Subscribe to download
byteframe12-Trashy_Turnstyle

This is my opus! Well, maybe not so much as I finish. It's also a test of my pirate asset packs. Originally envisioned as a large sprawling indoor storage facility, without an accessible outdoors, in the city, it was going to be filled with different kinds of doors, powered by point_value_remapper. Players would explore each themed room, and doors would unlock sequentially. I was going to have one with the Primer time machine in it, and a Breaking Bad pile of money, so please try to imagine that.

As for the PVR door, it seems to be bugged when using AnimationCycle in steamtours (dunno about hlvr), although I got a decent neutered version of the Aperture door to function (with minimal vscript), and it is featured once in this level, but all that discouraged me. Then, I was like, 'how will I light all the rooms?', and planned to all have windows for the sunlight, and think of ways to get hacky with lighting (I probably would have learned alot). So that's how it started, and a screenshot shows this, but then I got obsessed with the ground/road mesh template and it took on a life of its own from there.

My mapping efforts consist of obsessing over a few pieces of geometry/templates at the start (the zen at this stage is my favorite part) before I bridge the whole space together and fill it up with stupid props. So, not much new it terms of overall technique here, although I took my time with this one, and I think it shows. It's still just another silly open space VR diorama, when I want to someday soon want to make a cool house with rooms and stairs and doors and stuff. That might end of being a better application of all the doors and cabinet interactivity work than the storage facility idea.

I spent three nights working on the ground mesh (going to sleep happy, and waking up not liking what was on my screen), and you can see earlier iterations (woof). I learned about Shift-G (to repeat actions), but also learned that using hard right-angle shift-rotations for curves really just doest work in most instances. It's too sharp and screws with the quads, and I have to merge all the vertices, blah blah. I'm still more of a fan of the interpolated bridge. I might still employ this tactic to create something like that IRL sidewalk photo, but would need to move the pivot out a block of some size. None of this blather prolly makes any sense.

This map basically consists of only two pieces/kinds of geometry; the ground mesh, and a 128x128 bed-rock grid I would cut up for the center buildings and extrude up for the platforms at the top of the hill, and then later use as the displaced center earth areas that hug the buildings. I can make a crazy ass n-gon from all the edges of the inner curbs, but of course cannot subd or paint on that. I tried extending, and angling down from the curb, and it kinda works, but doesn't reach the center without alot of effort, and either way it was going to need alot of terrain massaging, and is much simpler to use the base faces and displace them upwards. I also displaced downwards on the road/grass mesh to the bed rock mesh in the lower corner apartment area, but thats a bit rough, and if I toggle off DP, some of the edges turn red. I don't know how the engine computes visibility and what not with displacements, I hope it all gets melted down to polygons beforehand. I took a glance at visibility type stuff in the wikis, but didn't apply any of the suggestions.

I learned how to use UV-peel to make the road texture work, that was fun. At the time of this writing I have not cut the road up into sections to see if makes the (brick sidewalk) peel look better. I didn't initially think to move the templates along the z-axis to create a hilly vertical effect, but ♥♥♥♥♥♥♥♥♥ it came out nice, albeit it entailed the aformentioned terrain displacement work. I wish I had (slightly) randomized the distance of the curb to the road in the templates to add more 'jitter', while right now the length of the asphalt is exactly 256 around the whole level.

Upon the last night/day of the initial ground construction, I started with the ground mesh being circular, such that the grass areas only extended out to the length of each original template piece. I then had the idea of actually having the template grass extend beyond the planned 3072x3072 space, and clipping it off after bridging to make the grass fully extend out to the edge. I was chuffed to see it worked great, although I had to make the four corner plates perfectly 45 degree angles with the corner of my big square to ensure everything was a quad. It was mysteriously required the merging of two vertices at each corner, I guess as a nature of what how the clipping tool works. In the end all my faces in this map are quads and edge-loopy-compatible (at least prior to beveling and subsequent fix ups), and I'm proud of that. Near the edges of the map the grass quads get very rectangular and that makes for harder painting, so I threw in an edge loop to give it a boost. This implies that maybe in the end, the whole thing would be better as just simple displaced terrain from the bottom, but the road briding them self was alot of fun, and (sans outer grass I suppose) could make a sick road system now.

Additionally, I also started to reuse my photon map, bake lighting on selections, and use lower quality probes to speed up iteration on lighting. Even starting from nothing though, this map bakes and builds relatively fast, I suppose because the Source 1 props are all low poly, and because I'm only using the environment light and one spot spooky spot light. Perf is acceptable to me, I'll see if I can arrange the foliage for less overlap, but I tried to apply it sparsely enough. Sometimes it runs at full frame, other times its in reprojection. Yikes. That's about as technical as I can get when making maps where everything is visible to everything else and I (presumably) have no room for optimization there.

I wonder what else I learned or new buttons in Hammer I now know to click, I forget.

As for the asset packs....hmm. My trepidation on actually using them is expressed within my various comments on those submissions, but I think overall they hold up enough.The first thing to dedude is whether or not they look too terrible in VR/source 2, and I suppose many do, but I knew I would end up with this anarchonistic mess of asset quality.

The second problem I discovered was problems with some of the model pivot points. When I was still studying doors, I wanted to use a door gate and rotate it with PVR, but the pivot point thing was in the center of the model, and unless there's some way to change that without editing the model, I figured I was going to have to fix it, and reupload. Nah. I considered just removing the asset pack submissions from the workshop and working off local content, and then only including what is used in the map at the end, but I don't know how to do that cull. Decided to take what I have and use what works, skipping what doesn't.

Other problems include the human characters not having eyeballs and broken lod levels. I had some dancing models, but since my asset packs dont have sounds, they're ragdolls. Some foliage props had flickering shadows. I got scared, thinking I had too much crap in my scene, but I disabled shadows on the offending props (trees/bushes mainly) or replaced them.

Some decal/overlays didn't work unless the target was world geometry, and it appears the source1 skyboxes don't work either. I'm hoping this is the only real use of these packs and somebody does a better job in the new engine.

Didn't have time to do floating water or flying birds. The music is by elysis.de, which is some ambient tracks I love and have been listening too since I was young. I foist them upon you so you can be present within my slightly depressed mood. This took about 1 month of work, and missed BlackCatAppreciationDay release by 10 hours.