rotation. So you're using the Vectorizer to act as an interative dispersion mechanism in the overall DualMode Paint preset.