Home 3D Art Showcase & Critiques

[UDK] Diorama : Old Gate

This is a personal project I just finished. I'm really trying to get my skills up for creating 3D game assets. I learned a TON doing this thing! Very fun.

I started from a concept that I saw ... here on Polycount, I think, actually. I clipped out a small gate in the background to use as a reference:


Untitled-1.png

After awhile it sort of took on a life of it's own and the reference became more of an inspiration than a blueprint.

At any rate, here's what I ended up with!

FinalShotA.png

FinalShotA_Wires.png

FinalShotB.png

FinalShotC.png

FinalShotD.png


This was a great chance to really get some time into learning Mudbox, 3DS Max and Photoshop better.

Any thoughts or comments would be appreciated!

Replies

  • DDuckworth
    Offline / Send Message
    DDuckworth polycounter lvl 6
    wow well your textures came out nicely but the wireframes look extremely heavy, especially on the ground/stones and the roof tiles, looks like you just decimated them and didn't hand optimize or bake out a NM to them, the piece looks nice but it certainly doesn't seem game res to me
  • WarrenM
    Yeah, the floor stones were some of the first things I did and I didn't optimize them nearly enough. I could definitely do the same look with less polys there.
  • DiegoTeran
    Offline / Send Message
    DiegoTeran polycounter lvl 11
    i think everything is a big to heavy on the poly side ;/ textures are great and im sure even if u take off at least half the poly count u will still get the same result
  • Oniram
    Offline / Send Message
    Oniram polycounter lvl 17
    you may want to try and play with lighting and atmospherics a bit more. as of now the light seems untouched aside from the light shafts setting, and the fog in the scene sort of takes away from what this could be. on one hand you have the very saturated roof giving off a big bloom, and everything else looks a bit washed and worn (im assuming because of the fog). the textures look nice but the "moss" or whatever it is dripping from the pillars doesnt sell to me. it looks more like a green burn mark than anything. the grass seems as though it is just a solid color on a plane (with an alpha or whatnot). i really do like the underlying rock, though i think it could use some more beating up, and i think it would look really cool to have some roots or something protruding out of it.

    so here are my suggestions.

    take your light and give it some sort of a yellowish/orange color and bring up the intensity.
    remove the fog, or just tone it down. its too much.
    because this is somewhat of a fantasy piece, go into your World Settings and bring your Environment Intensity to about 2. and your Diffuse to about 10.

    change the environment color (it actually looks as if you are reflecting a standard checker material.. giving off that blue/purplish color from lightmass. id say maybe lighten up the environment color to somewhat of a lightblue/grey.

    fix the plants.. maybe grab some flats (or make your own, thats always fun.. just scanning plants. :D ) and model off of them, or just apply them as an alpha. just so they have some shape aside from green-looking sticks.

    id say maybe also add a blue pointlight in the center. something with a small intensity and falloff, but something noticeable.

    most importantly, add a postprocess volume and start playing with colors. adding DOF wont hurt either. :D


    happy travels.


    EDIT: OH YEA! REDUCE THE POLYCOUNT. :P most of that can sell off of flats with normal and displacement (bump offset) maps.
  • WarrenM
    Gah, blue light .. can't believe I forgot that!

    Thanks, that is massively great feedback. I'll see about updating the scene...
  • DiegoTeran
    Offline / Send Message
    DiegoTeran polycounter lvl 11
    sean show off lol
  • WarrenM
    OK, so I'm taking the feedback here and trying to put it to good use. My primary task at the moment will be to reduce the poly count. I'm redoing some of the earlier assets and simply hand optimizing other ones (target welds, etc). I'm making progress and will post shots as I go along.

    I'll also try and incorporate the visual feedback as well. Thanks guys!

    We started at :

    12,358 triangles / 47,495 inst. triangles (according to the UDK primitive stats browser).

    So far...

    OptA.png
  • Oniram
    Offline / Send Message
    Oniram polycounter lvl 17
    better. but still a bit dense in the middle. would be best to just hop into max (if you are using it.. if not there are other methods) and use Pro-Optimizer, making sure that in editable poly you have PreserveUVs CHECKED!!! that way it will break down your mesh and still keep the same unwrap. :D the shingles on the roof need be taken down in tris quite a bit. and the stones on the wall on either side of the gate should be taken down as well (personally i would just leave that pretty flat.. normals can do the job). visually tho its looking better. the colors are coming out a lot more now without the fog.
  • WarrenM
    Oh, it's not done, believe me. I'm going to hit every mesh and get them as low as I can as I think this is a great exercise for me. Will be multiple updates as there are less and less wires...
  • EMC3D
    Offline / Send Message
    EMC3D polycounter lvl 14
    Your wireframe is way too much!

