Discussion about professional game models

[RO]

lol your right I have seen tons of those examples. I have found out that when I work on a game title my projects end up not as neat as if I was doing a personal project. Mostly because of time constraints. So you just end up doing things as good as it will get within that time period. Time constraints really kill you.

That is why some companies prefer to get modelers that are quick with good quality instead of slow but awesome quality modelers. I have seen this happen time and time again when companies hire people.


At the end the consumers usually have no clue about game art and will not notice errors like these. Games are mostly about game-play and a few art assets can be messy and not really harm that at all.


“WOW” has tons of these errors but the game is most likely the best morpg ever.

Edit: HAHAH at the cars wheeels!!!

[MightyOne]

LOL, NICE THREAD...

see... my point of view.... the leaks, bad shapes, the seams, jaggies, will not affect the game play, in the first example, doom3, you will shot in the enemy so fast that IN GAME PLAY, u see the enemy shots... 5, 10 seconds later it is history, is so fast that don't make any diference (? make it sense?), on the game industry (on movies too) u have rigid deadlines, u need to do your job then go to the next, it need to be fast... do your work, game industry is a serious business...

anyway a big % of players dont have the critique eyes that u me, and the people like us (who really understand artistc development in games, movies, tv etc...). Maybe i would be wrong (man if u understand it you will understand anything).

Conclusion... if in the game play it will not affect, why you need to do perfectly? it just need to be good... some major titles a lot of things is really ugly, but a great part of it you have dificult to see..

Sorry my BAD (today is tooootaaaalll bad) English
post your opinions

[xX_eXiGe_Xx]

When you/we did'nt have a clue about modeling, animtion w/e, then we where mostly thinking about gameplays. Therefor, how the models looked where always O.K for us, sometimes we even "wowed" on the "graphics", you know?

[Gadric]

sometimes, the engine makes mistakes....sometimes the artist. Not ever you have the time to make it perfect :D

On the screenshot, you can see a "little" bug on my gfxcard :rolleyes:

Gadric

[xno]

Just being curious, what game is that Gadric?

[CodeNothing]

"professional" is a loaded title. All it means is someone was paid to make it, not that it is of high quality.

Your right, its good to point things like this out to 3Ders trying to make it into the biz. It can be very frustrating to post a hard worked on model only to have dozens of people nit-pick a hundred different things on it.

[itsallgoode9]

ya know, it's kind of funny because since I started working in the games industry i've been really worried that my skill level is actually gonna go down due to stuff like this....where you just have to get it done quickly....many times even if that means it looking like crap. i've been getting alot faster, but i don't think i've goten any better and when I get a chance to go start working on my own stuff that's more realistic and complicated (as compared to the lmitations of game engines) that i'm gonna be kinda lost :D

[LiveWire]

this thread makes a really good point. i remember when i was at university learning 3D, i looked closely at models from different games for construction examples/ideas and to help learn techniques. what i found shocked me. as has been said i would post on the web and have every little seem and bit of stretching pointed out, but then i look at the 'pro' models and what do i see?

UT2003/4:terrible deformation (a result of the same animation set on every character of the type no doubt), large seems, smmearing and on some even obvious textel resolution variation. what's funy though is that you never notice this stuff in game. i guess it goes to show that in the end it's how it's seen in game that's important.

Doom3 on the other hand, you could see faults all through that even in game. the low poly model sihlouettes ruined the ormal mapping effect many times for me, and i saw a lot of texture errors, such as stretching, and plainly obvious seems right down the middle of the character's heads! (the normal maps screwing up), not to ention the obvious symetry of the textures.
while i loved the work in game in UT and never noticed the errors, Doom3 on the other hand were so plainly out there that i wondered how it ever got by QA.

[Neil]

"professional" is a loaded title. All it means is someone was paid to make it

That is my point. These artists got paid to create this work :)

I know deadline exist, but still texture stretching and such only take a minute or so to correct as do deformations and such. So that's why I get frustrated seeing that exist in the final model. And I purposely tried not to pick out errors that would be associated with glitches.

