Hey everyone!
Thanks for being patient with me for being a day late on the progress update! I had some technical difficulties that needed solving in order to get these character renders up to spec :D. This post will actually be a three-parter, as again there’s a lot to talk about as we push on with v0.06 and edge nearer to the engine upgrade and various overhauls we have planned. The fun bits are spread throughout, but I feel the technical details are also important to give you an idea of the growth the project will be experiencing after v0.06!
Our new artist, Ubercharge, has introduced us to a lot of new technology this month that opens doors for us both in terms of improving quality as well as improving efficiency. Many of these things I had considered but seemed over my skill level, however with his guidance TK and I have learned a ridiculous amount about AAA workflow and CG production in just a few short weeks. Admittedly, this all made some waves as we had to pause to orient our methods of development for v0.06 in light of what will be a lot of improvements in software/workflow soon thereafter during the aforementioned artwork overhaul, but overall, it’s been an incredibly productive month!
July Achievement Overview
Here’s an overview of our achievements for July. What I don’t go into detail on today will be in the posts tomorrow and the next day!
Transition to Redshift
Sorry to hit you straight out the gate with technical stuff, but this will help explain some of the dramatic improvements in visuals you’ll be seeing throughout these posts! The Inner Circle guys have already had to suffer through most of this :D.
One of Ubercharge’s specialties is being able to quickly create physically realistic materials. He does this using a physically based renderer (PBR) called Redshift. He has pretty easily convinced TK and I that transitioning from Mental Ray to Redshift for map creation is a good move, and with TK’s help he already has the basis for TK’s map creation workflow working with it. It has impressed us enough that by the end of the month we had decided that Redshift will most likely be our sole rendering engine after the artwork overhaul.
Obviously, our game’s art is highly stylized, so why should we be looking at PBR? Basically, stylization doesn’t mean the materials have to look simplistic or that lighting has to be flat. Overwatch is a great example of a stylized game that uses physically based rendering yet maintains its own look. Ultimately, the switch will offer room for growth we don’t have with our current technology.
Here are some of the potential benefits of switching to Redshift:
That’s not to say it won’t take some tweaking to get it to fit what we want to accomplish, but testing in this regard has been one of the themes throughout our work this month. Here’s a simple lighting and environment test render I came up with during my experimentation with the engine. Not bad, yeah :D?
Nvidia Iray Rendering Engine
At some point last week Ubercharge gave TK and I a two hour presentation via Skype on physically based rendering to sell the idea of using Redshift for maps. Afterward I felt more confident with trying out Daz Studio’s included physically based render, Nvidia Iray. I had originally tested it out with the intent of figuring out some good liquid materials, but got to thinking how much it’s been bothering me that our map graphic style and character art style have been diverging since we began doing maps with 3D. I went ahead and ran some tests on the characters with Iray and discovered I could quite easily get something very near to our character art style that doesn’t require any post processing. In that example, the image on the left was done using our old render and required compositing and post processing, whereas the image on the right (a relatively quick practice render) was done using Iray. At that point, I showed what I had to Uber and he was excited enough to download and learn the basics of Daz and Iray over the next day or so. Here’s a set of images he captured while testing out its capabilities.
Using this knowledge as a starting point, TK and I conducted further tests and learned about its built-in render management system. While initially successful, we found some major bugs with Daz’s implementation of the engine later on that would spark my interest in converting our characters to Maya to see about using Redshift for all our rendering purposes (we’ll talk about that in the post the day after tomorrow). While that’s all well and good, it won’t happen until after v0.06. In the meantime however Uber has already been helping me use Iray to refine our artwork style by focusing more on improved materials and advanced lighting and reflection techniques. An example of this is our newest work-in-progress character, which I’ll talk about more below. Using Iray for v0.06 artwork will be a great learning experience and transition for me as we develop the style further during the artwork overhaul.
Iray offers us a lot of benefits that the previous rendering engine, 3Delight, didn’t have. All of these (non-Daz related) benefits and even more are available in Redshift, so they apply there as well. Some of these include:
If you’re wondering how those liquid material tests turned out, check back tomorrow when we talk about our liquid simulation experiments!
Our Newest Character/Enemy Design!
Finally, time for some eye candy! The artwork has been updated since our Inner Circle preview by the way!
I tossed around the idea of adding a female police enemy in to v0.06, and went for it after I realized I could work in a neat little story event as an introduction. She’s still a work in progress, but I’m closing in on a final design; most of what’s left is improving the hair, adding her weapon and adding insignias and text to her outfit. Here are the latest high-resolution images!
About the Artwork
So, you’ll notice straight away that the art is different for this character. That’s because this is the first character I’ve created entirely with the intent to use a physically based rendering engine. I had expressed concern to the IC crowd over mixing art assets created between the new/old rendering engines for v0.06, but as I mentioned above I figure I’d rather get the experience and have the opportunity to develop the new artwork style in advance of the artwork overhaul than to simply strive for consistency with an artwork style we are probably growing out of anyways.
She is in a lot of ways an experiment due to having to create new PBR materials and to test various lighting situations to see how they behave. We’re really trying to avoid the use of cheesy post processing filters post artwork overhaul (such as heavy outlines without the stylized assets to support them), and one area this makes things a bit difficult is layering portraits onto backgrounds or map artwork and not having them clash. When everything gets run through a filter it makes things mesh quite nicely, but with this style we have to rely on informed lighting choices and more advanced techniques.
Some things I think are obvious improvements are the possibilities for improved detail in materials, Fresnel reflections (this is a staple in getting an animated look out of modern 3D), accurate translucency (see the light shining through the hair), and way better skin due to subsurface scattering.
Keep in mind the style is still in the early stages and will probably develop quite a bit, and also that this is simply one example, but definitely let me know what you think so far! Oh, and just to clarify, the “new style” is entirely about the rendering, lighting, materials/shaders, and processing methods, and has nothing to do with body shapes or character anatomy.
That’s it for part 1! Check back tomorrow for AltairPL's update, Ubercharge's map artwork progress and more!
Faithel
2017-08-05 02:14:33 +0000 UTCSubjectivity
2017-08-04 01:18:00 +0000 UTCtsukiouji
2017-08-03 23:11:38 +0000 UTCNot your bussines
2017-08-03 10:31:43 +0000 UTCblabla
2017-08-03 08:32:37 +0000 UTCFallenbur
2017-08-03 07:07:49 +0000 UTC