View Full Version : Environment,Need feedback!

01 January 2008, 08:51 AM
I was wanting to start making environments for a while so here is what i have been working on lately. it has fairly simple models no complex geometry. this is how it looks in unreal engine 3.(sky is not mine)
the inner corridor does have roof but it can only be seen from the bottom. Let me know what do you think and what should I add to make it look more interesting?

01 January 2008, 05:33 PM
Great modeling, really good trextures but only an ok start on the lighting.

- It looks like you turned up the ambient light level? I would turn that off, it is washing out the dark areas.
- Think about doing more dramatic lighting with less lights. Especailly in the hallway it is too bright, hardly any shadowy dark areas are present. Contrasting the light areas with dark will help the depth of the hallway as well as break up the repeating structures.

I did a paint over to help illustrate what I mean, since its easier for me to show it then to write it up.

I also wrote up an explanation of some basic light settings and techiques for 3ds Max for the most part they might apply to Unreal3:
Something tells me you could make good use of negative lights in this scene, that is
if Unreal lets you use a negative light to subtract light form a surface.

01 January 2008, 08:38 PM
looks great! just a lot of open bare spaces

01 January 2008, 01:38 AM
Thanks for taking time and commenting:)
I tried to fix the lighting(it has always been my weakness). Also changed the sky.
I was using point lights in the hallway, i changed those to spot lights(i had to use 2 lights for each bulb,one projecting at bulb and the other projecting on the floor, i don't know if that's right approach). I was using fog in the scene probably that was contributing to a washed out look, i turned it off. Waiting for more comments:)

01 January 2008, 09:05 PM
Looks good. Seems like you have all the large details looking good. I might recommend adding some other details of different sizes. Things like vegitation, flags, wires, crates.... etc. Things to make it feel like it is actually a used space.

01 January 2008, 03:31 AM
thanks nathanielbell;
after reading your post, some grass happened to grow in my environment alongwith some props:) . I will add some grass in hallway too.(maybe some other tiny props as well)

01 January 2008, 08:23 PM
I worked on it a little more.......

04 April 2008, 07:27 AM
hi again
here is my latest work basically i used casey's gas station for reference with a little bit of my own imagination.(note: sky dome is not mine)

04 April 2008, 10:04 PM
here is wireframe. as you can see that polycont doesn't look as good because of the fences(all of them being at about 10000tris).

04 April 2008, 11:14 PM
curious to know where 30,000 of polys are going in that scene? it just doesn't look like it should take that many polys to make that scene. Seems like it should be 1/2 that amount or less.

oh, i just saw your fences are taking of 10k of polys. If that's the case, you should make different design choices, i'd say. Your fences are just random eye candy detail, that in reality doesn't need to be that high...those are objects which could be replaces with something else that does the same thing in the end.. You could, for example make chain link fence, brick wall, or other kinds of "walls/baricades" that don't take up 10k

and if you really wanted to, you could use the extra polys for stuff way more interesting than repeated fences...more detail on the gas station, broken signs, big chunks missing in the asphalt etc.

04 April 2008, 11:15 PM
It looks good, but excessively dirty/grungey. And at the same time, the geometry is so straight and sharp. Tweak the edges so it's not so squared off and perfect.

A side note, doesn't anyone like to do colorful, non-post-apocalyptic stuff anymore? ;)

04 April 2008, 11:41 PM

Actualy i was wrong those damn fences were taking 17000 some polys.the gasstation is at 5750 tris in which 2100 polys are for pumps. and the lightposts are 666tris each(hidden in the following image). I will get rid of those poly hungry fences and replace those with something more efficient. And i will look into other areas that are using excessive polys.

04 April 2008, 06:39 AM
I worked on it a little bit,broke some parts of the building. I will make more broken peices of concrete/brick to be added on the floor. I also added some props from my previous environment just to fill up some empty space. this is mental ray render I will export everything(changed parts) once I am completely done with it..... and I got rid of those expensive fences:buttrock:

04 April 2008, 01:33 PM
Hey man, that's off to a nice start.

The pumps feel a little bland to me right now though. Just not much to them, and kind of feel out of place because of that.

