ZBrushCentral

Zbrush, Ideal for games?

ZBrush Usage Questions
#1

i just downloaded and saw zbrush’s power. it can create almost any models without pain. but as i have found that most of models created by zbrush are very high res to be used in games. so do you guys have any idea about polygon reduction or using some other method in zbrush to add polygons to only those areas where detail is needed? i saw zbrush’s optimization tool which decreases the polygons but it cannot be used on a model created with zspheres. and yea is there any polygon/vertex editing tool in zbrush so that we could edit faces manually etc. And 1 more question do you guys use any other package with conjuntion to zbrush to achieve something before starting out in zbrush. thanks.

P.S.S. : Whats the ideal num of polygons for a 3d game?

#2

You may want to check out this box modelling tutorial thread posted by Aurick.

http://209.132.68.66/zbc/showthread.php?t=20370

However, the capabilities shown in this tutorial aren’t available in the current demo.

Also, check out the adpative skin ZScript tutorials that come with ZBrush or any that are posted on the forum which can give you lower poly models when using ZSpheres.

#3

NiceMan, it all is very dependent on the architecture you’re working with.

If you’re making a game that will only show two models on the screen at one time, then perhaps the sky is the limit as far as polycount goes.

ZBrush is best for character sculpting, so for a typical character you could go for anywhere from 2,000 to 10,000 polygons.

Current first person shooters use anywhere from 4,000 to 10,000 polygons per character model. If you’re talking about a game console, such as xbox or ps2, then the count will be much lower, probably in the 1,500-2,500 range.

When you’re making games, however, there are different kinds of artists. Probably the most rewarding 3D modeller to be at this time would be a source artist. That would be the person who creates the highly detailed high polygon models and material.

This is later downsized so it can be used in a live fashion, but now with normal map technology, and other things on the rise, you can get high poly looking results on low poly models. In sense, the source artist is no longer ‘rejected’.

Before, source artists would make a huge model for an example on how to build the low poly version, and the high poly would never make it to the game. At best, the high poly model would make it on box art or a magazine cover because it looks better.

Now, you can see both. Example: Look at DOOM 3. They use normal maps to make low poly models look almost like the high poly version.

#4

In theory, ARIZONA is correct, a high poly artist may soon be more coveted. But as of today, normal maps in games are simply a gamers dream and are not yet popolar enough to be demanded in all, or most games. The hardware support is not there at all. Sure modern hardware is somewhat capable. But Doom3 just came out the other day and most high end systems can hardly handle it, and Doom3 usually has very few lights running, imagine what will need to happen for a more realistically lighted scene. Possibly normal maps can make it into every Xbox game (Such as ESPN football 2005) because it’s hardware is capable, and the releaser knows the profile of the machine it will be played on. As for PC gaming, this is not the case and the releasers try to make the game for every hardware profile incluing the oldest hardware profile, so that their game can sell as many copies as possible.

The intent of ID with Doom3 was to push the current hardware to the max, and do something that had never been done. The result took 3 years in the making, and not even ID wants to spend that much time on their next game, and very few game compnaies are capable or willing to invest to have that good of graphics.

For at least the next two years, GAMES MEAN LOW POLY. It’s hard enough to find work as a game artist. Finding a job as a HIGH POLY game artist would be next to impossibe. And the jobs that are available for high poly game work would be reserved for the gurus that are highly capable of doing such work.

SO: In the future High Poly & Zbrush will find their way into the Big picture. But for now, it is only our dream.

#5

I don’t think so…

Right now I can name off a few games on the PC that use normal maps…

Doom 3
Half-Life 2
Joint Operations: Typhoon Rising
FarCry
Quake3 (yeah, you heard me right.)
s.t.a.l.k.e.r.
Halo 2

DOOM 3 is still a baby, and when the mods start showing up it’ll bloom. Half-Life 2 is going to be a monster because all the C-S people will grab it up real quick.

The trend is already going to the consoles.

Look at it this way, man…

The three contenders for GAME OF THE YEAR all have normal map technology as a feature.

DOOM 3, Half-Life 2, Halo 2.

There are leaked Quake 4 screenshots of source models that Raven is using. High Poly artists are now essential, and you can’t get that detail out of a program unless it starts with Z and ends with Brush.

#6

I really disagree. YES the big games are using them. YES that is why they are game of the year nominees and spark interest, because they are stepping into new ground as far as hardware capabilities go. Xbox is the only console capable of normal maps. I love the idea of normal maps and high poly content.

On the other hand, these are the also the #1 game companies that have been spending 3 years or more on the games that you mentioned, and they are the cream of the crop studios. If you are suggesting that someone has a chance in hell of getting on the Valve or the ID team, you’re crazy. The guys that get there are the guys that have done all of the TRADITIONAL low poly game work to get where they are now. If you are that good at low poly, we’re pretty sure that you are capable and have toyed with high poly PLENTY.

