This hydrant look familiar to anyone? :P
polycounter lvl 8
H'okay... so as the title says... look familiar? 
Yup, it's the art test for DE, which I took yesterday... and not sure if it's crap or just... slightly less than crap. I kept thinking I had more time until the 4-hours rolled around and it looked like GARBAGE. And I nearly started to freak out. But I remembered from EricV's post of it, how people were saying to just take some extra time for it. I asked the HR lady I was in contact with, and she told me to take as much time as I needed and just say when I completed it.
Which meant... it turned out slightly less horribly than I was afraid it would. I have to wait about a week to hear back on what the art department thought of it. It took me a bit over 5 hours to complete to where I thought it was submittable.
So here it is, in all it's fire-hydrant glory:
Dunno what to make of it... I KNOW I could have done better, given say... double the amount of time. Or if I did the same thing again. But this is what it is!
So, basically, what do you guys think? Passable, or far below acceptable?
If the latter, please elaborate :P
Thanks!
-Mezz
Yup, it's the art test for DE, which I took yesterday... and not sure if it's crap or just... slightly less than crap. I kept thinking I had more time until the 4-hours rolled around and it looked like GARBAGE. And I nearly started to freak out. But I remembered from EricV's post of it, how people were saying to just take some extra time for it. I asked the HR lady I was in contact with, and she told me to take as much time as I needed and just say when I completed it.
Which meant... it turned out slightly less horribly than I was afraid it would. I have to wait about a week to hear back on what the art department thought of it. It took me a bit over 5 hours to complete to where I thought it was submittable.
So here it is, in all it's fire-hydrant glory:
Dunno what to make of it... I KNOW I could have done better, given say... double the amount of time. Or if I did the same thing again. But this is what it is!
So, basically, what do you guys think? Passable, or far below acceptable?
If the latter, please elaborate :P
Thanks!
-Mezz
Replies
( I have become very picky lately after failing a couple of art tests)
I like it in general, and I like your site too. Plus you got to do all that inbetweening that im jealous of... so nyah. :poly124:
Also the scale seems weird. The model reminds me more of a chibi character (huge head). Not sure though cause I didn't try to search for a reference. But ya.. the bumps make the scale look weird too. Just look at the topside and the bottom.. the bumps are different scale and the bottom obviously looks better even though it is the topside that most likely will be looked at more. Looks like an UV issue actually.
For an asset as small and simple as this, for a current-gen console game, you'd be mad not to render down from highpoly. It just gives you a better end result.
You could also stand to optimise the lowpoly a bit more - as has been mentioned, the way you've modelled the chains is wasteful (use a single-sided loop of faces, or an X cross section of 2 planes as someone else mentioned), plus those rings around the base could probably lose an entire edge loop around the middle and nobody would notice.
Then you could put those polys back into rounding out your cylinders - those two large cylinders coming out of the spherical top are only 8 sided, which is painfully obvious. They should be at least 12-sided, maybe even 16. Anything less and they'll look angular.
The texture is ok, but generally a bit noisy and lacking any specific directional dirt/damage. For example on a real fire hydrant, the base where it connects to the floor would be dirtier, and there would likely be water streaks or rust stains coming down from where the cylinders join together. Currently it's just "all over bumpiness", which doesn't really make any sense.
Still, it's a pretty good effort, and I bet if you made a new one from scratch (and i mean completely new, don't re-use anything from this one), you'd probably finish it faster and it'll look better.
Hope that helps.
The texture looks less meaty then others I've seen, but needs work. The pieces where clearly textured independently as the wear and tear don't really match from piece to piece. It happens quite a bit when people try to dump a ton of noise all over their normal map instead of baking from high to low, like mop suggested.
Going back to the noise of the normal map, grit and scratches are only valuable if you have something to contrast them off of, like some untouched areas. What makes a ruby valuable? Not everyone has one. If everyone could find them laying in the gutter no one would care. The same goes for scratches, the more you add the less meaningful they all become.
More scratches doesn't mean more detail, it means more mess. I haven't seen the supplied concept only peoples end results but even if the concept is noisy I think it would be a good departure not to vomit scratches and meat juice all over these...
I'm not really sure why there is a box behind the end caps? It looks like you tried to texture it like it wasn't there? Whats the box for and why wasn't it left out?
The geometry is very stuck together and intersects badly. You might have been able to hit the target poly count by using a quad sphere and extruding/cutting geometry into the needed shapes instead of using a geo sphere like a voodoo doll.
