Nat Sarkissian on the Generative Landscape

Over the last couple of years, the San Diego-based generative artist Nat Sarkissian has come to prominence for a number of projects that reveal nature as a generative system. Recently, he broke new ground with his work Eucalyptus and Sagebrush (2023), auctioned as part of Cure3, an exhibition organized by Cure Parkinson’s in association with Bonhams and Artwise with the critical aim of finding a cure for Parkinson’s Disease. He followed that up with two projects, Sunset from the Bluffs and Hello Modulo, that exemplify his range as an artist and skill as a coder. We were delighted to have the opportunity to discuss Nat’s work in greater depth — exploring his process and inspirations, while seeking to establish the true nature of generative art. 

‍Johnny Dean Mann: To start, we wanted to talk with you about some of the recent work that you’ve been doing. Projects like Eucalyptus and Sagebrush and Sunset from the Bluffs approach the look and feel of pastel or crayon. What is the creative process behind that work and what is the appeal of trying to capture hand-drawn elements?

‍Nat Sarkissian: I actually spend a lot of time — when I’m trying to come up with something to get interested in or just to brainstorm — on Pinterest. And I think it was late Fall last year, I kept coming across these colored pencil and pastel landscapes. I’d seen them before and always kind of liked them, but something was clicking. A lot of the shapes are not very complex, at least the ones that I was looking at, but there’s so much that comes from the texture of the medium that was adding a really nice richness. I think that’s what I was drawn to.

I started analyzing them, thinking about what it would look like to simulate this and what was happening on the paper that was so interesting compared to flat colors on the computer screen. Initially, the creative process was more analytical than creative. I was thinking about putting together a toolset of basic shapes because, at this point, that’s what everything I work with is: rectangles and ovals put in the right positions and sizes. I’m thinking: “Okay, if I can make the toolset of those basic shapes rendered in something that looks like that, then from there, I can hopefully just do whatever I want.”

These scribbly shapes kind of mirrored the actual experience of learning to draw as a child. 

Eventually, I wanted to try and make something more legitimate out of it instead of little doodles. The first thing I thought of was some sort of landscape. I merged a couple of the concepts that I’d been working with on previous projects with this new toolset into a proof of concept that I was happy with. Eucalyptus and Sagebrush was just a couple of steps past that.

‍JDM: A great deal of generative art has a kind of precision and cleanliness to it, but there is a parallel idea of working with code to make something that looks very organic. You’ve talked in the past about the Californian landscape as a source of visual inspiration, but are there other visuals that you find yourself returning to in your work?‍

NS: I’d say broadly, most of my time outside in nature does that to me. 

Sometimes, nature has a very direct influence — where I want to try to make those hills or that view. 

But there have definitely also been occasions when I’ve realized that I was headed in a pretty clear direction anyway. I mean, that’s how most of these projects came together. I absorb a lot simply by observing art online, but also at a couple of layers deeper, where it’s even less conscious. Another pretty major source is my cousin, who’s been an illustrator for a long time and with whom I grew up making art. 

‍JDM:  What was your approach with the toolset? How did you create the brushes and those organic textures? 

‍NS: I looked at it pretty straightforwardly. I was thinking about filling in, let’s say, a rectangular area with oscillating scribbles, conceiving of the pen as an object accelerating back and forth — every time it gets to one side, it’s got a new target on the other side. As it goes, the pen is depositing tiny circles that mimic pieces of crayon or lead. There are a lot of parameters and things you can do to tune that shape. For example, as it’s moving it’s dropping points within a radius of the pinpoint, so widening or narrowing that can make it seem like a thicker crayon versus a hard pen or pencil. You can do the same thing within a circular region or arbitrary quadrilateral region. After that, it was really just a matter of trying a bunch of permutations. 

‍JDM: Is there any analog reference for the color palettes? Did you take reference photographs of any particular local area for your landscape work?

‍NS: Regarding the color palettes, I think I used one reference: a painting I really like that I pulled a handful of colors out of. The rest of the palette emerged independently. For the most part, it was a matter of sorting palettes according to a combination of brightness and temperature. So you get hot yellows and pinks for places that are supposed to look hot and sunbleached. But on the other side of the same hill you get cooler temperatures that feel like shade.

JDM: You have a very distinct sense of light and shade in your works. How do you achieve that in a technical sense? Is the base topology a kind of 3D mesh with a particular light source?

