Jared Fischler did a breakdown of his amazing scene made as a part of a class at Gnomon. The main focus was put on visual storytelling.
Jared Fischler did a breakdown of Identity Theft made as a part of Christophe Desse‘s Texturing class at Gnomon. This time the main focus was put on visual storytelling.
Intro
Hi, I’m Jared Fischler! I’m originally from the East coast of the U.S where I studied at the University of Connecticut. During my studies there, I had the opportunity to participate in a number of exciting projects including pre-vis for a Broadway show and the creation of an interactive media wall at Boston Children’s Hospital. These experiences allowed me to get a taste for CG, which eventually landed me a job in New York City working as a 3D Generalist for a biomedical visualization company.
I spent a year in the city before I realized that my passions were more aligned with the film and game industries. I wanted to be in a space where 3D was used to tell captivating stories, and reels from companies like Naughty Dog and Blur Studios motivated me to apply to Gnomon.
It was obvious that something about the school inspired an enormous level of creativity and passion. So, I packed my bags and flew out to LA! I’m now in my 3rd term of the Modeling/Texturing track and I love it so far.
Class Goals
The featured piece was made for Christophe Desse’s Texture II class here at Gnomon, a class I was really excited for because it gave me an opportunity to create a more intimate scene that focused on the story. Storytelling with subtle details is really effective and a piece like this was the perfect opportunity to build something with a lot of beautiful detail in it.
I learned so much from Christophe’s Texturing class. He taught us how to take advantage of Substance Painter and Redshift’s speed to quickly iterate and therefore produce much better work by the end of the 10-week term.
The most valuable lessons I learned, though, were about how to work through creative ruts and keep my eyes on the end product. Through critique and guidance, Christophe helped me push my project to a level far beyond what I was able to achieve before taking the class.
The Concept
I decided to realize my own concept rather than interpret another artist’s work since there was a very specific emotion I was trying to illustrate. I wanted to capture a moment in time, as if someone had paused a movie mid-shot. I wanted it to be jarring, urgent, abrasive, and a bit uncomfortable. I love making work that pulls you in with attractive lighting and composition and then makes your stomach turn.
So instead of my usual process of hunting for inspiration on Artstation, Pinterest, etc, I decided to start by writing a narrative pass, which gives me a wealth of conceptual gunk to pull from while I’m developing the piece. I actually ended up referring back to that document up until the final composite in Nuke.
My narrative pass
Modeling the Assets
The Final Models
My initial blockout
For the face model, I began with a sphere in ZBrush and worked towards a likeness of Ofelia from Pan’s Labyrinth. While my character isn’t supposed to be her, I like to choose one main reference when sculpting faces to ensure that I end up with a unique face rather than a generic one.
ZBrush face and hull sculpt
After finishing the face, I sliced a chunk of it off and brought it into Maya to create the panel lines.
The wires were a fun challenge. Things like wires, pipes, and vines are such amazing storytelling tools. They can create order or disorder as well as lead the eye around your image. I used wires in this piece not only to show the audience how the head and computer devices were related, but also to create a vignette around the image to direct the composition. To model them, I started with a helpful little tool called FB Wire Jumbler found on HighEnd3d.com. The tool uses user-placed guide objects to direct the flow of generated curves which are then extruded into tubes based on user inputted diameter and diameter variation. I used the curves that the tool generated and re-extruded them with a few different NURBS profiles that I made. Repeating this process with some twisting and different NURBS profiles allowed me to create bundled wires really quickly.
Left: Guide objects and WireJumblers generated curves.
Right: My final wire bundle with marvelous designer sleeves
Onto Marvelous Designer!
I’ve actually been looking for an excuse to learn this program and this piece was the perfect opportunity. The workflow for the cable sleeves was actually fairly straightforward. After importing the bundled wire mesh into Marvelous, I dropped a rectangular pattern onto the wires where I wanted a sleeve, let the simulation run, and then stitched the ends together. After the cloth was loosely wrapped around the wire, I shortened the rectangular mesh in the 2D view, and let the simulation run again. Doing this allowed the cloth to look like it was really tightly wrapped around the wires.