Lastly (and this is a general statement to a lot of people) when are people going to realize that when MoP (and quite a few others) say its easier to knock out a high poly and bake it down, he's not talking out his ass... High poly modeling shouldn't scare anyone. Its not harder, its easier and lazier then doing low poly. The restrictions are off and you can knock up a few simple details quickly. If you use a few simple materials (often procedural) you can even get a huge head start on your diffuse!
/off bitter old crank
if an art test says take as long as you need i would be tempted to go for longer and do a really good job, companies want skills and eyes, speed is great but can be learnt/beaten into you
Some really great critics, and I agree with it all. The problem is, I've been so used to taking my time (and I mean taking a LOT of time) that I wasn't really ready to deal with the timeline, I suppose.
I know I could have done a much better texture and spec and normal, but I ended up having to rush through all that and just hope to make something passable. I also like the idea of making high poly first, but didn't realize that it would work in this case. Now that I know, I would definately give it a shot next time.
The alpha of the chains did turn out really brutal, I agree. I wasn't totally sure how I'd go about it, cause i did a chain once before and it turned out way too thin, so I tried this out instead. I would have gone back to fix it, but it would have taken me too long... -sigh-
Now that I've done this, though, and basically made a pile of crap, I really am inspired to do some more fast-modeled props. I hope they'll turn out better, especially without the pressure this time.
-Mezz
p.s. Kawe--it's hilarious how you think that the head looks too large and that it makes it look 'chibi', cause I tend to have that same problem with characters
Alex
Well, I guess it might not be a bad idea, what's the worse that could happen? But at the same time, it's been a few days since I took it, wouldn't it look weird if i asked to do it again now?
And then still, I worry that I'll once again take too long with it... and if I do poorly time-wise a second time, they'll definately not think I can handle their pressure... Which given, might be true... But still... hrmm hrmm...
Well, I have 'til Monday to decide and e-mail again, what do you guys think? Ask to retake, or let them go with what they have?
You have to think that they see quite a few hydrants and even if you manage to tow the line there's going to be a bunch of other people who did the same.
"I liked doing your hydrant so much I thought it needed friends, so I started building a scene"
I think it is all about how you perceive the world. It is hard to paint what you really see instead of painting what you think you see. Drawing a face is IMPOSSIBLE cause I always project it so it is straight ahead even if it is rotated. that's how my brain interprets faces.. and in this case I think it is cause we think heads/faces are more important than the rest of the body thus it gets chibiish.. heh.
One question though--So I've been trying to model a new prop, and I took the advice to make the high poly and then get the low poly from that so it'll have better details, better normal map, etc. However, I might be not going at it the best way...
I made the base in 3-D max, then exported it to z-brush and modelled all the detailing. I exported it again from there and tried to bring it into Max, which basically killed the poor program. I might try again with it at a lower geometry level, but otherwise, is there another way I should have gone about to model the HP? One that doesn't destroy Max? Or, considering I was doing some fine ornamental detailing, what I did was best and will just have to compromise the amount of polys that Max and my slow ol' computer can take?
Thanks again for all the advice everyone
P.s. That's actually really interesting, Kawe, I hadn't really thought about it like that but it makes sense. I definately do put the highest importance on the face, so I guess that's why it always turns out biggest on my characters. And even more importance on eyes, so I have trouble drawing small eyes.... Hmm, something interesting to ponder.
About the chains... I'm a little confused about the techniques suggested.
What do you mean by making 2 strips in an 'X'? I'm not exactly sure what this looks like/how it's supposed to work. A simple illustration would be very helpful, if anyone can provide!
The same goes for using 'a single-sided loop of faces'.
Any help with explaining this would be great, thanks!
As for the chains just like doing a bush with intersecting planes.
Like this:
(But with a better chain texture, that wasn't done in 30 seconds)
And yeah, I tried to bringing the piece I was working on into Max at a low geometry level, and it managed it alright. But yes, my computer really does needs lots more ram... I just need to remind myself to go out and get it :P
Thanks, Vig
Rares, it could work that way if I use the z-mapper, but using that thing, to me anyway, is the most fickle feature of z-brush that sometimes works, and then other times comes up with a million and one reasons why it just won't work. When it does, it's great, but I try to avoid it due to major headaches.
Thanks, though
I'd model out the highpoly, then do a quick optimise on those cages for the lowpoly (or build a new one from scratch, for a prop like this, that might be faster).
Then bake down normals, ambient occlusion, along with some colour info for selection masks, and finalise the textures in Photoshop.