On the short wall that surrounds the entire piece, it looks like you added a strong highlight on the edges of the top concrete piece. You should really just do a high poly model and get a nice bevel on it, then project it, rather than fake highlighting. Unreal does an amazing job with normal maps, real time lighting, etc. The highlight on it feels pretty unconvincing from 90% of the angles. The concrete itself (Judging by the last image posted) is quite noisy as well.

Off to a good start though, keep at it!

04 April 2008, 06:42 PM
Wow! thanks for the solid critiques. I will be working on the things you pointed out. I thought I would probably get away with the pumps but I guess it doesn't work so I will rework it(what poly count should I aim for the pumps?)
about the concrete being too noisy is probably because its mental ray render and the one that looks less noisy is unreal engine (its same base texture)

04 April 2008, 07:07 PM
Polycounts vary big time. Depends on the importance of the prop, it's location, quantity, etc. Since you only have a couple of them in there, I'd say you'd be safe being anywhere from 800-1500? However, since you are making this to just pimp out your work, you want it to look as awesome as you can. I don't think any developer is going to pass on you because some of your props are a little high poly or high res texturing. Frankly, I usually do a lot of my personal work at about twice the resolution/polycount than I would for an actual game (within reason). So if you are using a 512 for that pump, go crazy, use a 1024 or whatever. That's just my personal take on it. Some people would disagree, because you do want to show you're profficient with lower res stuff (and that's a very good and important thing too) but if you just want to showcase what you can do in realtime, might as well take it the whole nine yards. Plus, if you need to show profficiency, you can always go back and optimize :).

Just my 2 cents.

04 April 2008, 01:39 AM
Thanks again for clearing things up. I wish you always critique my work :deal: .
I made new pumps and I am really happy with what I have now. Its at 1214 polys(only 154polys more that the old one) and a 1024res texture. The following picture only has color map as I have yet to make normal and spec maps.

06 June 2008, 03:19 AM
I am back again. Here is what I have been working on. 2nd building's texture appears a bit blured and the 3rd building is wip,so far it has 3 X 1024 texture maps. I will add another storey on top and then roof. Should I keep the model of window frames or leave it to textures for upper windows? critics are welcome

*Images removed for faster loading.

06 June 2008, 11:39 PM
Here is a bit update on texture.

06 June 2008, 05:19 AM
very good start man :) I think the only you need to worry about is breaking up the texture and structure patterns of the buildings. You can pretty much put props and or anything that compliments the structure. They all have a good start - thumbs up for me :)

btw- I personally like this structure- looks intresting :P

06 June 2008, 09:44 PM
Looking good. The first one (with the red/orange bricks) looks good, but the bricks are so clean (especially when put up against the foundation texture that's all grimey). I know you are tiling the crap out of it, but adding in some general dirt grime would help a bit, and then using some decals and small static models to kind of break up the rest of the areas (to make it feel unique, and break up the large spaces).

Just my opinion. The latest shots you posted are looking really nice.

06 June 2008, 10:49 PM
have to agree about the clean bricks, they also need somthing to break up the pattern a bit and make it more interesting but I guess that could be done with foliage later on. Another thing about that brick texture is that there's a slightly darker/pinker row of pricks that gives away the tiling, maybe colour edit it a bit more to blend in? or was this some sort of intended effect?

06 June 2008, 04:04 AM
Euclidius: Thanks for liking my buildings. That building was completely made from my mind I didn't have any reference. For the current wip building I am going to make some props(Benches, Trash Can, Post Box and Street Lights. If you have any in mind let me know).

Josh: Thanks man, Actually I made that first building quickly to test the reflection vector on the windows. I am going to redo the textures once I am done with my current WIP.

Swift3D: Thanks, I have no idea where those dark pink layer of bricks came from and it wasn't intended. Since I will be redoing it and scraping current texture so hopefully it will look better.

In the meantime lets forget about the first building(for a while) and here is the progress of my current project. Its not completely done yet but is close

06 June 2008, 04:54 AM
That latest shot is looking great man. I have no real critiques at this point.

06 June 2008, 04:59 AM
Thanks bud :)

06 June 2008, 03:09 PM
Yeah, the last building looks quite nice :). But don't listen to that guy about adding crates xD

06 June 2008, 12:00 AM
Lol no I wouldn't add crates to this building.I took it in engine and here is how it looks. windows are using reflection vector based material(it looks a lot better while walking around it than the still images).
I specially need critiques about lighting as I think that it is my weak point.

