Hi there I recently got the Stingray model from Trickster to work on it in my sparetime. So I decided to post my progress here so yo guys can see I'm actually doing something. So here we go: Turret: Left is the original, right the lowpoly one. Polys: 35.818 -> 424 Engine: Original: 83.637 Lowpoly: 6.812 (dont know yet what i will do with these tubes back there) So if anyone got constructive comments tell them to me otherwise leave this thread!
Cool, keep it up. Also if you're contacted by anyone named Mr Die or someone wanting to advocate Albany or The Blackholian state ignore them.
The turret looks well done as a low poly, but the engine seems like it lost a little too much detail towards the back end. Do you think it would increase the polycount by that much to add a little more detail back in around that area? And an indent at the front to keep the turbine texture a bit more "set-in" could be nice too.
The engine has a few improvements, for one the array of spikes at the front can be made into cones, the original geometry is more or less a cone, and making the lowpolys into cones would remove the need for an N-gon on the end, probably saving you about 5-6 polies per spike, it'd also look better. The pipes can probably be broken down into pentagonal prisms with smoothing groups, allowing you to use few polies, and use maybe one or two bends to get the curved sections, personally though i would suggest removing the pipes and spikes and just baking the detail onto the flat surface below, at least for the LOD version anyway, you need at least one very low poly version for viewing at a distance, because planes will not usually be occluded by the game's optimisation system. Oh yes also as mootant has added the parallax technology, depending on whether it can deform along face edges, you might be able to use it to get almost all the detail, for example the spikes could just be left as is, and the rest of the cylinder could be deformed inwards to produce a very high fidelity effect with very few polygons. Definitely consider using it because there are places on this model which could benefit from it greatly.
hmm I will do that with the front of the engine I'm not quite sure about those spike though the most polys on the low one are in the tubes at the back atm as you can see from the original that they go all the way to the back and have around 12 to 16 sides I will replace them with something else one other thing: I didnt mention to do the textures myself yet I mean I know how to make an alpha channel and use that for parallax but I dont quite know how to bring the tex on the model (yet) I agree to realise the details on the spikes with parallax mapping but please realise that the diameter of the engine is about 3/4 of a human and as I've seen from many examples: parallax mapping for big eleveations on a texture looks just damn ugly so I would say I should do them as real models
No, parallax mapping doesn't look ugly here. Try it first. Your base mesh will need to be a simple cylinder for most of the length. Try baking displacement map, look at it in xnormal and be careful with UV mapping. Do not even think about parallax mapping in terms of what you currently have as low poly. All these mini cylinders are totally redundant when you're using it.
yeah well ive seen that on games like metro 2033 and even on the screens of the test here in empires: the plane gets 'transformed' but since the imagetexture is still projected from the normal it becomes distorted (I just realized I might confuse this with parallax occlusion mapping...)
Nothing becomes distorted, texture on the edge represents more space in 3d space. If you account for that with UV mapping there won't be any problems.
BAM and on we go!!! after having a few struggles with buggy boolean operations in 3DS Max(all solved by restarting said program), i've improved the engine so far it may look different but they got still the same length and concerning the suggestion of taking cones instead of cylinders: it doesn't matter i tested it. no difference in polycount, even when you make them with a pointy end with radius zero not sure what i go on for next, i think the turbines are pretty easy and contain a lot of polygons oh nearly forgot: i deleted nearly an amount of 100k polys of bolts on the hull so we are now on ca. 41.000 polys (it was around 350k once)
... Did you just scale a circle down? Because if you do that then yeah, it won't make any difference, because that's not a cone, it's an irregular cylinder. A cone is when you connect all the edges to a single vertex, not superposition half a dozen vertices...
O snap, a modeler that knows what he's doing :D Gj, Can't wait to see everything complete and in game.
used the predefinded cone but can try that see if there is another primitive i can use without making an editable mesh but either way a real cone looks ugly and it wouldnt be that much less of polys if i would hav to do it then as the last effort
If you aren't using editable meshes, how exactly are you optimising it? I don't think the standard primitives give you much choice in removing excess polygons for example.
You know the game can only actually load things as a mesh, and you'll probably need to convert it to a mesh to export it as an SMD which you then need to compile?
