WIP Italian Deli in UDK

tonysladky

Over the past couple months (trouble staying focused and motivated, frustration at overscoping), I've been working off-and-on on this environment. Now that I'm into texturing lighting, and coming up on my Shader R&D phase, I like to think I'm close enough to done with (hopefully), my first good environment that I can start fishing for critique.

Going back in time to the early stages, my early Hi-Res props:







Middle of summer, reducing, UVing and baking (all glass is UDK glass material as stand-in until I begin shader R&D yet to build my own), and preliminary lighting and layout:






Preliminary layout and lighting before cutting down the scope:




Where I'm at now. Rough textures, more finalized layout:

(God, that is a horrible capture on the bread slicer...)

tonysladky

Updated layout:

Here's where I realized the need to rescope due to too many assets repeating too often.




Adding in the architectural elements (molding, rafters etc.). Those emissive planes at the end of the rafters are eventually going to be decal textures detailing where the rafters intersect with the bricks and whatever I decide the back walls are made of:


Edit: The molding in the corners is a known issue. I haven't rescaled it since raising the ceiling, hence it cuts off too early.

Crit would be very appreciated, especially any modeling crit since I've gotten a few surprising* negative comments regarding my modeling on my blog, and I figure if anyone can be impartial about whether my modeling technique is flawed, it's a community that's almost exclusively focused on prop and environment modeling and texturing for games.

*surprising in that this anonymous, though well-spoken (ergo, obviously not the usual blog comment spam) commenter was the first to ever accuse me that my modeling technique is flawed to the core, a crit I had never heard from faculty, classmates, or portfolio reviewers at GDC, who, busy though they may be, would have no reason to attempt to defend my ego by mincing words.

pixel scholar

The important thing is that you are, in fact, modeling. It's easy to adjust one's technique to fit the requirements of a particular engine or studio as long as one continues to work and develop as an artist.

The only problem I see from a quick glance is that (and I might be wrong here, others please correct me if I am)the circular elements in your models are capped with tris that all originate from a single vertex right in the center which, I was always told, is an inefficient way to model in engine assets due to the fact that that sort of an arrangement would cause the engine to have to calculate that central vertex over and over again.

That was a few years ago though and the way UDK handles polys may have changed. Just wanted to throw that in.

Keep modeling!

tonysladky

Thanks for the tip. I usually try to avoid those just because it still shaves a couple tris off the overall tri-count. I think you're mostly seeing that in the Hi-Res models because I neglected to get any screengrabs of my Lo-Res wireframes. I know I've still got it going on in a few places (I wasn't too concerned about shaving a couple extra tris off the <500 Tri tables, for example). There's probably also a couple spots where I kept the central vertex in place so that I could have a not-quite-flat top on some of my cylinders (knobs on the meat slicer as kind of an extreme example, but I think most of where I'm doing that is on the hinges and hydraulics of the door).