Male And Female Templates To Come From Squawks This Weekend

[Domarius]

Squawks has developed male and female templates that she's going to package up and give to me on the weekend. Sorry I didn't let you know earlier but we weren't sure when she was going to finish.

 

Both male and female come with fully naked versions and clothed versions, the clothed versions are meant to be used as a reference for making future clothes. The female clothes are a barmaid version, and a thief version, while the male clothes are some generic townsperson clothing.

 

The heads are seperate already, and ready to go through the process to make them head swappable in Doom 3.

 

But as meshes, they have switchable parts that you can use while modelling.

 

You can switch their hair (both male and female) because they are bald and come with seperate hair.

The arms switch as well because the dress already has the arms on it, so you just put the hands on the ends (the hands swap as well) They also come with full arms for costumes that don't have sleeves, like the female thief costume she's made.

 

She said not to touch the geometry of the faces too much because she'll be making different diffuse and normal maps so that you can interchange the look by swapping textures. You can use those textures as a guide for making new face textures. When making new textures, keep the eyebrows etc. in the same place because the normal map will have to be modified if you change the positions of anything in the diffuse texture. Which you can do if you're up to it.

 

Her goal is to make 4 to 6 basic head types that you can make variations of by just changing the texture. There will be a 20-30 year old type, elderly type, child type, fat types, etc. and of course each type comes with its own textures (because you can't swap textures between types).

 

She said she's made the current faces as generic as possible.

 

And she knows we don't have much use for animals, but she says she has a high res horse she made for something else that we can have, if we need horses, but it hasn't got a low poly version yet. She's offering it as is, if anyone wants to make a low poly version.

 

The basic Model template

-----------------------------------

• Head - swaps out to different types, generic faces of 20/30yr, 50/60yr, children.
  
• Body - Basically is the clothing /dresses/thugs, ranging from top to ankle.
  
• Feet - interchange boot styles etc
  
• Hands - Any model that has clothing that covers up to the wrist will have interchangable hands/gloves what not
  Models that show skin from elbow to hand will be joined.
  
• Hair - different hair types swap between what you want.
  

Obviously some will have the hands joined and some might have feet joined depending on what the person is meant to be wearing and how much skin shown etc.

 

As a guide for poly count, the female with the thief clothing on is 4,458. So they are ready to go, no need to delete unessecary faces (and thus screwing up the maps) when putting attatchments together.

 

AND... look ma, no mitten hands! They have all their fingers. She says that you shouldn't have to resort to mitten hands with a budget of 5000 polys.

 

Hey I just thought, the bonus with all the body parts being seperate meshes, is that we can apply different materials easily to each body part, for armour vs skin collision.

[Domarius]

I know ideally we want to have her here and talk to her etc. but she really wants to work through me. At least this way we get some good quality models instead of ascottk having to scrape the bottom of the barrel with models like that demented noble guy. And these templates will really make it much easier for future modllers to stick to the same style and develop different varieties of people really easily.

 

And once I get Doom 3 installed on her computer she can come closer to meeting us halfway because they will already be sucessfully exporting to Doom 3, just not with all the features ascottk has been adding.

 

Also she asked me to apologise for her for taking so long, because she's been very busy.

 

And the pictures will come on the weekend, she's going to send me a complete package of everything.

[Domarius]

I begged her for one pic, so she caved in and made this sneak preview.

 

Heh... sneak...

 

[sparhawk]

I don't really like that scar. Reminds me to much of T3. Otherwise it sounds quite good.

[greebo]

She definitely has some talent here, I really like the overall look and the face proportions are very well done.

[Domarius]

She was asking me how she should do a hood, and I sent her T3 pics of Garrett... so that might explain where she got the scar from It's just on the diffuse map and can easily be removed.

[New Horizon]

She was asking me how she should do a hood, and I sent her T3 pics of Garrett... so that might explain where she got the scar from It's just on the diffuse map and can easily be removed.

 

The beauty of modelling. Looks very nice from what I can see in that little preview.

[Gildoran]

AND... look ma, no mitten hands! They have all their fingers. She says that you shouldn't have to resort to mitten hands with a budget of 5000 polys.

Yeah, it really bugged me that everybody decided to switch to mitten hands when there were plenty of less noticeable things that were wasting far more triangles...