06 June 2008, 09:29 AM
That's looking really nice but there's some fairly obvious repetition that bug me a little bit. Marked it on the attached picture. If you could use a decal or two to cover it up or maybe shift the textures around a bit to get a bit more variation I think it would benefit the end result.

06 June 2008, 01:02 AM
Thank you for pointing that out. I will work on those areas and probably redo the textures on the other building.

06 June 2008, 07:49 AM
I tried to put decals on that and fixed some scale issues with door and some texture changes.

06 June 2008, 09:56 AM
Looking good. Really like your new models and there nicely textured.

06 June 2008, 01:59 PM
Looking really good man. Keep going with that direction, it's really getting better with each pass.

07 July 2008, 04:30 AM
Thanks Jeevan and Josh. I had been really busy for the past couple of weeks so didn't really get a chance to post reply.

07 July 2008, 05:51 AM
I made this model over the weekend. It has about 2880Tris and one 1024res Texture.
Its not 100% done yet I have to make a few more elements in this scene.

07 July 2008, 01:19 PM
Hey man, that's looking really good. Especially to be on one 1024x1024! Only a couple comments.

There seems to be a lack of normal map showing. I don't know if you have one on yet, or if it's just not very visible. A strong bump and spec on this piece could really be the icing on the cake!

Secondly, the awnings you have kind of stand out from the realism a bit more than anything else. I think part of the reason is they look like they are emitting light :) Also appear to be quite clean compared to the rest of the building. I would add some dirt/grime on there. I would think the awnings in particular would be one of the easiest targets for weathering, dirt, damage, etc. Could be as easy as lowering the saturation on some of the green areas (to make it look like rain water has faded the dye away), adding a little overlay of dirt, and not making them react so brightly to the light.

Regardless, this is a strong looking piece so far. I can't wait to see an update.

Vertex Groover
07 July 2008, 02:13 PM
Thats a very clean and effective model. But yes, take what Flewda as suggested on board. Thats about all I can fault it with.

That, and maybe adding some pavement...?

07 July 2008, 05:31 AM
Josh: Really great comments just like always, thank you. I feel like I am addicted to your critiques:)

I worked on awning texture as suggested. The building does have a normal and spec map but they were done really quickly(in 5 seconds) as I was tempted to take this model to engine so they are not really showing at all but I am going to make better normal and spec map. I will post an update after I get those completed.

VertexGroover: Thank you I have that in mind and I will be adding the pavement so this building has bit more personality:)

07 July 2008, 04:50 PM
Great environment art, man! Keep it up!

Could you post some texture sheets when you get a chance? And maybe your UV layouts? I'd like to see how you're getting so much detail out of a 1024 map.


07 July 2008, 07:24 AM
Time for Update,

I made sidewalk and a quick floor as well as adjusted lighting,Hopefully normal and spec maps are displaying better.

Also a couple of textures for Artan. original size is 1024X1024.

07 July 2008, 05:02 PM
Thanks for posting those!

Building is looking great. Keep up the good work!

07 July 2008, 05:16 PM
The normal is still a little less present, imo. For example, the relief/design detail at the top of each column looks like it's painted on, and not even remotely close to 3D. The normal just isn't popping there at all. I would consider doinjg a high poly for that piece if you can, at least a general shape high poly, then do the detail normals with crazybump. The bricks themselves look little differnet to me (in terms of normal map), perhaps push it a little further?

Also, about half wa up that same column, you can notice some horizontal lines going through all the bricks. Almost looks like some errors where you were making the bricks seamless?

The awnings look MUCH better now, I like them a lot. Nice work. One thing you MAY want to consider (just an observation) is that you have some large wrinkles on them, it might be cool to see some smaller/tighter wrinkles, especially towards the edges where the cloth is tightly pulled around the framing? Just an idea, I don't think it will make or break anything.

Really though man, aside from the normal, anything else is really splitting hairs. This is a solid piece, and would be a great addition to a city environment.

