Art Test Critique

cay030

Howdy all, Looking for some crit's on a Art Test. Really excited about and nervous. Anyway if you lovely Polycount viewer's can lend a crit and view this little badboy of an Ice Cream Truck, the Tri's limit was 8000, i'm at 7464. Wasn't sure if i could post the Included image, if this is not allowed please edit the post and remove the image. Thanks and i look forward to the info.

Chris Y

Concept


Render's

Em.

Didn't someone else just have the same art test up?
It would probably help if you posted some wires so everyone can see exactly what's going on. I don't have much time, but one of the first things I notice is some ugly smoothing on the windshield, side window, bumper side, and on the "E" in Cream. Probably ought to check your smoothing groups.

cay030

I thought someone posted one, i did a search but couldnt find it on the Truck, anyway , thanks and i will post a Wire and will also check the smoothing groups. Thanks

Wells

heh. yes, looks familiar

not seeing where the 8k triangles went, but I'll wait for a wireframe to crit

cay030

Wireframes

Spark

Cay030, you have alot, and I mean alot of wasted polys that you could redistribute once you clean things up. And that is another thing that you want to make sure to do, is have clean wires (The word cream is just a mess) I would suggest you go back on this, and rework the areas that realy need it. Before you progress to the uv/texture stage, and it becomes a more difficult job.

Spark

danr

there's a reason they set these "difficult" poly limits, 8k being quite a big limit for a boxy van... they want to see that you can use the big budget creatively while still being efficient - those are two staples of good poly modelling. They want to see what elements of the concept you choose to model and why, and how it fits into the overall shape of the model. They want to see that you've worked up the elements that really give the concept its character. They want to see good poly density for the various shapes, and the ease of manipulating it for UVs, lods, creating potential damage models, etc.

If you can hit a good representation/interpretation of the concept with exactly 8000 triangles with not one single triangle wasted, that would probably be ideal in this case. They'd probably prefer that than if you've saved a few hundred off the budget

But what they certainly don't want is you to fill the budget up with a load of tris cos it was quicker and easier for you to do that then spend time cleaning it up.

Thinking about it, that's quite a tricky art test ... looking at it from the pov of someone who sets these things, that's a hell of a lot harder than it first appears. There is a solution, but i'm not giving that away , wouldn't be much of a test if i did ... sorry it's not much of a critique, but i thought it'd be just as useful to know how these things usually work. good luck with it

SubPablo

watch your proportions. there are quite a few major elements in the concept that you've scaled down or under emphasized to make it look more realistic. Unless they stated it was for a more realistic style, try to match the proportions as much as you can. Right now, your model and the concept look like they're from two completely different games stylistically.

cay030

Thanks for the replies, yea there was questions i had looking at it,the text on the cone for one, i kept thinking there was an easier way to do this. Thanks keep you posted

bounchfx

why is the windshield curved? in the concept it looks flat at least to me, and that would probably save you some tris. I'm not sure how closely you are supposed to follow the concept though..

and it doesn't look like 'smoothing' errors as much as it does triangulation errors. (I'm pretty sure that's not the same thing)

cay030

Here's a wireframe update

Jesse Moody

there is still some really weird geometry that I am seeing. It doesn't look very clean either. Lines aren't straight, 5 sided polys or more.

It looks like there are a ton of polys on the horn.

With 8000 tris you should be able to make one hell of an ice cream truck and right now it isn't at that stage yet.

The tires have a ton of wasted tris as well.

vertexguy

Not a bad start, but you still have a ways to go. Take a close look at the proportions of your truck verses the concept art. There needs to be some serious revision there. It's also very hard to tell where you're using your poly budget from this single wireframe but you definitely don't have enough detail in the model to justify 8000 tris. Look at areas where polygons are not seen (like the bottom of the cherry) and remove them. Also if you have any chamfered edges, they are too subtle to see and could be removed to save more polys.

Motz

To be honest there are incredibly obvious differences. I don't intend to be mean but these are really simple. Just put your wire frame pic up against the concept. It's a very simple boxy model, and your geometry isn't even remotely clean.


1. Why is this a box? At the most it should be a single alpha-ed plane. It doesn't even look like you need an alpha plane from the concept as it is flush with the body of the truck in the concept.

2. The proportions of your cone are completely off. Just look at the concept, it's much bigger.

3. You are correct in that the detail here for the door and window don't need modeled, unless you have left over poly budget. A normal map would pop these out just fine. You are missing the mirrors.