 

For example, why the heck was this belt-buckle fully modelled? Altogether, including the center part and both sides, the belt buckle should have taken 2, maybe even 6 triangles (by having the belt-buckle drawn with an alpha-test material - the design philosophy of D3 is to put the detail in textures, NOT model them!), but instead it looks like it takes 100. The models are full of things like this and everything is subdivided more than necessary... such wasteful use of triangles is why a single guard entails a 20 fps hit! Don't get me wrong, oDDity is an artistic genius and I wish I were even half as skilled as him, but I consider his judgment for low-poly modelling to be ... exceptionally lacking.

 

Edit: Sorry... this has been building up and I just felt like venting.

[Springheel]

Hair - different hair types swap between what you want.

 

That would be very useful if it were set up so mappers could attach different hairstyles using the def_attach system.

 

And these templates will really make it much easier for future modllers to stick to the same style and develop different varieties of people really easily.

 

Sounds good. Hopefully these new models match the style of the already existing characters, since the existing guards are going to be the most-used models in the game. Not having had any opportunity to communicate with her about these issues is...unfortunate.

 

I don't really like that scar.

 

I assumed it was a tatoo--she actually reminded me quite a bit of that T2X character.

 

AND... look ma, no mitten hands! They have all their fingers. She says that you shouldn't have to resort to mitten hands with a budget of 5000 polys.

 

I'm not sure that having some characters with mitten hands and some without makes a lot of sense. We'll have to see how that looks, since going back and redoing all the hands of all the other characters is not an option I want to consider.

[SneaksieDave]

I don't believe we should adhere to some 5k poly limit for characters. Putting more than one of our AI in a scene starts to hit the framerate hard. There are plenty of places I have been tempted to simplify the existing AI (open them up and look at them - there are places where actual spheres exist, and sometimes beneath the skin!) but I assume that would screw up the UVs and such. They could probably all have 1k polys shaved off and still look as great.

[Gildoran]

I completely agree with SneaksieDave. 5000 polies may be ok for a special monster, but if we want to have any chance of putting 5 guards on the screen at the same time without performance taking a nose-dive, they really need to be trimmed down.

[SneaksieDave]

What would be needed to do this? I believe Blender has a md5 importer, but surely that doesn't include the UV. And I assume, screwing up/losing the UV would set things back so far that it's not worth it. And personally, I'd rather not have to use LW - I don't have it, don't know it, and don't care to learn it - Blender is tricky enough. Any modellers have input on this topic? Should we use a separate thread for this?

[Domarius]

So the existing guards are roughly 5000 polys?

[SneaksieDave]

I believe that was the "stay under" goal, but I don't know sure numbers, just observation that our models are much more complex than D3, and I believe they hovered around 3-4k. I just looked into the Blender md5 importer... and it ain't good. It sounds like it barely works, so there will definitely be no UV support. So much for that.

[Domarius]

Sounds good. Hopefully these new models match the style of the already existing characters, since the existing guards are going to be the most-used models in the game. Not having had any opportunity to communicate with her about these issues is...unfortunate.

She's been using pics of oDDity's models as a reference from the start. So whatever results we get from her is as good as if you told her.

 

I'm not sure that having some characters with mitten hands and some without makes a lot of sense. We'll have to see how that looks, since going back and redoing all the hands of all the other characters is not an option I want to consider.

I don't think it'll be that noticible during gameplay, not enough to warrant re-working anything.

Look at these two guys;

http://www.realtimecg.com/template.php?id=...zm.php&n=18

http://www.realtimecg.com/template.php?id=...us.php&n=18

I didn't notice that one had mitten hands and one didn't untill I went to look closely at the topology so I could learn something to make my own model. I really doubt I would have noticed during gameplay. I mean I might eventually make the observation at some point if I saw enough closeups of hands, but it certianly wouldn't look strange during gameplay.

[Gildoran]

BTW, I forgot to say this earlier but... I look forward to seeing more pictures of Squawks' work!

[Springheel]

I completely agree with SneaksieDave. 5000 polies may be ok for a special monster, but if we want to have any chance of putting 5 guards on the screen at the same time without performance taking a nose-dive, they really need to be trimmed down.

 

What system are you guys using?? I have only a mid-range system and I don't notice any serious problems from that many AI unless they start going into loops because of some stimulus they've detected, and that has nothing to do with polys.