07 July 2008, 05:47 PM
Josh: I am going to work on it further tonight and fixing things you mentioned. Should I model a lower poly relief as well? I don't have crazybump as I had beta version and its expired and at this time I am on tight budget. I am using Xnormal and then photoshoping to make normal.
I agree about that horizantal line I will fix that.
I am glad it almost weekend, I will have much time to work on this

07 July 2008, 05:53 PM
If this building is more or less a large prop, I wouldn't spend the time/polys making it geometry for low poly, but yeah a quick high poly for a normal will go a long way. Also, Photoshop's normal filter works really well, there are some tricks you can use for it. I suggest though starting by making a white background, and painting all the mortar areas dark gray or black, as a base. Maybe give the black lines a gaussian blur (.5 or so ought to do it). Then run the filter on this. Then, run the filter on the diffuse itself, and overlay it on top of the normal you generated from the black/white hand painting. Also, you may want to take the layer you made a normal from the diffuse, and go to its curves options, drop down the menu to the blue channel (instead of RGB) and make it's output to 128. Then will remove the blue channel from the overlay. It'll look odd by itself, but with overlay on and having it ontop of the hand painted normals you created, it'll make sense. There's some debate about how effective this is, but in my experience, it's worked well.

Hopefully this made sense to you. If you have any more questions, you can hit me up on my AIM, i'd be more than happy to help out.

07 July 2008, 02:15 AM
Thanks for the photoshop trick.... I have been playing around with the normal in photoshop, I think I got pretty decent result.
I wrote a tutorial on how to setup reflection vector in a material in unreal ed. Here is the link. Please let me know what do you think

07 July 2008, 03:59 PM
Looking good but you should really model the ornaments on the pillars. They look like they should have some depth to them and not be flat against the brick (grey ornament on pillars, brick background)

07 July 2008, 05:26 PM
I am going to agree with urgaffel on this. I thought maybe a normal would cut it, but I think it's going to need to be some geometry. Shouldn't be too bad in terms of polycount. Just something simple to give it some depth.

07 July 2008, 09:04 PM
yes i agree to urgaffel & Flewda it'll look much better if you model the ornament or just cut into it to give some depth and also i think the bricks look slightly thin and slightly wide or is it just me but keep up the good work

07 July 2008, 05:29 PM
Thanks for the input.
Here is the result with really low poly ornaments.

07 July 2008, 07:19 PM
LoL, you're gonna hate this comment. I'm not saying you NEED to change this to call this piece done, but rather something I would address.

Back to the ornament piece. The texture resolution/clarity, and overall quality of the photo you used isn't really holding up well. It wouldn't be horrible by itself, or further away from the camera, but having it with nice crisp brick textures right behind it, really makes it stand out as being lower pixel density. I don't know if it's the UVs didn't get enough space or the image itself you used for it was too low res and had to be scaled up. Either way, it does stand out to me. But again, this is just being pretty nit picky, especially if this was made with the mindset of it being a background building or a non-gameplay important building for a level.

Seriously though, it's looking awesome, and I wouldn't blame you if you just call it done. However, since it's for your portfolio, and you want that to look as awesome as you can, I would try to addrerss this one last little thing.

My fault for not saying it earlier. It didn't really hit me until today for some reason.

07 July 2008, 05:39 AM
Actually I had the same feeling about it when I posted this update and even after modeling those ornaments it wasn't looking right so I made a little change in texture and I think now I can call it done:)

Edit: Don't think that I hated this comment. Even if you would have pointed that out after a month, I would have still gone back to fix it.

09 September 2008, 06:41 AM
Here is a test I did recently for a studio. No job though.

Vertex Groover
09 September 2008, 08:29 AM
No job? Thats a shame :( it looks really good.
3300 Tri's does seem like alot though, but I'd assume they'd give you a limit for your test.

Could we see the wires?

(I'd ask who the job was with but with the UNSC kinda gave it away...)

