There is a natural "rule" that can really help to guide you toward a masterful image, which is pleasing to the eye
To find out more about this exciting news please visit our new website.
Our blog has also moved to https://blog.therookies.co.
Hey guys, I'm going to give you a brief run-through of the process I typically go through when making an environment piece. I appreciate the interest, so I'll try my best to make an awesome tutorial for you all. First I would like to acknowledge the concept artist who did the original artwork, Jens Holdener.
As a 3D artist, whether, in production or concept, it's good practice to have about three or four of your favorite artists in mind when you work. Take note of their workflow and what they do to be awesome. Some of my inspirations are , ,. I make it a point to allocate any tutorial on Gumroad or Gnomon Workshop I can to really study their process and what they do to be faster and more efficient.
This sounds pretty rudimentary, but spend some time just kind of analyzing and studying the concept you're working from (if any). As a production modeler in training, it's kind of ingrained in me to be like a copy machine when it comes to learning the artwork that I'm going to be spending the next significant amount of time over, so I like to look through it and figure out what really makes it rad. In my case, I really dug the pilot chairs and GUIs, the tracks on the floor that hold the chairs in place and allow them to swivel back and forth, so I noted that in my head that these aspects would be important to nail down later on.
I'll go into Photoshop and start making some notes on materials, inventorying all the little parts and details I may have to model, deciding which software to approach some of the parts like cloth or foliage. During this phase, I'm always in Photoshop or on the internet and haven't opened Maya yet. I'll scribble notes and doodle all over the concept piece, usually saving a copy to keep alongside the reference as I'm working. I'll end up keeping a copy of the scribbled notes, the concept by itself, and an inventory list in the same folder so I can flip through them easily in something like Windows Photo Viewer.
Still having not touched Maya, I like to go on the internet and spend about 3 to 4 hours just scouring for cool tech-related directly to my concept.
So in the case of the Dead Space cockpit, I'll spend some time researching and looking for cockpit chairs from say, an F-15 or C-130, grabbing pics of the actual chairs with all the belts and straps, and then some of the smaller parts of the chair, like the bolts, grips, fasteners, etc.
I'm also taking note of how everything seems to come together and operate. I like to make a couple of pages of collages of holistic parts, then the smaller parts, getting multiple angles.
Lastly, I'll get some separate pages of just material callouts. Now, you may think that hours and hours of reference gathering sounds tedious, but trust me, the work really pays off and is worth it's weight in gold.
Ok, now you can open Maya and set up your project. If I had to choose any one aspect of what makes a cool art piece, be it a classical painting, illustration, 3D still image, or whatever, the single most important thing that will make or break your artwork is the composition. You can have the most insane modeling, with the most intricate of details, but if your composition is weak or just kind of so-so, then your final product is going to be weak or so-so.
Set up your image plane in Maya from a created camera. Make very simple shapes at this point, nailing the perspective of the piece by lining up the grid to the ground from the image plane, while matching the first cube you modeled to a reference point on the piece. I immediately picked out the foreground chair and made a simple rectangle. I like to drop a real-world scaled human basemesh into the scene to accurately scale the other objects in the scene, and I'll put this guy on his own layer, to add and remove him. You'll end up keeping this simple geo on it's own layer, so that when you're ready for the actual modeling phase, you can easily hide or even dump it from your scene.
Please note that you can also adjust the focal length in your camera accordingly to match the perspective. A higher focal length is like extending the lens of a natural camera but has a flattening effect on the image, while lowering this same number shortens the lens, bending and stretching the perspective a bit. Use this to your advantage. I usually lower the focal length to around 35, sometimes even below, to make the perspective just a bit more dynamic. You want to spend time and energy really nailing your composition because this is the foundation of making your rad artwork. Lock your ShotCam once your perspective is set, so you can use this to build into your final piece. I have a bad habit of moving my cameras, so this is essentially protecting me from messing up my perfectly set cam and ruining all the hard work.
