WIP models / concept art / creature system

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
Re: models / concept art

Great stuff - I'm limited to view that from work, but caught a really nice vibe off the model. I can somehow just picture those little guys darting around a DF-like base :)

And at 300 polys it'll take a few of them to rival the terrain poly count...

Very curious to see how we'd move forward with something like that on implementation and motion animation. Maybe we'll have actual creatures in-game soon!

Yay! Thanks! :D
 

begla

Project Founder and Lead Developer
Contributor
Architecture
Logistics
Re: models / concept art

Looks... Awesome! Love the style. :D
 

begla

Project Founder and Lead Developer
Contributor
Architecture
Logistics
Re: models / concept art

Immortius said:
Very nice. :)
Immortius – you up for writing a Blender general-purpose mesh importer? We could split up the part regarding the (skeletal) animation system. :geek:
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
Re: models / concept art

Behold! Photographic evidence of my theory that the snazzy dwarf model reminds me of Jake the Dog from Adventure Time:



Which is also a terrifying reminder of why I don't do art. Even yellowed over the eyes and failed to black the nose/eye rim. But I'm at least trying to learn Gimp rather than use Paint! :D

I also hereby name the dwarf model "Jake" :D

In more serious news we seem to be nearing critical mass on the art angle, Blenderers are coming out the woodwork, which is full of win. So we need to get started on some stuff! This was the exact goal for the next milestone (milestone 4 - Interaction)

* Blender -> Terasology like indicated by begla above - https://github.com/MovingBlocks/Terasology/issues/210
* Render refactoring - needs a little cleanup & prep :) https://github.com/MovingBlocks/Terasology/issues/211
* Rough creature system foundation - ES relation? ES branch close to live? https://github.com/MovingBlocks/Terasology/issues/212
* Models! We have Jake the Dwarf, and it would be wonderful to have a few similar ones. Glasz is tinkering here (need GitHub user!) :) https://github.com/MovingBlocks/Terasology/issues/213
* Creature definition files (existing item) - https://github.com/MovingBlocks/Terasology/issues/128

I'm sure there's more, but I don't want to go nuts slicing and dicing until more discussion has had some time!

One last thing: I'm thinking about us coming up with a race that's unique to Terasology, which is somewhat of a challenge, as just about everything has been done already. To really make it "fit" with the idea of a voxel world I figure it has to have a special relationship with blocks - so what about some sort of rocky/minerally/Silicoid'y creature, that's possibly able to fold itself somewhat into the shape of a block (to hide from evil, but stupid, creatures, or as a parlor trick, or to go to sleep), maybe able to swim in lava but sink in water (heavy), maybe able to squeeze into a 1-block tunnel, or even walk in semi-block mode? Crazy people on IRC brought up the Mimic from Dark Souls: http://www.youtube.com/watch?v=NJbZ2thxCaI (less chest, more block!)
 

Attachments

Immortius

Lead Software Architect
Contributor
Architecture
GUI
Immortius – you up for writing a Blender general-purpose mesh importer? We could split up the part regarding the (skeletal) animation system.
Sure. I guess for non-animated stuff we should just consume one of the standard mesh formats like .obj rather than rolling our own - unless there is any specific features we need? Can't really think of any. Maybe can use a generic format for animated meshes too, would need to investigate.

The main bit would be creating a Mesh asset class and integrating it into rendering.

Creature definition files (existing item)
In the Entity System world, creature definition files are subsumed into general purpose Entity Prefab file - you just define what components should make up a creature, with what settings.

Rough creature system foundation - ES relation? ES branch close to live?
The way creatures work at the moment in the ES, there a bunch of generic components all creatures would have (Location, CharacterMovement, Mesh, Health, maybe Inventory, probably some sort of Creature component), plus one or more AI related components. Then you add a system that updates the movement component's driver fields (desired movement direction, jump request), uses items in the inventory, and reacts to events (collision, damage, death) for entities with a specific AI component (or components).

The player is the same of course, except instead of AI components they have Player components.

It is also tempting to split things up further, with the AI/player being a separate entity that controls the character entity.

As far as how far the ES branch is from being live... Basically the branch lost a lot of functionality when I started converting over players, so I guess getting most of that functionality back would be nice. At the moment the main bits which are missing are:
* Tools
* 1st person rendering stuff like held item
* Some camera stuff like walk bob
* Torches
* Some Block update listening (not sure why it was up to each tool to notify of block changes rather than the world provider?)

Probably some other stuff too, lots of code touched on the player.
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
In the Entity System world, creature definition files are subsumed into general purpose Entity Prefab file - you just define what components should make up a creature, with what settings.
Are they in Java or config files? I figure like blocks we should strive to eventually make it possible for players to define their own extremely easily by simply providing a plain text config file (so far usually a Groovy super props file)
 

Immortius

Lead Software Architect
Contributor
Architecture
GUI
t3hk0d3 has them loadable from JSON files at the moment, which has a similar syntax to a Groovy config file.
 