09 September 2008, 03:31 AM
yaahh I am quite suprised too? :surprised I have been following your work for quite some time now and I think they are really good - welp I would say good luck but you don't need it, so I'll just say, keep it up man! ;)

09 September 2008, 05:10 AM
I was surprised and disappointed. They didn't say a single positive thing about this piece and just told me that "it wasn't upto the quality bar they were looking for(unsure whether this comment was about modeling or texture) and they didn't understand the design choice I made on the sides of generator that weren't visible in concept(I was supposed to be creative) and some of the lines in diffuse texture could have been more crisp, they looked like hand drawn."

Definitely a good learning experience though.

Heres wires(the polycount budget was between 2800-3600)

Vertex Groover
09 September 2008, 08:01 AM
I was surprised and disappointed. They didn't say a single positive thing about this piece and just told me that "it wasn't upto the quality bar they were looking for(unsure whether this comment was about modeling or texture) and they didn't understand the design choice I made on the sides of generator that weren't visible in concept(I was supposed to be creative) and some of the lines in diffuse texture could have been more crisp, they looked like hand drawn."

Definitely a good learning experience though.

Heres wires(the polycount budget was between 2800-3600)

How negative! :sad: Personally, I could see it fitting in their most recent game quite nicely. The texture looks fine,the normals seem to be in the places that'd have the most effect and you've hit the poly budget right in the middle.

Well, don't give up, Work like this would pretty much get you a job at my place, and a lot of other places as well.
Keep at it, you're really hitting a professional standard of work now.

09 September 2008, 08:06 AM
Thanks man,

Never gonna give up:D(just sent you a PM)

09 September 2008, 12:33 AM
Hey man,

That piece really turned out sharp. I don't know why they didn't follow up with you. You know the comments I made about them via email, so I won't publicly comment on stuff like that, but I think it's going to be their loss in the end. I see a lot of potential in you dude, and you've been improving like crazy. So yeah, keep your head high. You'll land something soon.

09 September 2008, 08:09 AM
It had been all the critiques from you guys,(specially Josh, BTW I still have that detailed email feedback you sent me about each piece on my website), that have helped me to be at the quality level I am at now.

09 September 2008, 08:33 AM
A very simple building wip (usually found in south Asian countries)

09 September 2008, 07:26 PM

Don't feel bad man. I did that test for Bungie too. It was my best work I have ever done, but still not good enough. With mine though, they said I didn't model it to the right proportions they were looking for. (it was a bit long) My UVs were also crap, now that I look back. They didn't say anything about what I did on the other side, I just mirrored it and added some other decals. Kept it simple. But yea.. disappointing. I feel ya. Although, now you have a really awesome prop for your portfolio! All we can do is just move on and hope the next art test is our ticket in.

I won't post a pic on your post but you can check out what I did on my blogspot. (

also, I was able to find someone else that did this art test, and he WAS hired. You can probably find it on google. Type in Bungie Art Test or something. The guy has it on his webpage portfolio. Here is one I found that got hired. There are others though.

10 October 2008, 02:52 AM
Awesome model!
Btw... are you using parallax bump maps for the bricks? I think that could improve a bit the appearance.

10 October 2008, 07:57 PM
Great UV's. However I think you could use a bit bigger texture resolution on the building, then again maybe not, I dunno. I think also a bit more 3D geometry could be added, but then again I dont know on that.

But one thing for sure. Normal maps or parallax maps would really improve the look of the building. So far I like what im seeing, keep up the good work = ).

10 October 2008, 05:51 AM
whoa! didn't realize my thread had new posts.

GOBEE: not a problem, I wouldn't want to work under a blind art director anyways who can't see which piece is higher quality.

Jogshy: It has normal map atm. I will look into parallax mapping.

Mathew: Could you point where I should use more geometry?

10 October 2008, 05:55 AM
Time for another wip. So far this building is pulling from a 1024res texture and I still have room for the upper windows. I am going to change the entrance area and would use another 512res texture.

10 October 2008, 07:19 AM
Small update on the entrance.

10 October 2008, 02:53 PM
yeah, this building is coming along nicely.
Especially i like the 2nd floor window textures but in term of scale those white curtains look off to me. The reason is that the scale of such curtains should be small i think, but they are huge in size if compared to entrance for example.

11 November 2008, 07:57 AM
Thanks Matroskin I will get to it soon.

Update on the new entrance.