[AdamAtomic]

There are some things that I think are very forgiveable, especially deformation and stretching during extreme poses - usually these are only visible on-screen for a fraction of a second, so frequently they will only be visible in screenshots, not in-game. It is a sort of 4th dimension to UV mapping that needs to be part of the compromise. For example, that zombie's mouth, is it really worth doubling its texture space just to make sure it looks perfect during the apex of the bite animation? To me, maybe not.

However, things like stretched UVs on static objects like that machete, that's just bad, that is an easy fix, and things like that really can make a difference in the presentation quality of the game. My current clients are being real sticklers for stretched/deformed UVs, even on the LOD models, so some people out there really are paying attention.

And as for the DooM3 normal maps, I didn't even know you could actually see the monsters in that game :D much less any texture seams!

[RmachucaA]

hehe, very good topic :D.


I would like to add Codemasters Pro Race Driver 1-2 for some of the worst vehicle models, some are good, but most of them have either dreadfull normal issues or missing polys, holes you can see through the car and just plain louzy modeling.

[Hourences]

While you shouldnt go over the top with technical problems, theres a limit, you can and will always have some around. I always iritate myself that a large group of people seem to care more about technical problems such as a messed up uv thing than about the general styling, atmosphere, emotion, design, etc of the thing. If I had an 1 extra day to make something Id rather spend that day on being able of making a cooler design or a better and more artistic stylish design for my model than to put that time in getting some technical problems correct. Id rather have a very great styled thing with a couple of small bugs than a very average and thus boring mediocre model albeit technical perfect

Style, art, etc > strechted uv's etc.

However, lots of pro game modellers dont do that either :)
Basically bugs are inferior to styling and stuff, if you were a lead artist you should first look to if it was styled and designed well and only after thats done well you should complain about bugs.

[aeres]

I was looking at the last screenshot when I thought out loud "Holy sh*t! square wheels!"

[Augh]

I was looking at the last screenshot when I thought out loud "Holy sh*t! square wheels!"

lmao, you too huh?

That one truly is eyepopping. Is it from that Matrix game? Maybe better not to say actually.

Wow! O_O

[SHEPEIRO]

obviously GTA is the grandaddy of bad texture placement (stretching). but in a world of that scale its bound to happen. and it happens mostly in the country side where you dont spend much time. with greater amounts of detail this is going to increase in preponderence. but decrease in size on screen.

like the contrast between the zombies jaw and the hugh texture stretches in GTA, i believe this will become less problematic in future, but always depends on good quality checks. someting i believe is not done too much in terms of graphical probs, rather than programming probs, in Q&A

[capone_adam]

Before the last post I was going to put something like...

Square wheels...terrible, but does it REALLY matter to the average player?...as long as the car moves nicley do they mind? I decided not to post that in the end as I think graphics help to make the experience/atmosphere more richer.

I mean of course we easily spot these things but I would say the majority of gamers don't even know what a 'seam' is.

but then, shepeiro rightly brought up GTA...the texture job in these games are dibolical. But what would you prefer...imagine GTA with perfect textures/modeling but due to extra time spent on texturing/modeling they had to make the game 1/8 the size.

What I'm saying is...one of the most succesful game series of all time is a game that lacks graphical polish and doesn't care about perfections.

[Ingurum]

really nice discussion.

but just to mention that square wheels from the enter the matrix game are caused from the options.. on some pc systems the game had set an option to low as default and so it had square wheels.. i think now you just looks at a lod model or the engine scaled the polycount down ...

jeh jeh i didnt like the game really but i need to defend that poor artist right? ^^

[Headless]

I agree with Adam Atomic. Once you've got animation in there, you're not going to notice most of these. Take the Resident Evil example: that kicking motion lasts about a second in total, so you're looking at about 3 frames where you'll have the problem. You'd have to be a humming bird to notice.