This is the magic behind the image. There is a natural "rule" that can really help to guide you toward a masterful image, which is pleasing to the eye. Stick to a mantra of using this for every one of your images, and you'll be golden (pun intended). I've illustrated from this concept piece I'm working from how the artist has employed the Golden Ratio to guide your eye where he wants you to look. Notice how the vignetting narrows down your field of vision while the natural flowing curve in the center tells you what's important.
For the next phase of my process, I like to save this blocked-in scene as a new iteration, then start a new scene.
This is where the fun starts. Looking at your reference, decide which pieces from your collage best resemble stuff from your scene, and start to model the smaller bits that comprise it. Your goal is to build a library of surface details, bolts, springs, joints, cables, etc. that you can then use to decorate your scene.
This library is more commonly known among hard surface modelers like Vitaly Bulgarov or Gavril Klimov as a kit. It's never a bad idea to UV the parts as you make them, so that when you're later duplicating several different bolts or surface details, you can pretty much skip the UVing process altogether and go straight into lighting/texturing.
Even though this is a futuristic space concept, note in the reference collage I've made the bottom right images of the F-15 cockpit chairs.
It's much more beneficial to your piece if you stay firmly grounded in reality in your reference process, to derive such ideas into your work.
Take a moment to really check out the cool details of something that actually exists, all the essential straps, pulleys, bolts, etc. that go into making this highly functional for a fighter pilot.
These little details help sell the believability of your artwork actually existing in reality.
For the joystick as well as the chair and the consoles in their own scene, to avoid overcomplicating my workflow. I find this to be a good general practice, and it really gives you the flexibility to not only model in a very light scene, but to have full focus on any given model to then be exported into a "master scene." The master scene is where I dump everything, but rarely work inside of unless lighting and rendering.
Using a new scene for established sections of the scene, continue to model and refine from your block-in. I like to add in the kit parts that I made from my library toward the end. Always have not only your concept clearly visible at all times, but your reference collages. I like to have at least 2 made, one for overall pieces, and another for the smaller bits, like bolts and hydraulics. Marry the two together and your piece will take on a believability and life of its own.
Right now I'm blessed to be attending an amazing school like Gnomon, where I can always look to either my fellow classmates, who are talented artists, or Alex, my demo reel instructor and the director of Gnomon. Your classmates can see your work with a fresh set of eyes and make suggestions as to what they'd like to see or what they like. There's an old saying: "steel sharpens steel." This holds especially true in the CG community.
Don't be afraid to get a good honest critique, as this is gold when it comes to helping your work reach that level of awesome that you want. Avoid asking loved ones or your mom. They're always going to say something like "looks lovely, dear," whether it looks great or it's total crap.
Related Link: Gnomon School of Visual Effects
Start with a primary light, usually an area light, with "Use Light Shape" checked on, and I like to go with either rectangle or cylinder, depending on the effect I'm looking for. In this case, I made my primary light a rectangular Area Light. Since I'm rendering in Mental Ray for this project, I'll go with a physical light setup. In this diagram, I've shown some notable things I do to tweak my Area Light, the Key light in the scene. I change the shadow color to a dark grey which is slightly higher than black, turn raytraced shadows on, check Use Light Shape, and in custom shaders, I'll click the checkered box for Light Shader options, go into the Mental Ray Lights set, and choose "Physical Light." For the physical light's color in it's respective subsection, I'll change that to mib_cie_d, also in the Mental Ray Lights section.
The hidden power behind this setup is that now you have a physically based light that has a temperature control and intensity slider. the higher the temp number, the cooler, and visa-versa. The intensity is usually a number in the 10 of thousands. I typically choose something like 50,000, render, and keep tweaking this number as needed. Your goal for your key light is to have a prominent coverage and strong shadows, with no blown out areas, similar to real photography.
My shotcam has a similar physically based setup attached to it, as shown in the diagram. In a more dramatic piece like this, typically you would go with a focal length of around 35mm (think camera lenses for this number).
A little trick I like to use is to lower this number slightly, sometimes into the twenties, to twist the perspective just slightly enough to make it stretch and twist just slightly, giving a real sense of movement and dynamism. I always set my far-clipping plane to an obscenely high number like 5 million, in this case.