Thanks again, I'm off to go make that darn hydrant again now! Wish me luck
Also, increase the resolution of that shadow map . Blocky hard shadows don't help to make a nice image
Btw... did they allow you to use some kind of reference image? I think it's really hard to model an object without it. If you can, try always to use a photo as reference:
http://en.wikipedia.org/wiki/Fire_hydrant
google images like this one:
etc... until you like one. Then, try to copy it. Without a reference it's very difficult to put there the right dimensions and colors.
And, finally, try to use a good lighting setup and background. If you use the xNormal's realtime 3D viewer, take a look to the triLight values and add a diffuse cube map to simulate skylight illumination ( the diffuseGI slot ).
I'm certain the modeling turned out better, and thanks so much to everyone who suggested doing the chains as x-planes! It works out so much better that way
The texture... well, I tried a few different things this time, and tried to focus on bringing out the metal peeking through a lot in the reference. (To those wondering, yes, they give us a reference pic, I just don't know if it's ok to share or not, especially cause all the test details are written on the image.) But in the end, I'm not certain if the texture turned out that well or not. I think where I hit a wall with it is trying to get the old heavy paint chipping away effect that it has.
Nonetheless, it has been sent on to Digital Extremes with a plea to have them review this one after I already did it the first time :P I would still love to get some critiques back on this, though. I'm not sure whether or not I've made it good enough for them, so if not, I'll need everyone's helpful comments so that one day it will be!
Here it is: (and man do I not want to look at a hydrant for the next month...)
My main critiques:
- Why is the base section with "41" on it upside down in the UV-map? Always try to UV-map things the right way up if possible, especially if they're going to have text on them.
- Your normal-map appears not to be normalised, it looks like your "noise" overlay is not blended correctly. To blend a normalmap correctly in Photoshop, you need to use Overlay mode, and drop the Blue channel's output levels to 128 on the overlaid layer.
- The specular is far too uniform ... the red paint should probably not be quite as shiny as the bare metal. I can't see any of the surface variation from the normal map or diffuse anywhere in the specular. You should use the specular to really push "material" contrasts (shiny/matte).
- The normal-map is still too noisy, it just makes the whole hydrant look bumpy. I've never seen a bumpy hydrant in real life, they tend to be quite smooth. I'd recommend only having bumpiness around the base of the hydrant where it meets the ground, and around the parts where the paint has flaked away.
Hope that helps.-I realized a bit after I had started the textures that the main base was upside-down. A silly mistake that I normally catch, but apparently not this time... I would have changed it, but in the interest of trying to keep my time as low as possible, I mentally slapped myself and told myself not to do it again
-You're right about the normal map, but I didn't know there was a better way, ie the one you mentioned. Thank you for that advice, I'll keep it in mind for next time.
-I was trying to bring out the highlights of the metal in the spec, but perhaps I didn't do a great job... I'll try harder for next time.
-The bumpyness I put in was because there was a lot of chipping paint that caused raised surfaces. The noise was to hopefully show off some of that... but yeah, probably could have been better executed.
Well, they're both in... now I get to wait and see what the results are. :poly122:
Common causes of inverted faces:
Using the mirror tool. The geometry mirror tool inverts the UV's but not the normals (unless you reset xform). To correct it you must flip the UV's using the UVW editors mirror tool.
Instead use the symmetry modifier.
It allows you greater control and can be used to work interactively on the mirrored side. Meaning what you do on the left shows up on the right as you work. (provided you have "show end result" toggled on under the modifier stack).
I like to stack my symmetry twice around an edit poly modifier, so the stack goes:
+Symmetry (updates your model in real time)
+Edit Poly (where you make your edits)
+Symmetry (first symmetry)
+Editable Poly (Lowest level)
You can use just the lowest two but it puts a cage around your model and I find that really annoying. Going the symmetry sandwich approach gets rid of the cage.
With all that crap out of the way. Its looking much better. MoP has some good advice and I'd stick to it!
In addition:
- I think it would be a good idea not to bake in a bevel/emboss on red paint it's going to fight your normal map.
- Also think about reducing the amount of chipped paint. It will help you control the total amount of visual noise.
The chains are much better but the construction is a bit off. Normally you want the edges to match, and the cross should run down the middle.
- There are two ways you can do the chains.
Yes, I am using 3Dmax, and I've been told in the past it's better to use symetry instead of the mirroring. However, I've tried it, and always find I prefer to use the mirror tool. I hadn't realized that it flips my UVs... so I guess maybe I'll try using symmetry again instead :P
Thanks again to everyone who has helped me with my first art test... (lol well, we all have to start somewhere.) Now, back to the waiting for a response...