A'nW

Member
Contributor
Art
Hi there everyone! I took inspiration from glasz's OP and created a pretty final dwarf character. I liked glasz's start but felt that the dwarf should be a bit more 'chunky' and beefy looking. Then I decided to texture it. :) I think it would fit the games world pretty well. All the texturing is completely from scratch and I think that it looks pretty unique! See what you guys think.
 

Attachments

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
Cool! Very nice looking dwarf, especially with the texturing also done. I wonder how Jake the Dwarf would look with full texturing :)

Still curious how we would go beyond that with animation and modular "model anatomy" if that makes sense. But we probably need to evolve the game itself a little more for that.
 

metouto

Active Member
Contributor
Art
:shock: ...... :eek: ........ :D ...... Jake, the Dwarf ..... you are the coolest dwarf @ Terasology !!!! Good job glasz & A'nW ;)
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
They're actually two different ones, glasz with the initial lower poly model without textures (the one I've named Jake), and A'nW's higher poly textured dwarf - I'm curious what's the poly stuff on that dwarf, A'nW? And really curious to see the lower poly model textured :)

Oh, now I suddenly see that there are more stats if you scroll down on that dwarf.
 

A'nW

Member
Contributor
Art
Thanks for the comments guys. Yeah, I wonder if glasz is going to texture Jake the dwarf.
@Metouto Thanks for the comment, glad you like it! Though "Jake" is actually the name given to glasz's model.

@Cervator I'm glad you saw that you could scroll down on my dwarf picture! It does look like you don't need to. :) Hmmm... taking a look at the wire-frame of my model again I can see there are a couple of unneeded edge-loops I missed that would reduce the poly count. I could also get rid of several quad faces to take the poly count even lower if needed. It will probably be a while before a character like the dwarf makes it into the game though anyway...
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
Okay, bigger post time! :geek:

First a giant disclaimer: I don't know a thing about art or artists

I'm realizing I need to highlight that more as otherwise I'll totally miss stuff and end up making somebody feel bad or something :)

I really like both dwarves posted so far, and especially would like to see how the first one would look textured.

Having two of them does bring up a tricky point tho: How do we collaborate and make choices in the artist space?

This is where my coder nature defeats me. It is easy to collaborate with code, since it is just an implementation of a concept that can be tweaked till you don't even recognize the code, yet it remains the same concept. Immortius did major refactoring to the block manifestation system I started, but I'm thrilled to see that as now it just works even better with more solid code (btw, I suck at code too). That's not so easy with art - re-use is far trickier. Add to that the international nature of the project and the potential for language issues / misunderstandings.

So I'm figuring (correct me if I'm wrong) that we need to arrive at a general poly count / level of detail range and other guidelines like that to have the models be fairly similar. And maybe we can use dwarves as the proof of concept example to see how dwarves at different levels of detail would compare with each other? In theory this might be doable even without a more concrete game concept, since the models need to fit nicely with the world, regardless of the theme or flavor to the world.

With guidelines in place then we could actually divvy out specific art issues so at that point we only have one person working per race or other grouping, with feedback from others or outright requests for alternative takes by the author?

At the same time, can we actually use multiple similar models for the same thing? Could we have mountain dwarves and plains dwarves, looking differently (like the two examples in this thread) in the same world, without clashing in style? If that would be a concern, I'm not an expert :(

Begla indicated on IRC that the poly count for either model here should be well below any point where the game would have any sort of trouble rending a pile of them running around. So that's great. I'm figuring it isn't a big issue either to have multiple variety texture sets? Almost as in DF where different professions = different looks

(Also, so I can put it somewhere: plug for maybe having a race inspired by http://en.wikipedia.org/wiki/Pequeninos?)
 

A'nW

Member
Contributor
Art
Big response time! ;)
Okay, bigger post time!

First a giant disclaimer: I don't know a thing about art or artists

I'm realizing I need to highlight that more as otherwise I'll totally miss stuff and end up making somebody feel bad or something
Don't worry to much about that. I, at least, understand, and think you brought up some good points here.

I really like both dwarves posted so far, and especially would like to see how the first one would look textured.

Having two of them does bring up a tricky point tho: How do we collaborate and make choices in the artist space?
[Bold mine]

Boy do I hear you. That may be about to become a large issue. We really need to figure this out before we get to deep into the art side of things.