Also, these are the games that are in the spotlight and are from HUGE studios. For the REST of the world and the small to normal sized studio, they don’t have the time, resources, or money for licenses for the engines or the tools, and want to target the entire realm of PC users, not the top of the line users. They represent a small portion of gamers and customers.

Look at normal maps right now as the Dodge Viper. They show it in all of the commercials and it sparks people’s interest. Yet how many people do you know that drive one, or do you ever really see one on the road. No . They are the mascot for Dodge.

So if you have a passion for designing cars, you may very much like designing cars like Vipers at home or for fun. But if you want a JOB in that industry, you’re going to be working on the TRADITIONAL models, that the normal people are going to buy and use.

#7

Firtst of all, normal maps are becoming … well … normal. Even the low-cost engines like 3D GameStudio, DarkBASIC Pro, etc are able to utilize normal maps … they are simply a form of a shader. So, it is not just the AAA engines that enable people to utilize normal maps.

In any case, I do not personally think that ZBrush, in its current form, is a good tool for real-time 3D games and applications. It is not a very good low-poly modeler (even though you can do it … it just is not as friendly as doing it in a traditional modeler like Silo, MAX, etc).

Secondly, ZBrush is not suited for painting a skin for a real-time 3D model. Often times RT3D models are limited as far as the size of the skin. Skin sizes may be something like 512 x 512 pixels or 1024 x 1024, etc. Texture space is at a premium and if you are planning to display more than one model and also have textures on your level geometry, then you need to keep skin sizes down. Size the skin size needs to be kept down, then a good UV map is a must for any RT3D model. Extra space on the skin is givin to important areas like the face and much less area given to unimportant areas like the underside of the shoes. We usually use a formula like 1/4 of the space for the face and then fit the rest in and scale it according to the details needed.

ZBrush, as far as I know, does not have a way to assign more texture space to one area of an object than another. Therefore, you need to create a UVW map in another app outside of ZBrush and then import that model into ZBrush for painting.

To conserve texture space often UVs are overlapped. For example, why have two undersides of shoes, especially if the texture on them will be exactly the same? So, the skin artist will place the UV for one sole over the top of the other. This frees up room on the skin for other elements. Sometimes we will use only half a face and mirror/overlay it on top of the other half. There are all sorts of tricks like this to maximized texture/skin space.

ZBrush does not support UV mapped models with overlaying UV faces. If you attempt to paint them you will either get strange effects or ZBrush will lock up and then crash. So, if you UV map your models this way (as is very common for game models) then you cannot paint the model in ZBrush (as it is).

Oh, there are some tricks that could make this work (deleting the parts of the model you intend to mirror, create the UV map, export from your 3D app, import into ZBrush, paint it, export from ZBrush, import into your 3D app, duplicate the missing parts and then merging the model), but it is a bit more work than taking the traditional route of simply exporting the UV map and painting it in Photoshop.

I would love to fit ZBrush into my pipeline for RT3D content creation. It looks like the only way I will be able to do that is if and when I start using normal maps. Then I can use ZBrush to make high detail models from my low poly ones.

Personally, I think that ZBrush has a long way to go to be utilized well in a game environment.

#8

Hi Folks!
Rick Vazquez here, from Liquid Development. – www.liquiddevelopment.com – I’m glad to see this discussion here. Dan is right. :+1: He’s accurately described the value and momentum of ZB in use for normal mapping game characters.

This is my first post on this forum - in fact, I must admit I’m fairly new at this altogether - but it’s my first one since the job posting for Liquid Development. If you’re reading this and haven’t seen Liquid’s job post, just do a search for our company name here - it’s in both the main and common forums.

Curious, there are a number of people doing really good work in ZB, I mean just fantastic artwork! But there don’t seem to be many with the Max/Photoshop/Zbrush technical ability combined with great texture skills that can do what Dan is describing. We’re in need of just such a person, and I’m kinda having a hard time finding people to even qualify for a test.

So… I guess I’m posting here to ask for some help.

An artist doesn’t necessarily need game character backround, but needs to be able to follow the steps, use the software and make great normal maps, textures and shading. For our current project, the models will be already made for the most part. So like Dan described, artists will need to be able to UV in Max 6, bring to ZB, and so on…

Thanks for the topic. And I look forward to any feedback.

Rick

#9

I don’t think it’s such a long way, and I suspect that the folks at zbrush are already discussing this concern. Integrating zbrush and normal map generation for lo-poly game assets is a priority-one issue for me right now. Here’s to hoping that our presence in this forum topic is helps to urge zbrush to address this need. Soon. :smiley:

#10

wow