WWW.BLENDERNATION.COM
Behind the Scenes: Discoveries
Behind the Scenes: Discoveries By Alina Khan on November 21, 2024 Behind the Scenes Dive into the sci-fi world of Alexander, an Austria-based concept artist, as he illustrates his work with special texturing techniques.INTRODUCTIONHello, my name is Alexander. I am 35 years old and based in Vienna, Austria. I picked up Blender and various illustration software such as Photoshop about two years ago. So, due to my actual professional background in the health sector, I did not have any training or education in the 3D industry. Having the opportunity to introduce myself and show the creative process behind my renders is a great honor for me.But enough about me now, let's dive into the project! Thank you in advance for reading through my process.Enjoy!INSPIRATIONThis image was the result and my entry for the monthly Blender competition on their Discord channel. The theme was Machinery. By now, I have a structured workflow that I follow every time I start a new 3D project. The first step, like the headline already says, is finding inspiration. A really good habit to develop is to browse art pages and media sites on a daily basis. By doing that, you will expand your visual library, which makes the process of creating a basic layout/composition later on in Blender much easier. So, places like Artstation, Reddit and Discord servers are a great way to start.Personally, when I read the theme, I immediately thought about some sort of sci-fi-like environment. So, with a basic idea in mind, I started to collect some references, most of which were actually screenshots of two games I played very recently: Dead Space Remake and The Callisto Protocol. Games in general are a very good source for getting some inspiration as well as references since the environments are already there for you to observe! So, take advantage of that!After gathering a bunch of screenshots and references from the web, I put them into PureRef. Its a godly piece of (free!) software that lets you build a moodboard for your projectshighly recommend it. So now, it was time to open up Blender and start blocking out my first ideas!PROCESSHardware Specs: Nvidia Geforce RTX 4080 Intel i5 13600K 32 GB DDR5 RAM @ 4800 MHz Software and main Addons used: Blender 4.1 Plasticity Photoshop Random flow addon Box Cutter / Hard ops addon Mesh Machine addon Botaniq addon Fluent Materializer addon Creating a composition:Before hopping into modeling assets and texturing them, the first thing I usually do is get an idea of my composition by blocking it out roughly and setting up my camera. In this render, I knew I wanted to create a big environment so it was clear to me that my camera will have to be tilted upwards with a low focal length (I go for around 23 mm for large shots). I usually display the thirds in my camera view to get an idea where busy spots and focuses should be as well as spots where the viewers eye can rest.This process is very important in my opinion, because focusing on too many things at once (composition, lighting, modeling, etc) can be overwhelming. Also, with a basic layout of the scene, its much easier to place your real assets later on. By doing this, you can already get a sense for the scale of the environment and for the assets themselves. So, a good thing to do is to always implement some human-size reference cubes in the scene, as seen in the blockout screenshot. I always model my assets to human scale, because faking size can be very distracting for the viewer in the end result. For example, just scaling a railing up will look not natural."Basic layout/composition of the scene"After the blockout phase is done, I move on to the basic lighting setup. Here, I try to find cool angles and shadows which will make a nice atmospheric lighting, combined with volumetrics. I usually go with 1 strong light source in my renders (for indoor scenes, of course its different). With volumetrics in your scene, you can also create a really good feel for distance and how big everything appears to be. The settings of volumetrics have to be quite low though, otherwise it can go south really quickly. I also tend to place the volumetric cubes a little bit in front or further away from the camera as it affects the image overall, when the camera is also inside the volumetric cube."Initial lighting setup, changed later in the process"Modeling the assetsWhen it comes to modeling, I basically have three techniques that I use. Most of the time, I start with primitives and try to implement interesting shapes by cutting parts away, mirroring, adding parts and so on. A very good tool for that is Boxcutter/HardOps which makes Booleans very easy to use, and the Twist 360 modifier is an absolute time saver, giving really nice results.For the main ground asset, I chose a machine called Hydroponic. I picked up the idea from the video game The Callisto Protocol and really liked the design of it so I made my own version."Face mode of the Asset"As mentioned before, my modeling mainly consists of simple shapes where I cut or add certain things to get the look I want. The mesh is not necessarily a clean quad topology but when you do still images and do not have to morph the mesh in any way, it will render out completely fine.The second method is done with HardOps and its amazing feature, Twist 360. It allows you to create really interesting shapes and forms within a few clicks. While the base mesh is editable, you create more and more different assets of a very simple base structure, for example, this walkway.Taking it from thisTwist 360to this, in a few