11 November 2008, 05:53 AM
Your models all look good but there are some technical issues with all your models that are holding you back. I've gone over some of the best examples.

1. Putting decals on your model using poly planes might seem like a good idea. But you run the risk of Z-fighting. This is where the game engine or 3d program doesn't know what to draw on top. You can see it on the right hand side and also in area 5 that there is a drawing error. In most engines this will produce flickering (3d programs especially 3DS max are very forgiving, engines are not)

Z-fighting happens when you have two surfaces too close together. For some game engines such as Doom 3, this problem is reduced.

The general rule is the greater the draw distance the more sensitive to Z-fighting. If you add speed to the mix it gets even worse.

2. Polygon density. In any game, regardless of the poly budget, you want to be as efficient as possible. First thing you need to find out is "how big is my model/prop/character going to be on screen?" There is no point in making cylinders such as those in area 2 as round as they are if each polygon is less then a pixel thick.

3. There are two problems with this area. First, same as above. Too many poly's are being used in this bevel. there is no way you are going to see this in game. One Bevel should be enough and make sure its big enough to see. Don't be afraid to scale up details so that you can see them.

The second problem is smoothing. there are problems throughout this model with smoothing groups (hard/soft edges for Maya users) I've highlighted in blue the edges i think need to be made hard.

One thing I might say about smoothing groups is the fewer the better. The more you have the slower your model will load so you cant just assign them willy-nilly. Worse then that, the code for SG's in max is really stupid. if you use a higher number like 20 it will include 1-19 as well in the model. Even if the model only has one group. (if you use maya no worries)

4. This area looks like it might have some density issues as well. I'm not certain, but it looks like there is a bevel along the highlighted line. If there is it is too small to see meaning that it needs to be made bigger or deleted.

This cylinder crashes through the box it sits on. Which is fine but it needs to crash though all the way around and the top does not. This creates a visible hole in the model which is unattractive. The average gamer will see this as a mistake.

5. This is another Z-fighting problem. You can fix this easily by making the box section thicker. Or you could try welding the cross beams to the box making it one continued surface.

6. These two edges could become one with out changing the look of the model. You might even be able to cull more of these edges with out damaging the look. Its worth a try.

7. I would make these tubes 4 sided. They are so small that they are only going to be a pixel at most in game. (you could even try making them 3 sided.) The smoothing and texture will fool the user into thinking they are round. With any cylinder, if you cant see the cap you can get away with murder with regards to making it low rez. General rule is to spend the polygons on the more important bends in the tube. Rather then roundness.

8. There is no reason here for not welding the two surfaces here together. you already have the polygons. (it will actually lower your poly count.) Most importantly it will save texture space.

Overall your model is too many pieces. If you were one of my artists I would tell you to weld a lot of these elements together. Make it out of fewer parts. That doesn't mean make it one piece. Students tend to get it into their heads that everything should be one continues model. There is always a balance.

You will find that the fewer parts your model is made of the easy it is to texture and the more texture space you have.

In all the modern platforms texture space (or lack there of) is the limiting factor. Especially when you add normal maps, spec maps and what not. So its worth spending poly's if it means

9. smoothing group issue (similar to 3)

1. triangle fans. (also called polls) You may or may not have heard of triangle strips. Basically this is something that you will need to look up on your own time. (and you need to if you want to be successful in the games industry)

Generally a well though out model will be 50% less expensive then a poorly made model. (even though they have the same number of poly's)

Basically beyond the # of triangles is the amount of tri strips in a model to determine how fast it will load in game.

Some engines don't care about triangle fanes. Others do. The engines I used at Backbone did something called fan striping. It recognized that it was a fan and could make it one single strip. Many though would consider each triangle in the fan its own strip.

2. I wonder if maybe you are using too many poly's on these curves.

One general rule of making buildings is know your cameras. Most games are either from street level or top down. You need to concentrate your polygons in the area closest to the viewer. Either the top or the bottom not both. You have done many models it would be worth doing examples of both. This goes for texture space as well.

I have noticed that there is inconsistency in texel density in you buildings. This means that some things like bricks that seems much higher rez then other concrete details. You want to keep the size of pixels (from the cameras perspective) as similar in size as possible.

