When Imre shared the result of his first Blendergrid project, I was really impressed with the quality of his work. After going back and forth a bit, we thought it would be interesting to share this project with the community through an article here. So in this interview with Imre, he shares how he's using Blender both as an artist and an Architecture student to create some stunning renders :)
How did you get started with art and 3D?
Art has been a part of my family for a long time, so naturally it has been part of my life since I was little. I was always interested in maps of cities and ancient empires for some reason, and I used to draw my own cities, and design buildings that would go into those cities.
The first real 3D software I used was ArchiCAD at the age of 16, because a friend of mine gave me a version of it to use for a school project. The reason being that the teachers wouldn't understand what we did, so we could easily talk ourselves into a good note without much effort.
This of course didn't work out that way and I ended up working hours and hours to get something that looked somewhat like a building. And I actually enjoyed it. This was four years ago, and it took me more than a year from that point to stumble upon Blender. When I finally realized what kind of power I was dealing with, working on 3D projects became my main commitment.
I'm studying architecture for a year now and my passion for art is still growing. How I might combine my passion for art with architecture is still a question I haven't answered.
What is the process when you create a project like this?
My projects always start with a thought. Often a sense of mystery or excitement from something I get in touch with. Maybe imagination is a better word, as I mostly think visually with these things. In the case of this project, I was thinking about the concept of the future, and how we humans are capable of working hard on something that doesn't yet exist. We try to understand our situation to predict the future, so we can take measurements to prevent all kinds of problems before they even arise. Looking into the future is the base of our growth, and yet it's mostly a matter of guessing.
How do you visualize that? For me it had to be something powerful and a bit scary. What if something happens that turns your whole world upside down? When our guesses for the future all turn out to be wrong? Well that's what’s happening with the Corona pandemic right now. Building a future is all fun and fancy until it suddenly isn't. Until our vision turns out to be incorrect. What's left is an explosion of thick sight-blocking smoke, uncertainties and fire. Or well, hypothetically speaking, that's quite an accurate description for some of us.
I think this close-up is very powerful. This man is going face to face with his future, and inevitably, with his own death.
Of course this didn't just pop up into my head, and working on the composition and lighting in Blender it slowly developed into a clear idea. To have the most freedom while designing, sketching on paper is often the way to go in the beginning. But in this case I started out with simple blocks and lights in Blender. Because of the importance of lighting and composition, getting those aspects right in the beginning is crucial. Playing with smoke simulations less so, but it's definitely a fun part. First, I cut a sphere in two at the place of the explosion and used it as a hair emitter. Next, I simulated wind blowing through it from one side and used the result as the base for my smoke simulation. To get more volume, I added five simple spheres with an inward motion path and used the same wind to blow the smoke away from the middle. A simple blue point light creates a glow that lightens the scene.
Understanding how smoke reacts to movement helps a lot with coming up with these kinds of solutions.
Now this is just one, rather silly, way of making an explosion. But it resulted in something unique with just the right shape I needed.
A lot of time went into sculpting and making the materials, which I'll discuss later. Good to mention, I think, is the subtle use of objects from Megascans, who have an impressive library, free for all Unreal users, and in the highest quality there is. Some scattered rocks, a set of mushrooms and two vases did the trick for me, together with some of their rock textures. The man is from Renderpeople.com, who have a set of free models, and the headphone, phone and radio are from various 3D assets sites. There are many ways to add detail to your scene, so a little research is always important so you don't get stuck modeling everything yourself.
Several renders I made through the process.
At what stage did you think you had to look for an online render farm?
I've known about the existence of render farms for some time, but leaving the pc on through the night was always the go-to option for me. The tipping point? When I took a look at the render of the explosion in the morning and it was not even halfway done. It needed at least another day of rendering, and I needed to work. The convenience of a render farm is that in addition to the higher speeds, it also enables you to work on other things. Productivity wise it brings back the wasted waiting time to zero. That doesn't mean it's a golden tool. You can't just check the processed parts halfway through, which might be important in complex scenes with a lot that can go wrong. Luckily Blendergrid offers a money back guarantee, so you don’t get stuck paying for something you can't use. But it can still be frustrating. So instead of rendering everything in one go, I decided to split the scene up in two parts and rendered them separately. The smoke simulation was by far the biggest load, and because it didn't cast any shadows on the rest of the scene, I could simply render it on the render farm and stitch them together in post. These kind of solutions can require a bit of creativity and forward thinking, but can save a lot of time and trouble later on.
The smoke for example still had some problems after rendering it in high quality. But because it was rendered with a transparent background, I could simply retouch it with multiple layers and some fire effects without much difficulty. Of course this is much harder for animations and it's best to get your render as close to a perfect outcome as you can, but sometimes a simple retouch can save you tens of hours of fighting with simulations or in this case, give me an outcome that's near impossible to achieve with simulations in Blender. Knowing when to use a different tool might be the bottom line here, whether it's a render farm or fire brushes in Photoshop.
The edit process. As you can see, just stacking the image really helped with making the details visible.
What was the most challenging part of this particular render?
The most difficult part was to get the path to look rocky, while maintaining the somewhat strange shape. I wanted it to be stone like, but clearly something not out of our world. So I spent a lot of time sculpting, creating the demon-like creatures in the background and making the road look dangerous. Scattering little rocks on top of it helped with the realism, but what really made it work was the material set-up. I used the Procedural Noise Pack made by Simon Thommes, which I definitely recommend. There are a lot of options in the pack and it took multiple days to get my material the way I wanted, but without it, I’d have to rely textures because the way Simon uses nodes to make his pack is years ahead of me.
The stone structure has almost bone-like forms, but the material shows all kinds of cracks and rough surfaces.
Now that you are using Rhino for architectural modeling, how does that fit into your personal “cg pipeline” ? Will Blender still play a role or is Rhino covering everything you need?
The real difference between architecture software and most other 3D tools is that what you are making is not the real product, it's still just a concept when you finish all your work. The 3D model itself isn't art, it's just a representation of a building that might be worthy of being called art after it has been built. To sell your idea to the client however, you'll need to turn the model into some kind of art anyway.
This is one of the situations where Blender might have the edge on other software. The biggest advantage is that you can easily switch between Eevee and Cycles, where other archviz software often is either based on game engines or doesn't have the fast material edit options we have with the node editor. Of course there are expensive options like 3Ds Max and Cinema 4D, but considering Blender offers the same quality for free, with a huge market full of handy add-ons, it's definitely the better option for me.
Another reason to use Blender is in the beginning of the design process, especially because of the sculpt option. Sculpting is a simple, fast, and intuitive way of creating basic volumes and shapes. Combine this with an add-on that lets you import 3D Google maps data, and you have the perfect scene to try things out.
This was a great outcome for the design class that just started. Because of the quarantine measures, we couldn't go to the location in Amsterdam to get a feeling of the scenery we would be working with. With some effort I managed to get the location from Google Maps into Blender, and after clearing the building spot, I could simply export it into Rhino and send it to my professor, asking if it would be of any help. He really liked it, so now all the students can use it, in the software they are familiar with. And that's all made possible because of the awesome community around Blender. So yes, Blender can be helpful in many ways in the architecture pipeline.
Test render of the location.
Connect with Imre
To see more of Imre's work, check out his social media channels: