Modified TOS Enterprise

Originally posted by EBOLII
blah blah...little lonely are you?.....nobody to talk to huh?

Okay, just so I know where I stand, are you really razzing me for saying too much or were you being sarcastic about the fact that I posted a bunch of wireframes to illustrate my techniques without saying a word about them? I'm inclined to think the latter, in which case simply allow me to remind you that a picture is worth a thousand words.

On a serious note...I think it is a seriously perfect mesh thus far..What are your intensions for the Nacelles?..I personally like the 2nd season end caps that have the same deflector type tip and the hole exhausts

Actually, I'm not a fan of the hole exhausts. I'm not a huge fan the half-spheres either. The design I actually prefer is the open cowl and sphere that Franz Joseph used for his blueprints and that I modeled for my USS Constitution.

The only other modification I'm thinking about for the nacelles is changing the grill details on the inner sides to something along the lines of the TMP refit. I sort of did that on my Constitution using textures only, but I'm planning to model it for this one. I'm still going to keep the other features like the intercoolers and reactor loop, though.

I actually prefer is the open cowl and sphere that Franz Joseph used for his blueprints

I need to see it....unable to picture them?

Originally posted by EBOLII
I need to see it....unable to picture them?

Maybe http://www.sub-odeon.com/stsstcsmua/...stitution.html ?

Or were you guys thinking of the Ballentine Books Star Trek blueprints, not the Star Trek Technical Manual?

Michael

Another round of modeling updates to the hangar door area and fantail:




The hangar doors are in place, as well as the control booth above them. I used the same overlapping panel design for these doors that I did on the Constitution only I reversed the order so the innermost panels are in the center instead of on the sides. Why? Mainly because when I get around to modeling the door tracks in the floor of the hangar deck, this arrangement will conceal them in the closed position.

My doors are also shaped a little differently from the TOS filming model. Mine are a perfect quarter sphere, except for the offsets from the overlapping panels. The ones on the filming model are a little out of spherical true and their midline is actually slightly below deck level.

I also lined up the centerlines of the hangar doors, the control booth and the landing beacon on top of the hull. On the filming model, the control booth is off-center toward the rear of the ship and the beacon even more so. My beacon winds up to be obviously farther forward, but I prefer the design symmetry of having everything lined up. One nice side effect of reversing the door panel order is that the control booth winds up looking like it's still off-center toward the rear, even though it isn't.

Lastly, I added a "spine" to the secondary hull, similar to the TMP refit, and connected it to the base of the landing beacon.

Here's a couple more renderings for good measure:




By the way, total face count is 26,694; although, I think I can optimize that by a good 30% or so by the time all is said and done.

Looks nice. :thumbsup:
Honestly, I wouldn't bother optimizing away anything. I think the poly count is on the low side for such a smooth model anyway. For the level of smoothness and detail you're shooting for 100k polys would be just about right.

Originally posted by EBOLII


I need to see it....unable to picture them?

I don't have an image handy at the moment, but I'll try to post one this afternoon that shows the aft end of one of my Constitution nacelles.

Originally posted by Nova Class
Looks nice. :thumbsup:
Honestly, I wouldn't bother optimizing away anything. I think the poly count is on the low side for such a smooth model anyway. For the level of smoothness and detail you're shooting for 100k polys would be just about right.

Well, the optimization I was talking about was just cleaning up unnecessary polys. For example, the sec hull was cut in half at one point and then mirrored so I didn't have to manually duplicate all those details on both sides. That left a cut down the center of the lower "box" below the deflector and the cove below the hangar deck. Those are flat surfaces (from side to side at least) so at least half of those polys can go away. I also have surfaces on the details on the deflector boxes that are hidden inside the model, as well as the top and bottom edges of the hangar doors, etc. etc., all of which can go away because they are never seen. Oh, and there's quite a bit more that can be trimmed out of the smaller diameter parts of the deflector dish, as well as the straight section of the engineering hull itself.

EDIT - I went ahead and made most of those optimizations I just mentioned and the polycount is now 25,343. 30% was probably a high estimate, but I'll bet 10% or 12% is doable. Either way, I'm still confident I can keep the model under 100K polys.

Originally posted by EBOLII


I need to see it....unable to picture them?

Okay, here we go:

Okay, y'all ready fer this?