One last thing on buildings. To better fake distance and scale you should try and de-saturate colours as they get further from the camera. So if your shooting from street level the texture at the top of a tall building will be more gray (or maybe blue gray) then the more vibrant colours at the base of a building. This is pretty noticeable in real life. And worth overdoing in game.

In closing you obviously have some art skills. and your texture skills are good. What i have said here might not get you a job but it should make it easier to keep one.

I hope this helps. Let me know if you need clarification on something.

11 November 2008, 07:00 AM
Thanks Chris. Its definitely great to hear from professionals and getting advice and feedback. Specially such detailed one.

Actually it was test's requirement to use polygons as decals. Also I was trying to keep my polycount between the given budget.

I have known that its ok to have multiple objects in a model as long as they are "combined" before import in engine (I guess it depends on engine)?

Isn't it better to have those triangle going towards corner(triangle fans) as long as the model will be lightmapped?

I must say that I loved your feedback and I am keeping these points in mind :) and I look forward to hear from you again

11 November 2008, 01:58 PM
great post, cholbrow, thanks, very useful stuff.
Tri stripping is really a pain in the ass, it takes me a good amount of time to wait while exporter tri strips the level;)

I agree about Z-fighting; however, there is 1 thing i wanted to mention.

At work (i do stuff for the Wii) i use decals a lot because they give me flexibility and help to save on poly count and texture memory because i can use smaller generic textures and then add dirt etc as decals over it. It also allows to create fake shadowmaps if engine does not have its own shadow map (and in the case of Wii is happens often).
In my case max actually is not forgiving and feeds me with flickering but the game engine feels pretty fine with rendering coplanar surfaces close to each other (0.01-0.002 as safe distance). Even on DS i figured out that 0,1 m is a safe distance and in case of TPS or RPG on a small DS screen this is really small value.
I think it really depends on the engine and hardware. I agree that it might be tricky but I just wanted to say that it is not that bad and often beneficial. However, overlapping polys is definetely not the right thing ;)

Btw, do u have some links to info about fans? Cause i think that triangulation that salman-fas has is correct, i do it pretty much the same way and that's the way i learned to do it from others. But if there is some good info on more efficient ways drop it here, it is always cool to check out something new ;)

11 November 2008, 08:41 AM
Ok back on this project. Here is how it is so far. I have hidden the entrance part for now. I fixed the scale issues as mentioned by Metroskin. I will work a bit more on the main floor windows so they won't look as weird.

11 November 2008, 04:54 PM
looks great. @.@

11 November 2008, 10:24 AM
The building looks good, I'm a big fan of three point perspective

"I have known that its ok to have multiple objects in a model as long as they are "combined" before import in engine (I guess it depends on engine)?"

About combining objects. Making every model one solid mesh is crazy, but I do think that your field generator has too many elements. If i was looking at this model as an art test this is the main reason i would reject it. I do find though that artists that have a good sense of this are rare. Its just something that you have to experiment with.

Z-fighting sensitivity is based on how extreme your near and far clip plane are. (there is only so much resolution in the z-buffer) If you have to be able to see things close up and also far away (like say Zelda) you have to be careful. Games like FIFA where most of the game is from a distance or like Doom where everything is close. Well then you are safe. The quality of you engine can have an effect too. The last engine i used sorted polygons well but alpha in a really primitive way.

some engines can do something called z-buffer offset. Which allows you to tag your decals so that they don't z-fight.

I have noticed in recent versions of max the clip plane tool is busted, (although i haven't tried it in 2008) so that you cant adjust it.

as far as tri strip info goes there used to be some good stuff here on cg talk but it seems to have been deleted. I've got the best result by trying to even out the distribution of edges to vertexs. Spread the load so to speak. I find though that the biggest difference you can make with tri strips are with your hard/soft edge and uv seam choices. as these are static non skinned objects, how clean your geometry is can make a huge difference.

I hope this clears some stuff up.

11 November 2008, 08:15 AM

It does clear things up a bit. I did look for that thread on cgtalk but unfortunately the image links are broken. I will keep searching though. (BTW I use Maya).

CGTalk Moderation
11 November 2008, 08:15 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.