Why not make things that fit together?

[Darryl]

Hi Steve,

 

I've been following your models for a year or two now, but so far I haven't been able to really use them in a game.

The problem is that there aren't enough models that really fit together. We can't make a pirate game with just 1 pirate. Or a magic game with just 1 wizard. Or a war game with just 1 soldier. Or a space game with just two characters and a ship interior but no exteriors etc. etc.

 

The medieval stuff is the closest we have so far to usable - with lots of buildings that could be used and a male and female base character.

But even then, they are differently constructed so it's not really possible to let players choose to be male or female without having to code everything twice to handle different ways of equipping each model. Plus there are still some issues with weights and animations that could use attention.

 

Would be really great I think if you could just kinda pick a theme and flesh it out so there's the complete set of all the different characters, buildings and items needed to make a full game rather than just doing one or two things for each theme. I think you'd sell more that way too...

 

Medieval/fantasy would be my suggestion - just make a base male that works the same way as the female and make loads more clothes sets that are available for both and then make loads of buildings, props, trees etc. that fit with the fantasy theme.

[Guest Steve Tack]

Yeah, I have to agree. I'm working on a sci-fi game, and the only way I could use the sci-fi characters is if the world was populated only by weird alien guys and space marines. By the humans in my game would mostly be merchants and other random civilians. What would be useful is a larger variety of faces with more generic clothing that could be used for multiple game types.

[admin]

This is interesting, and informative - thankyou. I always enjoy feedback. With respects to the generic sci fi characters in generic clothing, do you have any pointers as to what your after here? More faces on the sci fi base character is certainly doable. Generic clothing - yes. With respects to the medieval content, and indeed the new Pirate - surely those all fit together. Counting all the styles in the medbase, medfemale, and the other characters, there are many many character variations. What other characters do you wish to see here. Thanks for all the support, its very welcome.

[Darryl]

This is interesting, and informative - thankyou. I always enjoy feedback. With respects to the generic sci fi characters in generic clothing, do you have any pointers as to what your after here? More faces on the sci fi base character is certainly doable. Generic clothing - yes. With respects to the medieval content, and indeed the new Pirate - surely those all fit together. Counting all the styles in the medbase, medfemale, and the other characters, there are many many character variations. What other characters do you wish to see here. Thanks for all the support, its very welcome.

 

Hi Steve,

 

Thanks for the reply. I want to start by apologising for going direct to the request without first saying again how great the models are in general - this is probably the best model resource on the web for people trying to make games on a budget.

 

I think for the medieval/fantasy theme there are quite a lot of models you've made that do fit the theme. The main problem I have is that they work in slightly different ways. From a game coding perspective it would be great if all the models had the same skeleton, same animations, same equipping method etc.

I would like to see a male that works as the female does, so you can make them naked and add clothes on top. Then things like the pirate could just be clothes packs and there could be male and female versions of it.

That way we'd be able to just have one character controller script that works for all cases.

 

I really like the sci-fi stuff, but for it to be usable I think it needs a few more kinds of clothing, props, ships, environments etc...

 

I suppose what I am really trying to say is that rather than working on one sci-fi, one modern, one 70's, one zombie... etc. etc. I think it would work better to just focus on filling out a complete game-worth of models along one theme and then onto the next theme.

 

Right now I feel like we have loads of great models for making a start on lots of different games, but not enough to really flesh out a full game in any one theme.

[Guest goya]

hi steve love the assets, so much great stuff to play with, having fun with the zombies

 

i would have to agree with the idea of going back and cleaning up and standardizing instead of doing lots of different ones in different ways.

 

so perhaps come up with base male/female body with a set of standard animations and have lots of addons to them done in a standard way.

 

so for example the run animation would always be named run so that it would be easy to create a controller instead of sort of having one controller per model.

 

also the models needs a bit of cleaning up with regards to making sure all the animations are there and they are properly named and that they work and that the models and folders for each of the models have same hierarchy with standard names.

 

basically do version 2 of each package with lots of refinement, cleanup, standardization. and not focus so much on many different models.