    Your pillars could be basic polycubes and still hold the same detail with a low number of edge loops, you have so much going on when all you need is the basic silhouette of the wireframe, it makes no sense.

    It's almost as if you didn't bother to do a lowpoly and used decimation master / reduce polygons from a highpoly.
  • WarrenM
    It's almost as if you didn't bother to do a lowpoly and used decimation master / reduce polygons from a highpoly.
    I know, right? It's almost like that. Almost exactly like that.

    ;)
  • Oniram
    Offline / Send Message
    Oniram polycounter lvl 17
    ha. i dont blame you. but usually when i decimate things.. ill do it 3 or sometimes 4 times to get it low enough. :P
  • WarrenM
    I haven't forgotten this, I just haven't done anything worth updating the thread for yet. I've got the triangle count down to about half of what it was but there's a lot more to go.

    I just grabbed a copy of WrapIt last night and am learning how to best use it. Never really messed with re-topology steps before, so it's interesting to say the least!
  • sltrOlsson
    Offline / Send Message
    sltrOlsson polycounter lvl 14
    No offence, but dude. Learn lowpoly modeling. Super easy things like the flat surface of the gate have triangles all over for no reason.

    Just my two cents..
  • WarrenM
    I'm trying! ...but it's a process. I'm learning all this stuff as I go and the feedback in threads like this is invaluable to me.

    I didn't mean to imply that WrapIt was going to solve all of my low poly modeling issues, that would be silly - but it's a tool and I think it's worth learning it as I see it recommended a lot.
  • sltrOlsson
    Offline / Send Message
    sltrOlsson polycounter lvl 14
    Sure, it's great, i use it all the time. Or i use MeshLab, same same.

    But if you do want to learn game asset creation you should get into the process of creating a lowres version for baking.

    Then you got the small details that's pretty heavy, but the round "platform" is very jagged if compared to the rest.

    I realize it's a process, we've all bean there, just don't get to comfy with easy shortcuts (:

    I do think it looks good though! Very nice little atmosphere..
  • Oniram
    Offline / Send Message
    Oniram polycounter lvl 17
    I'm trying! ...but it's a process. I'm learning all this stuff as I go and the feedback in threads like this is invaluable to me.

    I didn't mean to imply that WrapIt was going to solve all of my low poly modeling issues, that would be silly - but it's a tool and I think it's worth learning it as I see it recommended a lot.

    im actually very against WrapIt for low poly creation. i saw a demo of that tool way back when and the guy made a low poly of a horse from just boxes, and then turbosmoothed it. sure it gets the job done but using a tool like that doesnt really teach you about topology. i think the best way to learn topology is to mess it up. i use 3dcoat for retops, and if im in the middle of retopologizing a mesh, and i make a mistake, i basically delete everything around that area and start over. its a huge help when trying to learn proper low poly modeling. id definitely recommend 3dcoat, or even the step tool in 3ds max. they both basically do the same thing except the hotkeys in max are kind of weird for the built in retop
  • WarrenM
    I'm working on figuring out this low poly stuff. It's going alright. I've remade the same piece at least 4 times now and I'm going back in for a 5th this morning. I'll get it eventually! Thanks for the input, really, it's helping immensely.
  • sltrOlsson
    Offline / Send Message
    sltrOlsson polycounter lvl 14
    I'm working on figuring out this low poly stuff. It's going alright. I've remade the same piece at least 4 times now and I'm going back in for a 5th this morning. I'll get it eventually! Thanks for the input, really, it's helping immensely.

    Cool, i say post it up! Even though you know what's wrong and that it's shit, letting others tear it down some extra will make you success on the second or third try instead.
  • WarrenM
    Getting closer. I have a few meshes still left to make low poly meshes for and then I need to re-make the textures over the weekend.

    I think I'm getting the hang of the low poly modeling and I've been doing it all by hand which makes me regret buying WrapIt now, but as I do more organic stuff perhaps WrapIt will come into it's own.

    OptB.png

    Atmospherics, bloom, DOF, etc are all going to change in the end. I'm currently just focusing on my triangle counts, exclusively.
Sign In or Register to comment.