4. I'm guessing it's just not finished, but you have half a van there.

5. Again just look at the concept, the wheel well is very circular and even. Yours is not and the flow is quite ugly, try using a math generated object like squash/edit a torus to get that rounded form.

6. Missing bumper.

7. The whip cream on the ice cream, i would just throw it out at this point. Tons of wasted polys. Whipped cream isn't some symmetrically perfect mathematical shape, its just a gob of organic fluid over the ice cream. The ice cream is the same way. Yours is also way too large of a topping compared to the ice cream.

8. Why bother modeling this, it's flush with the van. Unless you have budget to waste. Again a normal will pop that out.

Look if I can find these in less than 10 seconds, simply by looking at the concept closer, you should be able to crit yourself. These are extremely simple modifications. My advice is to search for some tutorials on polygon flow. There are angled polys that don't have to be, wasted polys used for nothing etc. This is an incredibly simple concept, and they gave you and incredible budget to see where you can put in some creative thoughts. Add a freezer etc. Just look at the concept, over and over and compare it. Once you get the idea, then go nuts with creativity until you are good enough to bypass that. Long way to go, but keep pushing!

You have your gamer tag under your resume... your name is 'LordBallSack1'. You would have to be a drop dead amazing artist for me to even click on the resume if I was a professional art recruiter.

cay030

Thanks for the replies, will work on redoing this.

P.S. Motz as far as the Gamertag, i didnt think it was anything bad to keep on the site or it didnt have anything to do with my resume. I just had it up there for people i know. Its been removed now , thanks.

monkeyboy_garth

There are a lot of things wrong with the concept. The more I look, the more I find. Good luck. A literal translation could end in tears.

Rhinokey

its common to have concepts that are impossible to be translated literaly, but a good artist can translate it well enough for someone to look at the concept and the model and go "thats right"

in the concept the red and white fringe above the large window is painted on the van, and not a hanging fringe.

your front wheel wells are way too elongated.

keep working one it.

Dakkon

Usually when it comes to art tests like these, if they give you a pretty massive poly limit, as someone said earlier, they want to see how you use the polygons. In general, the way I do it is I'd model the object the same way I always would for a game, which would be a fair bit lower than 8k.

After that, try to think about how to improve the look of the vehicle to how it would appear in game. Beveling edges is one of the first things I'd do. Take a look at the profiles of the object. Does anything stand out as being really polygonal? I've modeled a PILE of vehicles for various games, and the one thing that always happens is you need to add a lot of polygons in the wheels to get them looking proper, so don't be afraid to do that.

Rounding off the windows a bit might be helpful as well. Anything that a texture can't really translate without being at a high resolution might be a good idea.

Also, I think you can take a fair few polygons out of that cone and add some fine detail to the lower area.

SuperOstrich

[ QUOTE ]

You have your gamer tag under your resume... your name is 'LordBallSack1'. You would have to be a drop dead amazing artist for me to even click on the resume if I was a professional art recruiter.

[/ QUOTE ]

How the hell does Poopinmymouth keep finding work?

Thegodzero

Name repetition, personality mimicking, and never wearing pants. That and a deep love for the term hot coffee.

cay030

Thanks for the replies guys, still wokring on this.

Dakkon thanks for the reply, how would you tackle the text around the cone, i thought maybe a alpha plane but that wouldnt give the depth needed.

Thegodzero love the environment work, and props you done. great stuff.

And yeah i guess LordBallSack could look bad on a resume page, LOL

Thanks

Eric Williams

[ QUOTE ]
7. The whip cream on the ice cream, i would just throw it out at this point. Tons of wasted polys. Whipped cream isn't some symmetrically perfect mathematical shape, its just a gob of organic fluid over the ice cream. The ice cream is the same way. Yours is also way too large of a topping compared to the ice cream.

[/ QUOTE ]

I hope that's not real ice cream they put on top of trucks O.o I agree with most other stuff motz said, but I also agree with monkeyboy_garth.

If you can't translate literally then you might ask what they want. But this is a test, so you can't ask them how big they really want the cone on top of the truck or whatever else looks like it can't be translated literally...So, I would translate as literal as possible. And just assume they want that truck in 3d..not your representation of what you think that truck should look like in 3d.

Really it depends on the company and the amount of artistic freedom they give you. They could hire you cause it looks similiar but good in 3d. Or hire you on your ability to match concept.