The pennant is modeled, not textured. I shapemerged it into the hull, separated the polys as their own object, then restored the original hull and floated the pennant object about a centimeter off the surface. Old trick a lot of other modelers have used for a long time but a first for me. Worked pretty well.

You'll notice I have also applied some film grain and upped the contrast slightly. Just experimenting a little bit.



TA-DA! Got all those access panels or whatever they are applied to the bottom of the hull. These, too, were shapemerged, but I went the extra step of modeling in a double groove around each panel.

One more angle highlighting the access panels and the navigation lights on the fantail:

Here's a nice close-up of those access panels:



Yeah, they look kinda cool with the modeled grooves, but this image actually illustrates why I should not have modeled them. You have to be at least this close to notice that the grooves are there at all. I'll probably go back and eliminate them, though I may compromise by raising the outline of each panel slightly, which will only require about half the polys.

Finally, here are two more "beauty shots" to show off the lighting and material effects I've used so far:





I'm especially pleased with that last one. I'm using three ray-traced omni lights and a hundred-odd shadow-mapped spotlights arrayed in a globe. The main hull material is just a blinn shader with a large noise map in the diffuse color channel. The deflector dish is textured with the Max standard steel material, but I tweaked the specularity and glossiness and adjusted the tiling in a concentric pattern to make it look a little like stainless steel. On top of all that, I added a film grain of .09 and increased the contrast to .85. Pretty realistic if I do say so.

By the way, the polycount is now up to 27,750, though I'll probably be able to knock a couple hundred out of that by redoing the access panels differently and a couple of other things.

The secondary hull is nearly finished except for the interior of the hangar deck and the "window boxes". I'll probably save all that for a later stage after I've made some progress on other parts of the ship.

Nice, Vector!:thumbsup:

I'm not sure I like what you've done with the pennants though, they just don't look Starfleet anymnore, more like... FLash Gordon.:p

The underhull panels are cool though and maybe you could separate them so you could animate some of their functions. F.E. the yellow one is the Warp Core access panel. You could model the shaft and core and do a cool ejection animation.... ah just dreaming out loud


I have to say... The render looks very good, but damn that's an awful lot of lights!!! I could've done it with 3 or 4. How long did that take to render?

Originally posted by Nova Class
I'm not sure I like what you've done with the pennants though, they just don't look Starfleet anymnore, more like... FLash Gordon.:p

Yeah, I'm not totally thrilled with 'em either, though I like them better than those 60s refrigerator door trims the filming model had. Don't get me wrong, they had their charm, but I really feel something different is needed. What I've done here isn't all that different from the pennant on the TMP refit except I've got three stripes instead of two. Still, I might try blunting the ends a little bit instead of letting them come to such sharp points.

The underhull panels are cool though and maybe you could separate them so you could animate some of their functions. F.E. the yellow one is the Warp Core access panel. You could model the shaft and core and do a cool ejection animation.... ah just dreaming out loud

Anything's possible.

Later today I'm gonna rebuild those panels by eliminating the outer groove and accentuating the inner groove. It will save a few polys and will also let me treat the panels as separate objects, just as you suggest, though I haven't decided yet exactly what I'm going to do with that. I've thought about building details behind all of the access panels and then rigging them to be openable. Could make for a pretty cool rendering to have the ship sitting there gleaming in the drydock lights with all the hoods up and all that powerful machinery exposed.

We'll see.

I have to say... The render looks very good, but damn that's an awful lot of lights!!! I could've done it with 3 or 4. How long did that take to render?

I just double-checked, the exact number of spotlights in the globe is 114. Render time for the last image I posted above was 28 seconds. I could get by with fewer, but I certainly couldn't achieve a soft, ambient light effect with only 3 or 4. It really does make a quite noticeable difference in the realism of the image.

Redid the access panels to simplify them and make them look a little better from a distance:

Ack... lose the film stock look--your model is too beautiful to hide her behind anachronistic film grain!

Well, I've been wrestling with the primary hull for the last two or three days, trying to do modeled gridlines in a way that looks decent yet doesn't blow my polycount all to hell. This is what I have so far:






First, let me say right off that those rings around the bottom of the saucer are still a little too wide and perhaps a little too deep. More of the latter, I think. I'll be fixing that in the next iteration.

Second, just in case you're wondering about the funky mesh pattern in the wireframes, I built the saucer as a single 15 degree slice and optimized it for less density toward the center. Saved a lot of polys that way. The saucer currently stands at 25,248 faces, which is more than I would have liked but not bad considering the smoothness and the modeled gridlines.