 

another thing is perhaps the community can send steve so that he could share with everyone the cleaning up that they have done

 

so if you found an animation missing and added it or misnamed and fixed it send him a copy so he could get update out to everybody

 

in the long run everybody benefits cause it would be high value assets and new releases would not be so time consuming to integrate since they would have the fixes in.

 

cheers!

[maconbot]

With respects to the medieval content, and indeed the new Pirate - surely those all fit together. Counting all the styles in the medbase, medfemale, and the other characters, there are many many character variations. What other characters do you wish to see here.

 

I also agree with this request and we have danced around it in various other threads. As far as the above mentioned fantasy models:

...what I found was that the base kits were overwhelming at first and very difficult to find out what types of characters I could make with it. I quickly found out that really the base kits are just all the different models you see in the store put onto one skeleton. I found that if I tried to wear one leg type with a different torso type it wouldn't line up and look really bad. If there are a crazy amount of character that do line up - a sample Unity Scene along with a text file containing what matches what would help a great deal. The Unity scene would contain each major variation of the model, although really having any minor variation might be the best bet. So we can really know where we are at rather than just assuming it is workable…with out trying to use it in a practical sense like we have.

 

And really to have a great pack…all interchangeable pieces should attach not just a few. As I opted for membership I still feel I've received the value and beyond of my membership…however if a person just bought the base kits I think they would be at a loss and feel the kits are disorganized from a coding perspective.

 

I think what we are experiencing here is the typical communication gap that exists between the artist and the coders. We certainly all need to come with one voice so he is not hearing 20 different voices and ideas. My main frustration was finding as described above - a lack of standardization of the models and textures and materials. Making a unified female and male character controller seemed to almost require as much work as building one from scratch as it just wasn't setup the way it needed to be.

 

Also yes the bone weights have problems especially in the female models on smaller framework game engines like XNA.

 

I agree with above suggestions. And I think doing the following will make your library more enjoyable for you as well as for us.

 

The List:

Single mesh nude base male and female

- optional underwear/bathing suit texture

Ideally the same equipment mesh fits both male and female

If this is too hard, the same look is available for male and female with 2 meshes but same material and texture files

- optional looks can be added later that are unique to sex later, however we need a base for a start that is the same for both

Standard Naming Convention and Nomenclature

- Single Track and Broken out FBX animations all have the same name

- Textures are all the same

- Materials are all the same

 

Using this approach we can begin to come up with different materials that allow for equipping our character. If we want to allow a sheep, bear, cat, bone skinned leg armor - this should be easy to do in the UV mapped texture. You also can soon offer 30 different textures that fit a basic leg and torso setup and have great value.

 

I agree start with fantasy, and then build these suits/armor sets/equip sets into sci fi etc. Keep the same base model offer different faces and hair…and let's clean up the UV maps while we are doing this.

 

From a texture artists perspective there needs to be a lot of clean up. Draw calls need to be reduced on static objects. Materials should be shared and reshared as much as possible. I recommend creating a palette of textures that you continue to grow. I would much rather have a building that uses 5 materials but those materials are shared across my 30 building scene…and thus only have 5 draw calls versus…each building have it's own texture and now my scene has 30 draw calls, huge file size and I am limited to desktop apps only.

 

I have some examples of great texture artists that I can share with you if need to see an example.

 

I appreciate you being willing to accept feedback. I love the models and your style…just a continuous improvement thing here.

 

Cheers,

maconbot

[Guest jwall]

I agree with all of statements above. Consequence of not having that is simple, as you wrote: "but so far I haven't been able to really use them in a game".

Exactly this problematic was open in topic "Medieval male/female unification". United male/female rig, same equipment, lots if necessary cleaning, etc.

 

Anaway I think, that this problem will be solved very soon

 

 

2maconbot: Material sharing is nice, on the other side it always leads in lower visual quality/diversity. So not solution for 2013. Proper solution is to prepare scripts for automatical making of texture atlas on the fly, so you will have always 1 texture per character. It can be done under 100ms, so no problem for any kind of game. For Unity I ve even seen such plugin on the forum, but it is simple to make in almost any engine. It is even possible to make dynamically mesh only from used parts (instead of having one mesh with hundreds of submeshes and turning on/off only required). Result much faster rendering (character is one simple mesh + one atlas).

