Rendering Tips: Combining 2D and 3D

Introduction and Career

I’m Sergio Castaneda, born and raised in Mexico and I’m a Concept Artist at One Pixel Brush, working for the videogame industry. I’ve been working on projects like The Last of Us Part II, Call of Duty, Arena of Valor, among others. I got my foot on the door last December when Shaddy Safady discovered me on Facebook and invited me to work at OPB. Before that, I worked on local studios doing other stuff like advertisement videos and motion graphics. I always had this dream of doing AAA games, but it wasn’t until last year I decided to go full time on it and started to work on my portfolio.

When you are working on personal projects, first of all, you have to be sure what you want to achieve, and you have to be excited about it, motivated by a painting, a movie, a photo, a memory or even the lighting entering through your window. For me, the composition will be the most important thing at the beginning. I love starting with simple black and white thumbnails, this way you know if your composition idea is working or not, if it’s worth the 20-30 hours you’re about to spend on your piece. I gather references by grouping: Subject Ideas, Light, Details, Mood, and Final Look. Once I have a bunch of decent thumbnails, I pick my favorite and I’ll stick to it, its shapes are locked from now on. Then, I proceed to create the scene in 3D.

Jumping on the technique side, after all the 3D work, an important step is to use 3D passes, Zdepth pass is one of my main tools, can give you atmosphere, depth, dark foreground, all of this following your thumbnail values. Also, Geometry Normals and Material ID will save you so much time, when you want to select certain areas.

The Workflow

I’ve been into 3D since college 12 years ago, so I’m very comfortable at this point. 3D is the best way for me to start my pieces because it gives you so much freedom! It is so flexible in terms of lenses, lighting, materials, moving objects around, etc. but it also can be a deadly trap if you don’t know your end goal, as I mentioned before, and you ended up wasting hours and hours, so frustrating. Therefore, I love my 2D thumbnails so much, because I can be sure my effort is going to pay off at the end. My rule is 90% 3D – 10% 2D. Adding photos as textures and little touches of paint over are easier than if I decide to go just pure brushes in Photoshop for a client piece, I will take so much more effort to do a decent result, and I’m risking too much, because maybe the client wants a different angle or lens, and there’s no way I can achieve this in the few hours I have on a regular day.

Working with Assets

I bought most of the assets but also get tons of free ones from 3D stock websites like Turbosquid, 3D warehouse, 3Dsky or Megascans. I download every model that is related to my image, but I ended up using maybe just 30% of them. I think the trick is in the mixing, you have to be clever of what you can mix to get a uniform result. I rarely use them as the original model, most of the time I just use the base and modify it or extract some parts to mix them with other models. This process is known as 3D Kitbash as many of you know.

Rendering

C4D and Octane Render are my main combo. For me, it’s all about having a procedural-nondestructive scene. C4D gives me infinite variations, when I need it: trees, rubbish, clouds, grass, shrubs, etc. Just one click on the random seed and I can have a totally different scene, this allows me to look for natural randomness I need in a matter of clicks. Or I can set up an interior set -walls, floor, doors, windows, moldings- with NURBS extrudes and sweeps that follow just one spline, and all of them will follow the direction of just a simple vertex when I move it. Octane, on the other hand, has this powerful real-time path tracing engine that makes everything the light touch looks gorgeous! You can see how light changes the moment you move it. My general method is to start a 3d scene with the big shapes of your image, it can be the room, terrain, a tree, mountains, etc., and then I go for medium shapes, then small, and finally details. Same with the light, start with one big light – the sun, for example, then add complementary area lights if necessary. For highlights and bloom, I prefer to do it in post-processing with Photoshop.

About the Eytan’s last Gumroad course

The 2D Paint tool from 3D Coat was a big discover! I used it to explore quick ideas and turn out great! Also, Eytan’s has a great method, when it comes to thumbnails mixing 3D and 2D. 

Afterword

I think photo-bashing is a great method, too, for quick sketches it is great! And, of course, if you already master it, you can mix it with 3D to aim for a quicker process. I keep using photo-bashing techniques as a post-process pass to get rid of the “3dness” of my raw render. In my opinion, it’s hard to compete with the freedom that 3D gives you, as I mentioned before, you can pay as much attention to details as you want. If you come from a photo-bash workflow with zero 3d experience, just try with big boxes and throw a big light on it, play with the shapes that lights and shadows give you, and when you’re happy with the composition just, hit render, and you are ready to go, start photo-bashing your 3D base. If you are a more experienced 3D user, you can try something more complex, download a bunch of 3D objects and mix them thoughtfully, with intention. Remember guys, it is not about the models you bought or download, it is about your story, how they serve to your purpose.

Sergio Castaneda, Concept Artist

Interview conducted by Kirill Tokarev