Speaking of gridlines, I'm still debating over those suckers. They're so faint that I'm trying to decide whether to model them in 3D at all, as opposed to modeling them flat or simply texturing them. One thing that bugs me is the lack of smoothing from one side of a gridline to the other, which tends to make the panels look a little like they are puckered outward around the edges. I can more or less fix that but it requires a lot of added geometry.

And lemme tell ya, getting the edge of the saucer to smooth correctly was a pain in the ass. That same lack of smoothing across the gridlines really showed up badly on the sharper edges. I wound up adding an additional edge on each side of the radial gridlines, just where they wrap around the rim of the saucer, which spoofed the smooth shader into making them look the way they were supposed to, at a cost of another 5,000 or so faces.

So what do you all think? Are the modeled gridlines worth it?

My two cents on grid lines, they’re not worth modeling. The ones on the model are only light pencil lines and in some place nonexistent.

Your work is excellent I look forward to seeing the finished product.

Alan
http://www.angelfire.com/scifi2/wizardofflight/

There, I think this is an improvement on the concentric rings:

I'd agree on the gridlines. They'e so fine that you'd only notice if you got really close, far closer than anyone is liekly to want to go. Better to make a hi res texture and use that. You won't even have to bump-map it.

Originally posted by Proximo
I'd agree on the gridlines. They'e so fine that you'd only notice if you got really close, far closer than anyone is liekly to want to go. Better to make a hi res texture and use that. You won't even have to bump-map it.

True, but only up to a point. Right now, the gridlines are defined by light and shadow, allowing them to blend and fade from certain angles and distances. A textured gridline won't necessarily do that. Is the difference worth the extra overhead? I don't know, I haven't decided yet.

Oh well, if you can solve that 'puckering' problem without too many polys, I say go for it.

*notices numerous spelling mistakes in is last post*

Oops...

Okay, here's an option using modeled gridlines, but they're flat like a decal rather than "etched" into the hull:



This allowed me to eliminate over half the faces previously required for the gridlines. The face count is now 17,952 as opposed to the previous 25,248. By Grapthar's Hammer, what a savings! And I don't think there's even much of a noticeable difference, other than the fact that the hull plates no longer look "puckered" around the edges. Thank God.

Here's one more to show how they hold up visually at a distance:



I think this will about do the trick.

I think that looks much nicer. Your work is simply marvelous either way.

So, how does that work exactly? Have you given them a different smoothing group?

Very impressive by the way.

Okay, here's the latest--and hopefully final--primary hull saucer with flat, modeled gridlines. I did some more tweaking on the shape of the cross section to round off some edges a little more and very slightly reshape the curvature of the lower "bulge." Somehow in the process of rebuilding it, I shed another 1,104 faces. The count for the saucer now stands at 16,848.




As you can see, I have also added my own version of the lower sensor array. I say "my own" because it does not conform precisely to the shape of the filming model. It's close, but as with other parts of this mesh, I have modified it just slightly to get what I feel is an improved look.

I think I've got the gridlines exactly how I want them and the overall shape of the saucer is right where I want it as well.

Opinions?

Originally posted by Proximo
So, how does that work exactly? Have you given them a different smoothing group?

No.

Basically, what I did was build one 15 degree segment of the saucer, then I assigned the faces for the gridlines a different material ID that's darker in shade than the rest of the hull. Then I arrayed it 24 times around the center of the saucer, attached the separate pieces, welded the vertices and smoothed the whole thing off.

I see. I'll have to remember that.

Added the interconnecting dorsal today and started playing around with some details on the lower sensor array.





For those of you who are interested, that last image was rendered using the ever popular Catmull-Rom anti-aliasing filter.

By the way, the total face count so far is just a fraction under 50,000. Half way to the target of 100K.

Built most of the bridge module today with some extra detail added to the turbolift:



Some of you may be interested to know that the teardrop-shaped B and C deck object was built with a spline cage.

I also added some temporary placeholders for the warp nacelles and support pylons, just to get a feel for the overall proportions:



And here you can see that I have modified the shape of the interconnecting dorsal to give the rear edge more of a squared-off appearance:



And one more just because it looked cool:

Got a lot done on the impulse assembly and took quite a bit of creative license in the process. I've altered some of the specific parts and pieces, but the general shape remains the same:




Opinions?

puuuuurteeeey!