Alternatively you've picked some examples where the focus is on serious amounts of optmisation because there's going to be a ton of stuff going on on screen. The Lord of the Ring example, for instance, may be off in that screenshot, but this is a game where you're supposed to have hundreds of those guys on screen at once. See if you can see the same issue here...

http://www.gamingnewz.co.uk/pictures/screens/1104749099_LOTRtbfme08.jpg

[maul]

actually the square wheels are lods.. i remember that you could check something with LODs in the graphic settings then it became like that.. Uncheck it and it will look round

[ArchangelTalon]

I think some things have to be forgiveable. There's a timescale to make things in and with the next/current gen games the workload grows ever larger.

Doom 3 I just find to be generally poor, when you have a game that relies on normal maps sooo heavily, then I think you really should make sure that they're good, especially with iD's infamous "when it's done" timescale.

On the other hand, you have to weigh out the effort involved with the end result improvement. After all, the enemy's on the screen for maybe 10 seconds before you kill it and move on (and in D3 you're lucky if you even see it once in those 10 seconds :p).

The examples you've picked here are not great - however I haven't a clue what's happening with that horse - but I do know what you mean. I've looked at some of the really big name games and thought "Wow, I can do a lot better than that" but when you're doing it day-in-day-out to probably tight deadlines on a project you may not be entirely enthused about working on, then I guess you tend to kind of slide/economise a bit.

[Psyk0]

actually the square wheels are lods.. i remember that you could check something with LODs in the graphic settings then it became like that.. Uncheck it and it will look round

Yes he's right those are LODs, still funny tho!

-"Alright who's the dumbass that made those square wheels?!?!"

-"Euhh...it's a matrix bug?"

-"You're fired!"

[JaMo]

ive been noticing how most of my favorite games have gfx problems. i dont even bother going around looking for them anymore because once you see it you will always see it when you play. metroid prime for example is one of my favorite games but it has some areas where there is a hair line where it looks like the uv is off one pixel or something. PoP:WW has the same thing. so yea i dont tear apart my games anymore unless someone ask me to.:)

even with all this im still worried because im getting ready to try to break into the industry and i know that imperfections are probably less forgivable when it comes to demo reels and trying to get hired.

[Headless]

Doom 3 I just find to be generally poor, when you have a game that relies on normal maps sooo heavily, then I think you really should make sure that they're good, especially with iD's infamous "when it's done" timescale.
Quite.

It's something I always thought about Halo 2: If you're going to normal mapping, you have to make sure that you're normal maps are high enough res or it'll just look rubbish (http://nikon.bungie.org/screenshots/halo2review_110704/singleplayer/cinematics/chiefandcortana.jpg). Personally I think UE3 normal mapping is the first example of it that actually looks any good. I think there are so many games that look so much better just because they got the base textures right.

[HellBoy]

too bad I can't get screenshot but in KILLZONE on PS2, when you're shot dead theres huge error in the players mesh.


anyway, could it be that modellers done great job but programmers let them down?

[gshouston]

I don't know where everyone's even going with this.

If you're wondering why everyone nitpicks your models to the Nth degree when you post them, it's because they're giving feedback. That's what you should be posting for anyway. If you're posting your models just to get people endlessly praising you for accomplishing something mediocre, then you're not a serious artist. I for one am pretty sick of people posting stuff in the WIP forum, and then saying "I HOPE YOU LIKE IT!! LOL!!1"

As for the doom3 stuff, yeah, there's errors on those characters, but until you can make something even that good, you've got really no right to trash it. Especially when the 3D art dept. consists of like, 6 guys.

[aeres]

I don't know where everyone's even going with this.

If you're wondering why everyone nitpicks your models to the Nth degree when you post them, it's because they're giving feedback. That's what you should be posting for anyway. If you're posting your models just to get people endlessly praising you for accomplishing something mediocre, then you're not a serious artist. I for one am pretty sick of people posting stuff in the WIP forum, and then saying "I HOPE YOU LIKE IT!! LOL!!1"

As for the doom3 stuff, yeah, there's errors on those characters, but until you can make something even that good, you've got really no right to trash it. Especially when the 3D art dept. consists of like, 6 guys.

