Just Writing About Buying ZBrush…

So I have the opportunity to buy ZBrush, only I’m not certain if I should. It’s $699, $589 at Novedge, which is almost too expensive in my world, and definitely not a decision I should make whimsically, just because I think it’s the most amazing piece of software on the planet. I need to write this out.

Let me be clear. This isn’t some kind of pro versus con list. I can’t think of any cons. What I need to determine is whether I really need it or not.

My current project is a very extensive series of graphic novels. I want it to look like an average comic, the popular style, with inked lines and somewhat flat colors. It’s not what I’d call cel-shaded, but it’s not realistically rendered, either. The main reason for this style choice is I’m doing it all with 3D software and I can’t afford to spend hours every day waiting for renders on every panel. I’d like to get to the point where I can kick out two or three pages a day, a 64-page graphic novel a month. Only then will I be confident I can finish this project in less than two lifetimes.

To understand in more detail what I’m actually trying to accomplish, there’s a very simple technique for getting well-inked lines from a special 3D render pass I need to explain. I learned it from RogueLDR.

First, I set up my scene and apply shaders with all their material properties to my objects the way I normally would, only making the surfaces look flat rather than shaded because a photorealistically rendered object would look terrible with inked lines on top of it. It doesn’t have to look too flat, like with cel-shading, but it can’t look too real, either. Basically, I have to stylize everything. After I add lights and my scene is ready, I then render it at double the size I need.

Next, I create a special shader and apply it to every object. The shader is set to render the internal areas lighter than the external areas, all according to the perspective of the camera, i.e., what it sees. It’s called “incidence angle” and it’s not quite as simple as I’m describing it. For instance, the internal cavities are rendered darker, too. Anyway, when I have it all set up, I again render my scene at double the size I need.

Now I take both images into Photoshop. I assign the colored image to a background layer and run the “high pass” filter over the other one, adjust the levels and Gaussian blur it. I set the layer to multiply, resize it all down to the resolution I need, and the result is an almost finished image, fully colored and inked.

Of course, a professional inker would probably hate it, but I have weighted lines nonetheless and that’s about all I can hope for without any post-processing and clean-up. Also, considering the fact I have almost zero artistic ability (in the traditional sense), it’s really the best I can do for now.

However, what I’ve noticed is this rendering process leads to better results if the objects have a lot of detail. I can’t just place a high-res texture or bump map on the object, either; the detail has to be in the model for the best results. It will work, of course, either way, but it’s better with a lot of detail, i.e., bumps, scratches, deformations, etc.

This is where ZBrush is useful. It makes it possible to “sculpt” a model very quickly, deforming the polygons into much more interesting shapes, especially for my rendering process. For instance, I can model a perfect 2×4 with beveled edges that will look great in a normal, photorealistic rendering style, but it will look boring and flat in my inked and colored, non-photorealistic style. If I take that board into ZBrush and sculpt in wood grain and mess up the edges a little, now I have a much more interesting board. The kind of board you see in most cartoons and comics.

Is it worth $600 to achieve this effect? Absolutely. Can I do it? I don’t know, at least not consistently. In order to sculpt in all that beautiful detail, I have to raise the polygon count on my models exponentially. ZBrush is special and allows for insanely high polygon counts, but when I bring that model back into my primary program (Carrara), it is most likely going to crash or freeze a lot. I can be smart and hide any objects I’m not working on, but my file sizes will be huge and when it’s time to set up the lighting in my scene, my render times will be unmanageable. There are ways around all of this, of course. I could use normal maps and displacement maps (like video games) to fake high polygon counts, but they take a lot of time to set up and that would slow my workflow to a crawl.

Really, it’s an easy decision for me, because I see the graphic novels as a sort of proof-of-concept for a series of 3D animations and getting them done quickly and efficiently is more valueable to me in the long run. I want them to look as good as possible, but maybe low-res models and finished images with less detail is the best I can do.

For Offline Viewing, Download PDF Versions Of All “Just Writing” Posts Here...

I’m going to buy ZBrush. There will be enough instances when I’ll need it that it makes the purchase very much worthwhile. Also, if you can’t tell, I really want it. (Have you seen the new “Fiber Mesh” video preview?!) It’s just not going to change my workflow like I hoped it would.

Sunday, January 8th, 2012

Michael Island