In the Mental Ray tab toward the bottom of the attributes, I'll go into the lens shader and attach a mia_exposure_photographic lens, another physically based lens. The cool thing about this lens is that it operates very similarly to a digital camera like the Canon 5D, where you can adjust things like the F Stop, CM2 Factor, Shutter, etc.
I find this lens incredibly simple to use with a little bit of basic photography knowledge.
When I'm doing a big environment scene like this, I really like to streamline the material workflow into an uber setup, which is why I typically lean heavily toward just mastering the uber shader, whether it be Mental Ray's mia_material_x (not a fan of the MILA material in 2015 because it seems to be, well...broken) or Vray's vray_mtl. With metals I like to crank up the IOR settings to a high number like 60, also turning Use Fresnel on.
I know that sounds crazy, but I seem to have perfect control over what my spec map is contributing into my material at that point. This is my setup. Always remember to check Alpha is Luminance for your spec map or it won't work. Basically, without this check box, your material will look flat.
In Mental Ray, setup is pretty easy as long as you keep a few things in mind. Your reflections refractions settings should always match the numbers you set up in your lights. For this scene, I only used Final Gather and not Global Illumination, due to the materials used and the simple light approach I employed. There simply wasn't enough need for GI that reflections or refractions couldn't already provide, so I left this unchecked. If there are things you can do to save on render time that doesn't hurt the quality of the image, then do them.
I went with Unified Sampling for this one, where in Vray I'll go with Adaptive DMC. I always use the Mitchell filter; it always seems to provide a nicer, faster render, and raytracing values match those of my area lights in the scene. Last but not least, in the Indirect Lighting tab, I usually tweak the accuracy, seeing how low I can dip without deterioration in quality of the image, as this saves in render time. Check both extremes, going higher to see how it's affecting render time, lower to see how much render time you can potentially save. You'll get the hang of it once you've tested this slider a couple of times, don't worry.
Once you've made these adjustments, just go through the scene, blocking in materials and rendering at your lowest settings, tweaking your spec map values (Reflectivity and Glossiness) as needed to achieve the desired look. Once you've got a solid working render and the materials are working well with the lighting, it's time to take your passes into a compositing package like Nuke.
For this piece, I used one of my go-to Nuke scripts, which contains a simple Grade, Color Correct, blur for the background planet, which then gets combined together and piped into a noise node, where I put a fractal noise over the entire image to break up the perfection of the CG render, and turn the opacity down to where it's barely noticeable. I'll spend the next 3 or 4 hours...minimum, tweaking these values, looking repeatedly at my image to catch the vibe and mood.
I consider this phase, compositing, to be one of the most important in really separating your artwork from a common CG render. This is your chance to become a cinematographer and wow your audience. Put some love into this and, provided the previous steps were golden, you'll be well on your way.
The bottom portion of my script has a section where I fake chromatic aberration, which is an effect naturally caused in camera lenses which;separates prisms of light. I basically offset the green channel from the red and blue, which are on their own nodes, to create this illusion.It's kind of a cheap version of more elegant chromatic aberration scripts which I'm sure are out there, but this seems to get the job done for me. Once I'm happy with what I've output from Nuke, I'll go one last time back into Photoshop, where I'll add very subtle details like lens dust from an image. I also use a plug-in within Photoshop called Magic Bullet, very subtly adding things like diffusion, flares, and vignetting to really polish and refine my compositing.
After hours upon hours in the final stages, tweaking values, going back and forth between Nuke, Photoshop, and Magic Bullet, I like to take an hour break where I try my best not to look or think about my project. This is done very deliberatly to ensure that when I return to it, I have a fresh set of eyes with which to critique my piece and make more changes. Bear in mind you've probably been slaving over this piece for quite a while, so things start to look the same, and it's harder to pick out mistakes if you make them. Do yourself a favor and take breaks.
I hope you've enjoyed this tutorial over the Dead Space Cockpit, and found the information in some way useful. By no means am I saying that my approach is the correct or best way to go about this workflow, but these methods seemed to give me good results in the end. Never stop learning, and take care guys!