Read the original post. This thread was created in the spirit of sharing the fact that even professional game meshes are not perfect, hence those of us shouldn't let ourselves down if we can't create perfect meshes ourselves (yet).

Why would you think if the viewer can't produce work of such quality immediately revokes his right to criticize? do you think all movie critics are filmmakers? Mind quoting anyone on this thread who actually "Thrashed" those screenshots?

[Neil]

too bad I can't get screenshot but in KILLZONE on PS2, when you're shot dead theres huge error in the players mesh.


anyway, could it be that modellers done great job but programmers let them down?

With Killzone, yeah they used a heavy amount of LOD with their characters and when you run up to someone really quickly it still shows the lowest level before it can load up the higher model. I don't think that one was an artists or programmers fault, more just them trying to do more than the PS2 could handle.

[Neil]

As for the doom3 stuff, yeah, there's errors on those characters, but until you can make something even that good, you've got really no right to trash it. Especially when the 3D art dept. consists of like, 6 guys.

But there is plenty of stuff on here that is that good or even better. Goes to show how many talented people are out there. Gotta always remember that the people that work at id are not gods, just talented guys/gals that are not perfect, and posters on here could fill those shoes with continued work and dedication.

[heavyness]

i don't think its the artist or the programmers fault. the majority of errors i find are because the code is taking over. like physics or lod. how many times have you died in a game that features rag doll physics to only fall through the floor.

also, some of these mistakes are only visible once they game is close to complete. sometimes some code has changed that will effect polys in a different way. even the littlest fix near the end of a dev cycle could mean a lot of work.

but like said before, most of these mistakes are never seen since their is so much going on the screen at one time.

[MFTituS]

This thread was created in the spirit of sharing the fact that even professional game meshes are not perfect, hence those of us shouldn't let ourselves down if we can't create perfect meshes ourselves (yet).

but the difference is, the pros can make perfect meshes if they want/have time.
dont use such flaws in games to explain your own mistakes. just cause they make some, you dont have to learn to create perfect meshes and textures?
first learn to make professional work, than you can try to make the best out of very limited time you get. otherwise if you cant achieve the quality, how do you want to make even solid work in half the time you normally invest for your models?
mistakes happen. you have many objects to do on your list, so you are using meshes of other artists made before, copy them, merge them together, and many things more to rush those things out. and mostly at the end you havent time for polishing all the stuff as it was promised at the beginning.
so texture seams, smoothing errors and many more persist till the final product.

i dont want to attack someone, only explain why such things happen.
at the end these screenshots just show the sad side of the industry. no money,no time. things are rushed out, errors stay, not everyone can employ only "good" artists, sometimes salaries are more important than the quality, graphics are outsourced to people/ teams you even dont know, polishing is a word you mostly know from hearing.....
and often all the stuff will not be recognised by the player...he just looks on his monitor superficially and either he likes what he sees or not.
so you have think twice if it is worth to invest extra time for such things no one will tap you on the shoulder.

but it is nice to see, that even the big/successful studios cant hold out against a trained eye. not only the small studios have to make compomises.

[vrapp]

MFTituS: Word. It's one thing working on a personal project with no deadlines and another thing entierly to work on a project together with tons of other people, with some material being done inhouse, combined with out-sourced assets and tight deadlines and no extra time for polishing.

[SHEPEIRO]

interesting topic

its just a fact that game models have to cope with alot more. than say a film or demo reel model which will be optimised for every scene/pose/whateva. its not surprising that as games allow you greater freedom *(in many ways, not just exploration, but view distance, actions, etc) these models need to be more and more robust. the games industry is always trying to push the graphics as far as they can, so they are always rough edges. until processing power levels flatten out this will be the case.

[aeres]

but the difference is, the pros can make perfect meshes if they want/have time.
dont use such flaws in games to explain your own mistakes. just cause they make some, you dont have to learn to create perfect meshes and textures?.....