Also, when running a simulation over geometry with lots of crevices, make sure you lower your particle distance! The default will likely cause the sim to go straight through some of your mesh.
Texturing
I started texturing early. I prefer to do this with all of my pieces alongside lighting and test renders, as it helps enormously when it comes to art direction. More often than not, the idea I have in my head for the final product is a fuzzy visual of lighting and the essence of the emotion I’m trying to convey through the piece. By blocking out lighting and texturing even before the models are finished, I ensure that I’m making the right decisions as I build out the scene. The quicker I can get half-baked ideas from my head to the model, the more fruitful my creative process is. This is where Substance Painter comes in! The smart material/smart mask system enables me to quickly establish basic textures on my objects so I can bring them back into Maya and see how everything is fitting together.
For the android head, I started with Substance’s Skin Face smart material. Having a base color, AO, SSS, and some micro height/roughness detail already established saves me the trouble and lets me get straight to adding character to the skin. After tweaking the smart material, my next step was to exaggerate the color regions of the face. This adds realism to the skin as it suggests different thicknesses of dermis/epidermis and also gives me a chance to add character and stylize the texture. I wanted this character to look vulnerable and sickly so I added paler areas along with saturated patches of red, blue tones, and a redder nose. Adding layers of brush strokes gives the piece a more stylized and painterly feel as well as helps to break up the uniformity that computer-generated noise and smart masks will create.
Next, I turn off the pore detail from the original smart material. I like to make my own pores with a combination of noise channels and the new skin scans that are shipped with Substance. It’s important that these layers are driving both height and roughness channels. On top of this, I add a fill layer driving roughness with a custom painted mask to make the face look more oily around the nose and forehead.
Skin Painting Process
Dirt & Wear Painting Process
Final Texture
Computer Assets. Substance Painter
Using Anchor Points
In the two screenshots below I’m illustrating a trick I like to use when painting height detail. After creating some wear with a fill layer and mask, I use an anchor point to reference that mask into a new fill layer. I then blur it and subtract the original. This creates a nice halo effect around the base. Using anchor points to do this allows me to change my original mask at any time and the halo effect will change dynamically. I discovered this trick when painting skin conditions at my old company, after noticing that most things affecting the skin had a surrounding layer of inflammation.
Wear detail with and without halo effect.
As you increase the original mask, all anchored effects are updated dynamically.
Computer Screen Workflow
The setup for the screens is pretty simple. I have a single sided mesh underneath that’s set up to be a Vray Light Mesh. On top of that I have a double sided mesh with a refractive material on it. I painted a diffuse and roughness map in Substance Painter to make it look dusty.
Two pieces of mesh used for the large computer screen.
Lighting
For me, lighting directs the creative process. I start thinking about lighting even before conceptualizing the assets. For this piece I wanted the story specifics to be pretty ambiguous. I wanted to give my audience a chance to be curious and come to their own conclusions to ensure that they could connect to and appreciate the piece.
The focus of the scene is the android head. Therefore, my entire composition and collection of assets were chosen carefully to justify lighting the head in a certain way.
I created a circular composition by placing the head in the left-most third of the image, followed by a larger and brighter computer screen, and ending with the smaller oscilloscope. To encourage the viewer to repeat this cycle, I completed the circle by adding the bundle of wires connecting the oscilloscope to the head.
I did this to direct my audience’s attention around the image before they notice the eye on the undamaged side of the face. To me this is the humanizing part of the character and the most emotive part of the entire image and it wouldn’t carry enough weight unless the viewer’s eye had traveled through the scene before landing on that section.
Conclusion
This piece was an amazing journey for me and a spectacular challenge as well. I want to thank Kirill for appreciating the care I put into this project and conducting this interview. I hope all of the 80 Level readers enjoy my breakdown!