clicks! Amazing, right? Twist 360and even more interesting shapes and forms!So, this somewhat unorthodox modeling method allows you to come up with lots of cool models in a few minutes, which you can then use to greeble up your main assets or big structures, to create a massive amount of detail on the one hand, and on the other to show the scale of your large objects even more!The last method is making use of an addon called Random Flow, also highly recommended. With this addon, you can greatly increase the big, medium, and small elements on your base mesh. In this example, I used it to create a very detailed structure that ended up being the big thing in the final image."Selecting the faces you want to be affected by the addon"After selecting the faces of the base mesh you want to alter, you can start working with random flow.As you can see, my initial base mesh was pretty simple (some cubes, cylinders, extrusions, and bevels on the edges). After applying the addon, the mesh is detailed with larger junks and smaller parts as well.TexturingWhen it comes to texturing my objects, I have three methods. They are procedurally generated; I use photo textures from textures.com most of the time, or Quixel Mixer for grounds.With procedurally generated materials in Blender, there is an amazing addon called Fluent-Materializer. I use it a lot when it comes to rusty, grungy sci-fi objects, and it gets the job done perfectly. Setting up the node system can be a bit tricky at first, but with a little practice, its smooth as butter. You will achieve a very detailed, realistic look that can be shown on any hero asset, and it will look very nice!"Hydroponic, textured with Fluent Materializer"and the results speak for themselves!"Bridge asset, textured with Fluent Materializer""Main structure textured with Fluent Materializer"The second method I constantly use is using photo textures and projecting them onto the mesh. This is very advantageous because your model does not have to have every detail modeled in or dozens of decals. Instead, the juice comes from a 4K, sometimes even 8K texture (when very close to the camera). The node setup compared to Fluent Materializer is very easy and simple."A wall asset from a different Project, textured with photo textures"The model itself is just a cube, with some unevenness modeled in to make it look more organic and believable.The third method I use is mostly for the floor/ground only. Quixel has a piece of software called Quixel Mixer. Basically, it allows you to mash up lots of different textures from their Quixel Megascans library. You can then adjust everything to your needs (be it displacement, roughness, metallic, etc.) and even add puddles of water. After importing into Blender, you can work with the displacement modifier to apply the displacement onto your mesh (subdivide your planes a bunch of times). The only downside of this method is that its very heavy in GPU load due to the displacement. So, I usually end up making one or two and alt+d ing them around the scene. Also, you should decimate the plane after applying the sub d and displacement. Another cool trick is to make the albedo, roughness, and normal smaller beforehand. There is an awesome free software called Infranview that gets the job done perfectly. You can reduce the size from 100 mb to around 6 mb within a few clicks. Just save it as a .JPG instead of the heavy .PNG files. The only thing I do not make smaller is the displacement map itself, cause thats where the juice is."Quixel ground asset"From Quixel Mixer"Imported ground asset into Blender"into the project, it looks fire!After I was finished with modeling and texturing, I went back to the blockout and started to arrange my assets in the scene. I usually end up showing a few compositional guidelines in my camera view, such as the thirds and the golden ratio, to get a rough overview on how to and where to place my objects. There are a few little tricks you can do to lead the viewers eye where you want it to go: Rule of thirds (place your main objects into that area) Do not split the image in half (rather, pick a side). Have strong leading lines to guide the eye. Cluster detail Include scale references or shapes in your image that the human eye can easily relate to. Have breathing room (e.g., empty space in your scene) to give the eye a little rest. Lighting sources should be placed not randomly but with a specific purpose (to show something important or to lead the viewers eye towards something). Use volumetrics (either in Blender itself or render out the mist pass alongside for further processing in Photoshop or any other software), but use volumetrics carefully as it washes out your textures when you overdo it. After rendering it out in Blender, I took it to Photoshop for color grading, lighting adjustments, and adding dust and some particles floating in the air. A really neat way to get rid of the 3D-ness is to overpaint the image. I usually go with the mixer-brush tool in Photoshop, and it gets the job done perfectly. Paint over the sharp edges to soften up your image and give it that slight painterly lookit will make a considerable difference!Below youll see the final image and a clay render alongside it.RENDER: Discoveries"Final image after post-processing""Clay render of the final arrangement"Thank you so much for letting me have this opportunity to guide you through my process of creating my 3D art. I hope you enjoyed the small voyage.Cheers, Alexander!About the Artist Alexander is a 35-year-old radiographer and hobby concept artist based in Vienna, Austria.Links
0 Commenti
0 condivisioni
10 Views