Ah, maybe my post have been misunderstood. I did not mean that one should use the flaws on professional work as an excuse not to improve one's skills, quite the opposite really. Viewing these images actually made me thought : Hey, they are not perfect too, I will be able do better than that....hence it really motivates me than make me think that flaws are acceptable.

But, that's just me.

SO...back on topic.

[DingBat99999]

Here's some input from someone who's made programming a profession.

This is where theory runs headlong into reality.

It's theoretically possible to create a bug-free, perfectly coded piece of software. Unfortunately, that perfect piece of software might also take 100 years to write.

EVERYTHING is cost-value. If I asked you: "What would you like the Pinto or the Ferrari", everyone would pick the Ferrari. It's only when I tell you that the Ferrari costs half a million bills that the Pinto starts to look attractive.

Remember, the goal of a game development project is to make money. I'm pretty sure that "Make perfect models" ranks quite a ways further down the list of priorities. And, since games are a commercial venture intended to be sold, I can't really argue with the priorities.

I'm NOT condoning shoddy work. I'm merely pointing out that Doom3, warts and all, made more money than <insert potentially perfect, cancelled game here>.

[Neil]

I'm NOT condoning shoddy work. I'm merely pointing out that Doom3, warts and all, made more money than <insert potentially perfect, cancelled game here>.

Very true. Everyone has brought up good points.
But wasn't Doom3 also finished on a "when it's done" attitude? They aren't under the crucial deadlines that everyone else in the industry is. Granted, they're not gonna release it in 2008 and have it looked dated already, but as a modeler you can fix stretching and seams in an hour or two and that shouldn't change the code or impose any new bugs since it's already loading the exact same stuff, just more polished.

[DingBat99999]

But wasn't Doom3 also finished on a "when it's done" attitude?

IMO, that attitude is for show only. Even id can't afford to stick with it, there's too much competition.

Personally, I thought id admitted that when they rushed Quake3 out the door to beat UT.

[florian]

Hi there

Just wanted to mention that I think, the doom3 models are one of the best game models ever. Compared to the rest, they are very low Poly, but the artists made the best of it.
Design is top and the "mistakes" you showed apeared because of the stencil shadow and the polycount. If something was not as cool as expected, it was the engine itself.
I'm not a programmer. But I guess its more hard work to make something good looking run smooth than using all dx9 features and have 5 frames per second.
Correct me if I'm wrong. :shrug:

[zeebit]

i thought this was gunna be a cool bloopers thread that showed the shortcomings of games, whatever they may be.

sucks it got into a bashing thing

-z

[Headless]

One thing a few people keep saying is "it only takes an hour or two to fix a problem like that", but they should realise that those "hour or two"'s add up. The artist may well have spent several "hour or two"'s fixing issues to get the model to where it is in these screenhots, but there's a point where you have to stop and move on to the next thing or you could be just taking that extra few hours forever.

I also agree with DingBat99999: With something like Half-Life 2 they have some pretty perfect art assets, but then that level of perfection took six years and cost 40 million dollars. Compare that to something like GTA:San Andreas, which has big texture issues but still massively outsold HL2, and being only two years in development was probably a damn-sight cheaper to make.

[sathe]

Its very hard to balance the development of a game, especially with so many variables, so i think games companines are doing a very good job and are steadily progressing however beauty is in the eye of the beholder so no matter what we see, were gonna have our own opinions.
Personally i think, knowing that games companies art perfect does ad a little motivation but when did they ever! get the reputation for being perfect in the first place?