‍NS: For every landscape piece that I’ve done so far the base has been a three-dimensional height map, using noise to give it a general shape. Ever since Reconnaissance (2022), I’ve also been using a piece of code that simulates erosion from rain. Once I have that 3D object in code, I can calculate the angle at any point along its surface and how that responds to either point light or directional light. In this case, it was a simple directional light. 

You can start off with a really unrealistic, blobby terrain. But then you can simulate hundreds of thousands of raindrops across its surface, carving more natural-looking peaks and valleys. 

‍JDM: The erosion process seems quite resource-intensive. How do you try to balance speed and efficiency with the effectiveness of the technique? 

‍NS: It depends on the piece. There’s a limit to what I can do and what works on my computer isn’t necessarily going to work on the average computer. The reason the erosion is so slow is because it is on the CPU right now. I have that somewhere on my mental to-do list to come back to the algorithm and put it on the GPU, because I know that it’s possible. Timing is also a big issue, but finding creative ways to deal with it has actually led to things that I liked visually and ended up choosing intentionally.

‍JDM: You’ve mentioned creating together with your cousin. Does your current approach still veer toward the playful?

‍NS: I think it’s definitely playful. That’s what got me into this space in the first place. I never really got interested in computer science and the nitty-gritty of the technicals. As a kid, I was looking at games and thinking: “Wow, I would love to be creative in that way.” It seems that there’s so much more depth there than in a simple painting or drawing because of the interactivity. The ability to build worlds was really fascinating to me — that’s really integral to my way of thinking.

‍JDM: There seems to be a “three-stage” resolution of fxhash projects. There’s the code itself, the midpoint where it’s either creating itself or you’re interacting with it, and then the final output, either printed or displayed. Do you see your practice as a three-stage process? Your early works don’t seem to fit that pattern, but now they seem almost to revel in it.‍

NS: I see that as a pretty good classification of the three stages, and I guess the rendering stage or the live stage is the one you’re asking about. If I put up California Hills in Late Sun (2022) and made you wait on a blank screen until you saw the image, it would be pretty rough. So I decided: “Okay, I’ll just show them every frame as it builds itself. At least that’s kind of interesting, albeit in a very different way to the final product.” Initially, it was a necessity. And if anything, I kind of saw it as a flaw: this code is egregiously slow, and you have to wait for all of this time just to see the output. But in later projects, I’ve learned to appreciate it more.

With Eucalyptus and Sagebrush, I started watching it render and found myself enjoying it. I felt like I was watching someone draw something in time-lapse.

I usually think about the final image as the art, so how you get there has been a little bit less important to me. It’s certainly not something to prioritize deeply. I wouldn’t ever sacrifice something in the final image to make the creation stage more interesting.

‍JDM: You describe your project, Artifact (2022), as “bringing light to unreal shapes and forms and uncovering artifacts in an impossible space.” Like Curtain (2022), it has a non-organic aesthetic that contrasts your other pieces. Is that something you’re looking to move more into, or is your main focus natural, organic shapes and compositions?

‍NS: To be honest, it is actually more of a “moving back to.” Artifact is much more representative of what I was doing for the last fifteen years up until I started working on fx(hash). I’ve been fascinated by doing landscapes over the last year, and I’m sure I’ll be doing more of that in the future. But I see projects like Curtain and Artifact as representative of what got me into generative art in the first place. What I’ve got coming up is actually more in that vein than landscape.

‍JDM: Where do you see the art in generative art? Is it the final product? Is it the process leading up to it? Or is it a combination?

‍NS: For me, the art is everything that happens after you click Run. That’s been it for me for a long time. I never got into super-efficient code or making things really compact. I have a lot of respect for people who do that but, for me, it is an entirely different discipline. Sometimes, I almost don’t even really care if the code is as optimized as it could be or if it’s pretty or what someone would think about it. That’s very much secondary to me.

I think that the running of the code and what it does after you execute it is the art.

‍JDM: Finally, we did a little piece on you for Cure³, where we described your work as “a string of loosely interconnected projects that form a sort of hazy narrative of discovery both internal and external.” We wondered whether you agree or disagree with that interpretation.  

‍NS: I think that’s a really reasonable way to see it (from the outside). The only part of it that I have more clarity on, being the person in it, is the through line connecting my projects, which is a lot more straightforward (barring Curtain, which is a little piece of my past that I plopped into the series). Looking at the code, there’s a reuse and iteration on concepts, with elements added to step each project forward.