 

So again, proper solution is to solve this on technical level, not on artistic, because it will have impact on visual quality with no performance gain (technical solution will be always faster).

[maconbot]

Thanks for the tips…I found this: http://u3d.as/conten...esh-baker-2/3DL

 

However if we are combining the current UV's into a mega atlas the size of any characters merged uv diffuse texture would exceed 4096 … as the maps currently are all about 2048 and we are talking a 2048 for a leg, 2048 for a hat, etc. I'd like to see some improvement on how these UV's get made up. It's something I am working on in my own art and will always be a growing need.

 

At first I thought, well with hard drives getting bigger and download times getting faster…who really cares in the long run. But I have come to realize, outside of the reducing the draw calls….that I will eventually want height maps, along with my normal maps and specularity maps. And in coming future we may even start wanting decals available. So that can get to be an extreme amount of file space. So less is more here.

[admin]

Improovement wise, its something that cannot be done with the way the uvs are utilised. Each mesh items requires its own texture so various shader effects can be applied if that's something you wish to do. Other than that, its very problematic on export when meshes share textures/materials, as they get grouped together and combined on export, therefore they cannot be split or each selected in engine. I do agree with jwall, in that these other areas would be better looked at from a coding view. As jwall says also.... stay tuned....everyone is in for a treat in the next few days!

[Guest jwall]

Thanks for the tips…I found this: http://u3d.as/conten...esh-baker-2/3DL

 

However if we are combining the current UV's into a mega atlas the size of any characters merged uv diffuse texture would exceed 4096 … as the maps currently are all about 2048 and we are talking a 2048 for a leg, 2048 for a hat, etc. I'd like to see some improvement on how these UV's get made up. It's something I am working on in my own art and will always be a growing need.

 

At first I thought, well with hard drives getting bigger and download times getting faster…who really cares in the long run. But I have come to realize, outside of the reducing the draw calls….that I will eventually want height maps, along with my normal maps and specularity maps. And in coming future we may even start wanting decals available. So that can get to be an extreme amount of file space. So less is more here.

 

Actually you need merging of skinned meshes. There was product, which merged skinned meshes on the fly and built texture atlas. It was done by user "Diablo":

 

http://forum.unity3d.com/threads/100318-Skinning-package-for-Unity-Pro

 

Sad story is, that Unity documentation suck (one of the reason why we left Unity) and especially this "skinning" functionality is not documented at all, so it is quite pain to get it working (but in reality it is quite short code, once you find out how).

 

4096 is quite enough for character atlas, but if you want to be really next gen, you can use Amplify virtual texture plugin. It is probably best plugin for Unity ever, it will reduce your memory budget, allow ultimate altasing (whole game can be one huge atlas) and you will have solved texture streaming and almost unlimited texture size.

[Guest Chris]

Agree with everything said. A single cartoon pirate isn't much use. Try to build up model packs and consistency so people can use it in real games. Also standardise your scale settings which I find are inconsistent.

[Guest StandSureStudio]

I agree as well. I have had all the older characters for some time but have been working on system things. Now I have been trying to get the old characters to work properly and it was really difficult. I am no 3D artist...my weak point....so I was thinking it was me. I have been playing with OBM/OBF of late and looks to be in the proper direction. My genre clothing wise is Roman/Greek, Elven, decorative piece plating, Ranger Leathers....so mainly a different attire style than what is here. Some are close which will be good for reverse engineering to see if 20-32k poly characters can make use of this OBM/OBF technique.

 

20-32k poly characters are for the PC/XBox1/PS4 type gaming machines.....the current packages here are not quite there BUT so far can EASILY be used to test the Character Customization and Morphing Sliders. I have tried both Male and Female with Mecanim and worked with no problem.

 

The issue I am going to have is the naming scheme used with the code is formatted heavily (used similar model naming as Neverwinter Nights 1 and 2) as well as getting these to scale to 4 different racial base types will be a huge challenge (Human/Elf/Dwarf/Halfing...for example).

 

Bottom line....scrap the old characters and use OBM/OBF for testing humans. So far very solid. Thanks Steve