[MFTituS]

thats the nice or sad thing about free camera in 3d.
in old good 2d games you have made graphics for one view. you have known 100% what the player could see and what not.
but know you can see behind the magic. it is nothing special, limited ressources, unlovely areas....the only difference is, that you know can see every little flaws, but they existed all the time....



posted by aeres
Ah, maybe my post have been misunderstood. I did not mean that one should use the flaws on professional work as an excuse not to improve one's skills, quite the opposite really. Viewing these images actually made me thought : Hey, they are not perfect too, I will be able do better than that....hence it really motivates me than make me think that flaws are acceptable.


ah ok, understand you.
i wanted to say, that my post wasnt meant to offend anyone. i have just tried to speak in general (what is hard for me in english), cause i didnt know into which direction this thread was going. so i just quotet your words, cause they came at the right time :)

so happy modeling
TituS


florian wrote:
Just wanted to mention that I think, the doom3 models are one of the best game models ever. Compared to the rest, they are very low Poly, but the artists made the best of it.

hm, maybe if they had not forgotten to put the textures on the models :)
everything looked like the same grey mass.....
then better the great models from ut....great models and the same quality work for the textures.

[wanzai]

i've been getting alot faster, but i don't think i've goten any better and when I get a chance to go start working on my own stuff that's more realistic and complicated (as compared to the lmitations of game engines) that i'm gonna be kinda lost :D

Hey, I can relay to that... during the last 4 years or so in the game dev I got faster more than I got better. After a while you do learn to get "good enough" results in the given time, but improving your quality from there is rather difficult, IMHO. By "good enough" I mean the stuff really can work well in the game but it's not "CGtalk frontpage" quality. And as for me, improving my skills from that "fast enough for production but not necessarily portfolio"-level mostly came from personal projects, if any.

Also I'd like to note the mind plays a big role too. You can really burn out and just stop caring about that final 10 % to perfection when you have to do your stuff quickly, constantly; and more importantly, when you realise that those details actually don't really matter. It won't make or break the game, and your extra efforts are unlikely to be noticed, maybe on CGtalk, if at all and only if you worked on those big names.

I'm not saying "don't strive for perfection", but these are the realities. If I think about it a bit, I could barely name a handful artists whose production game models are frontpage quality and manage to wow us trained eyes. Apart from that, right off the head I'd say I'm impressed by the models in MGS3 and Kingdom Hearts (2) - and RE4, but than again, even RE4 is an example here, and I didn't inspect those 2 other games now.

Hope I'm making any sense at this late hour...

Interesting thread anyway.

[KINGOMONKEY]

I think there is an interesting point being made here. I just saw this site and thought id suggest some peopel look at it. This is a game coming out from a Primo company and you would expect the texturing to be the same..
however cos its so high detail, you can really see the mistakes, especially where mirrored maps meet and where necks meet torsos.

It does help me to stop worrying AM I GOOD ENOUGH and just apply for some jobs, someone somewhere must want me?

anyway heres the link to the Quake 4 shots

http://www.gamersgallery.com/gallery/showgallery.php/cat/6072/page/1/perpage/12/what/allfields

[BoBo-the-seal]

Time per given asset is the number one issue but restrictions within polycount and more importantly texture resolution causes a lot of the problems. You’re going to see a lot of issues in consol games simply because the restriction on texture size is so great. You have to do your best with what you’re given. Also you’re picking apart screenshots. I have never seen a game designed for screenshots. Most games are designed for the player to be constantly on the move. When evaluating your work in a professional setting you have to make sure you’ve done your best to make something that look good within the game from its given camera (based on time given, texture and polygonal restrictions, and overall art direction) and at what point within the players action this asset will be seen.



When you are creating a piece of art for your demo reel its imperative that you spend the time to make it the very best you can with as little issues as possible. Honestly it’s expected that you make it as flawless as possible because your target output isn’t usually games (although it may be an exercise in game art production). Most people on these boards are making “game art” for demo reasons. These assets will never find their way into a game engine. Therefore their target output is really for renders. Since it’s for renders and to show off that one particular asset it has to be as good as it possibly can. It’s the focal point. Games on the other hand are a sum of its total. It’s silly to break apart any one asset when everything is created to work together (again based on the above criteria).

- BoBo

[dpizzle]

damn bobo...
thanks for the info
:thumbsup:

[CGTalk Moderation]

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.