[oDDity]

For example, why the heck was this belt-buckle fully modelled? Altogether, including the center part and both sides, the belt buckle should have taken 2, maybe even 6 triangles (by having the belt-buckle drawn with an alpha-test material - the design philosophy of D3 is to put the detail in textures, NOT model them!), but instead it looks like it takes 100.

 

Because the engine doesn't recognise the alpha when it comes to shadow casting, and I wanted things to cast a proper shadow.

Anyway, what's 100 polys in a budget of 5k

I was given that figure of around of 5k and didn't see much point in trying to sqeeze everything down to 3k.

Also, I was modeling these in the dark, before we had worked out all these facts and figures about how expensive several characters on the screen at once were, head swapping, attachments, locational damage materials etc

We haven't got the luxury of one piece meshes of guys in boiler suits like the doom guys could get away with, our designs are all wearing complex costumes, and alpha maps just don't cut it for quality.

Also, one guard does not give you a 20fps hit because his buckle has 100 polys instead of an alpha map, that's ridiculous.

Anyway, it's nice to see you obviously have a competent character modeler to take my place now. Her models are a bit generic for my tastes, the sort of models you see in any average game, she's obviosuly yet to develop a style, but they're good quailty.

[Gildoran]

Because the engine doesn't recognise the alpha when it comes to shadow casting, and I wanted things to cast a proper shadow.

The belt buckle doesn't stick out from the body that much, so I doubt the player would notice whether or not it casts a shadow. If worse came to worse, you could add some shadowcaulk (or maybe use forceshadows - though I don't know if that would work).

 

I was given that figure of around of 5k and didn't see much point in trying to sqeeze everything down to 3k. Also, I was modeling these in the dark, before we had worked out all these facts and figures about how expensive several characters on the screen at once were, head swapping, attachments, locational damage materials etc

Touche.

 

We haven't got the luxury of one piece meshes of guys in boiler suits like the doom guys could get away with, our designs are all wearing complex costumes, and alpha maps just don't cut it for quality.

I say we could slim down the models based off of looking at showTris in D3... I find myself doubting the player would notice if we were to go with a coarser, less subdivided mesh in many parts.

 

Also, one guard does not give you a 20fps hit because his buckle has 100 polys instead of an alpha map, that's ridiculous.

A single belt-buckle doesn't, but having numerous high-poly parts and a large number of subdivisions together can.

[oDDity]

Remember, that when I was designing these, we we also toying with the idea of cloth animation, which meant that that the cloth parts did need a certain amount of subdivision to be realistic, and in general, a certain amount of division is required for realistic bending in animation.

[OrbWeaver]

A single belt-buckle doesn't, but having numerous high-poly parts and a large number of subdivisions together can.

 

I think everybody is aware of this now, but it is the shadow polys that count rather than the render polys. Once the shadow meshes are created for the AI, I would be very surprised if there isn't a significant performance boost.

[Gildoran]

While I'm very interested in low-poly shadow meshes, I'm just worried that animation has an effect too, and shadow meshes could end up creating more vertexes to animate. But since shadows are more work than animation, I guess I'll have to hold my breath until we have some models that make use of them.

[ascottk]

While I'm very interested in low-poly shadow meshes, I'm just worried that animation has an effect too, and shadow meshes could end up creating more vertexes to animate.

I often wonder is there's a way to increase performance in animations. The md5anim conversion in d3 uses every frame in the animation even if the animation in maya only has 8 keyframes.

 

Suppose you have 4 keyframes in maya in a total of 48 frames (a hand at idle position, a key from idle to a raised position, the raised position, then back to idle position). The md5anim will use all of the 48 frames for that simple animation as opposed to the simplified 4 keyframes in maya. Then again, the 48 frame md5anim needs to account for the animation curves that occur between the keyframes.

[OrbWeaver]

Yeah, I'm pretty sure the Doom 3 animation system only uses linear blending between the MD5 frames, rather than the sinusoidal interpolation that the 3D app will use to create them from keyframes.

 

It might be possible to affect performance by using fewer MD5 frames during export, I don't know.

[ascottk]

The only thing I found on iddev are these export options:

 

-xyzprecision [precision]

Will take even tiny movements (translations) into consideration if you make this # lower. Default will try and compress down the animations getting rid of tiny movements.

-quatprecision [precision]

Will take even tiny movements (rotations) into consideration if you make this # lower. Default will try and compress down the animations getting rid of tiny movements.