This is where my coder nature defeats me. It is easy to collaborate with code, since it is just an implementation of a concept that can be tweaked till you don't even recognize the code, yet it remains the same concept. Immortius did major refactoring to the block manifestation system I started, but I'm thrilled to see that as now it just works even better with more solid code (btw, I suck at code too). That's not so easy with art - re-use is far trickier. Add to that the international nature of the project and the potential for language issues / misunderstandings.
Quite right. Art re-use seems to me to be much more tricky.

So I'm figuring (correct me if I'm wrong) that we need to arrive at a general poly count / level of detail range and other guidelines like that to have the models be fairly similar. And maybe we can use dwarves as the proof of concept example to see how dwarves at different levels of detail would compare with each other? In theory this might be doable even without a more concrete game concept, since the models need to fit nicely with the world, regardless of the theme or flavor to the world.
This is were things get complicated. Simply setting a general poly count (or other rather simplistic guides like that) is not going to cut it, yet we definitely need some sort of bigger guiding standard. I don't have much experience in this area either unfortunately (as I generally work by myself), but do have some ideas. First, there should be a "design sheet", some place where all the modeling,texturing, rigging, animation, etc. guidelines/tips/recommendations are. I already did this personally to some extent when I made my dwarf. I set myself a 12x128 texture, I made the model perfectly pixel proportional , etc. Stuff like that would go a long way in "unifying" the look, at least, of stuff. That would not really be enough though probably, at least not if we start getting a lot more than just 3-4 of us artists. :) The reason for this is that you could still get stuff that fulfills all technical guidelines and still looks "wrong". The only way to remedy this as far as I know is to have some sort of "Art director" or "Art lead" that could kind of "moderate" all the incoming art assets. I don't know how this would exactly work though being an open-source project.

With guidelines in place then we could actually divvy out specific art issues so at that point we only have one person working per race or other grouping, with feedback from others or outright requests for alternative takes by the author?
Yes, you could and that would be a good idea to break up the load. Again though, you are going to get really varying results even if everything is up to the guidelines standards.

At the same time, can we actually use multiple similar models for the same thing? Could we have mountain dwarves and plains dwarves, looking differently (like the two examples in this thread) in the same world, without clashing in style? If that would be a concern, I'm not an expert
Similar models for the same thing? Yes, if they look good together. Having a few different models for each character would be really nice actually. Here, though, I have to point out that I would not consider glasz's and my models to be "similar" in this regard as they both have different design guidelines (not saying that either is better). For instance, his has floating arms, hands, and legless feet. If you saw both of these in a game, at the very least, it would be rather odd looking (at least to me). I think that design differences like this are definitely a concern if we want a game that looks like it has a nice "cohesive world".

Begla indicated on IRC that the poly count for either model here should be well below any point where the game would have any sort of trouble rending a pile of them running around. So that's great. I'm figuring it isn't a big issue either to have multiple variety texture sets? Almost as in DF where different professions = different looks
Yeah, I would have guessed that their poly count wouldn't be a huge concern. As for different texture sets/skins, that is a great way to introduce some nice, simple variety.

(Also, so I can put it somewhere: plug for maybe having a race inspired by http://en.wikipedia.org/wiki/Pequeninos?)
Those interesting I guess...

All in all, I think you brought up some very relevant points. Stuff we need to figure out. I also have to submit a bit of a disclaimer. I am very much an art perfectionist so "world cohesiveness" is a rather important point for me. I really have to be carefull what I say because I don't want to make anybody feel to bad. :D Projects like this tend to attract rather a lot of inexperienced/new artists though. If someone doesn't have firm hand and everything gets included...
 

metouto

Active Member
Contributor
Art
A'nW said:
@Metouto Thanks for the comment, glad you like it! Though "Jake" is actually the name given to glasz's model.....
:oops: :oops: :oops: I thought that was Jack's other bother Jack the Dwarf 2 :oops: :oops: :oops:
 

Cervator

Org Co-Founder & Project Lead
Contributor
Design
Logistics
SpecOps
Poked about this topic a little on IRC again, but just to get something in the thread - yeah, I figured it'd be too easy to just get at a poly count etc, alas :)

AW - would you mind posting your full "design sheet" to see if others like it / can agree with it / tweak it?

Beyond that, I suppose we need a "feel" for the art - which yeah, again probably isn't as simple as "thin", "thick", "rangly" etc since all those vary by the model... would more examples help us at all?

On the two dwarves styling differently then yeah, I hadn't thought much about the differing details like slightly floating arms and such. I sort of like how the first one ("Jake") has sort of a distinct style to it, yet the second one might be more standard'ish yet more versatile?
 

A'nW

Member
Contributor
Art
Immortius said:
Monkey head demon slimes.

Have some basic obj mesh importing working.
Awesome! The obj mesh importing that is. :D
 
Top