Asset Workflow Test

[Daemoc]

This is a little project I started to test the asset workflow from Cheetah3D to Leadwerks.

^^This is the concept sketch obviously. This was really quick and dirty, but it gets the point across.

^^The 3d model did not take long either, UV's always take a while though.

^^The model with and without the bump map. I am not using a high res mesh for normal map generation. Just using hand painted bump maps.

^^More bump details.

^^Base texture test render with a little test model stand for poops and giggles.

^^More texture work.

^^I decided I wanted to cover the face so I did not have to worry about animating facial features. That will be addressed WAY down the road. If ever.

^^This is an openGL render with one point light and a 50% ambient light to mimic a game engine render. All of the lighting contrast is baked into the texture. I also modeled a little gun to test with.

I have successfully exported the test model from Cheetah to Leadwerks and everything was intact. Animations, normals as well as normal breaks. I had zero issues. There was a lot of trial and error along the way, but now that I know what works the next model I do will go much quicker. I am working on some base animations at  the moment as those are a big part of the next step. When I get some animations in game I will post actual in game pics.

 

I will also format the images better next time. There is usually an insert "thumbnail" option, but I do not see one anywhere. I hate spamming images that large on one page.

[Josh]

Very cool!

[Daemoc]

Thanks Josh!

Well, I imported the model with a few test animations in the game engine. That was a no brainer. I spent FAR too long trying to get a 3rd person camera to work though. I went through a few different tutorials and the idea was to learn the scripting language while doing something productive.

Yeah, about that. Talk about a lesson in frustration. I ended up having to DL the workshop script and modifying to suit my needs. I have the code in for all of the animations I need already, just waiting on the animations themselves.

Speaking of which, here is a little test footage of the walk cycle.

Animation Test @ YouTube

Also had a little fun with iMovie. There is no Project Nil, outside of testing that is. I am just playing around.

I am not an animator by any stretch of the imagination, but for my first walk cycle I am happy with it. It is not done though, I will be tweaking things as I go.

[carlb]

looking good and using the zone just seem right for the char lol

well done

[Daemoc]

Thanks bud.

I have most of the unarmed motion animations complete.

I have also modeled a little something to test the next stage of animations.

No texture yet, but I am not sure what kind of camera I want to use for the final project yet. FPS, TPS, 3/4 overhead...

I am leaning towards 3/4 overhead as it will be far more forgiving when it to content creation. Since this is just a test model though, I have a FPS high detail polygon count. Who knows, I still may go that route.

[Daemoc]

After getting completely side-tracked on cameras and character control over the weekend, I finally finished my jog animation. I am pretty happy with these considering I never really messed with animation until about a month ago.

The arm swing needs a little adjustment, the shoulders are a little too stiff and the head is bobbing a bit too much. I am sure there is more, but this calls for fresh eyes tomorrow. It's getting there though.

Edit: Removed Vids and uploaded to Youtube due to file size and to save on post clutter.

Jog Animation

[Daemoc]

Next on my list is to make a little "scene" to test the character in. At the moment I am using hand painted bump maps to generate detail. Again, without really knowing exactly what I want to do yet, I am keeping the resolution pretty low. More or less just experimenting. Once I start getting more pieces together I will start working on more details.

Only one floor tile and the stair frame so far. Obviously I need a few more floor tile's so it is not so repetitive. I also am going to tweak the floor grates to match the tiles a little better.

The hand rails are a bit too thick, need to slim them up a taste.?

[carlb]

that looks so nice for a test level

when you start adding dirt and rust it be great lol

 

[Daemoc]

Thanks bud. Aye, it does need some "texture" work, that's for sure. It's the little things that really make it pop.

Got a little more done today. I re-did everything I did yesterday, but kept everything cleaner.

The bumps have been toned down considerably, maybe too much? I also added a few more variations. I had to throw some geometry in there just to fill space, but those boxes are nothing I plan to use. Hand rail is thinner. I will work on textures for the next few days, then start adding details. Lights, electronics, pipes, wires, etc.

[Josh]

If you add an environment probe it will totally change the whole look.

[Daemoc]

I have weird issues with the probes and low light. Everything looks like it is made of bronze. Even the character model with no specular maps.

[Daemoc]

I had a major setback.?

I remember reading in the documentation that the standard unit in Leadwerks was a meter which is what I built all of my content around. The idea is to be able to put most of the pieces in place using grid snap. I was not even paying attention to the grid sizes in the editor, I assumed that could be changed to whatever anyway. After hours of fighting the grid snap and placing objects by hand I decided I had to rebuild this content around the grid sizes.

Lucky for me I noticed this early enough that most of this has been texture work so far. I have all of my objects resized and am back on track. I am now working on detail models.

[carlb]

looking good and well done

i love the look and style

[Daemoc]

Thanks carlb

Here is a quick vid of everything in action...
ContentUpdate4

Youtube really compresses the snot out of it, it looks a lot cleaner in person so to speak.

Everything now snaps together super easy. I have a slight scale issue with the smaller texture tiles though. For a single "square" the scale needs to be 0.1875, but the closest I can get is 0.187-8. It works up to 4 tiles, but beyond that the scaling error is noticeable. Consequently the single tile floor adjacent to the floor grates are in 1x4 sections rather than a continuous run. I might just say the heck with it and just use meshes for the smaller bits. The larger 4x4 tiles are perfect though.

These test props are super simple, but I am really seeing some potential here. I really want to do a TPS/RPG, but realistically I do not see that happening while I have a day job. This simple 3/4 content is doable though. ?

 

[carlb]

wow looks great and so smooth and love the shadow

keep it up and hope you do a great game i would love to play it lol

[Josh]

There is a resize tool in the model editor that might be useful.

the motion looks very natural, especially the jumping. It really looks like a little person is in there running around.

[Daemoc]

Thanks gents!

55 minutes ago, Josh said:

There is a resize tool in the model editor that might be useful.

I did finally dial in the model scale. It took me a while to figure out that the Leadwerks importer was actually expecting centimeters and automatically converting to meters. Once I figured that out I was good to go.

The scale issue I am having applies to the UV scale of textures applied to CSG's in the editor. One texture tile = 96cm in order for my content to line up properly. The 4x4 tiles match perfectly at 0.75. The 1x1 tiles equals out to 0.1875, but the editor rounds to the nearest thousandth. Again, staying under 4 tiles the scale error is pretty much a non issue. I also can easily fix this by just using meshes for single tile objects rather than CSG's.

 

[Josh]

I am interested to hear more about this issue. I do not think I understand what you mean. The UV mapping on CSGs should be perfect.

[Daemoc]

It is just the scale that I happened to settle on is not optimal for use with the engine. If I use any power of 2 square on a 192cm square CSG, the UV scale = 0.75 and it lines up perfectly. If you cut that 192cm into quads and apply a power of 2 square texture, the scale needs to be 0.1875 to line up properly. The only issue with that is the engine rounds to the nearest thousandth. Either 0.187 or 0.188. It is a very subtle error, but after 4 tiles that little scale error adds up and you can see things not aligning properly.

It’s not a big deal. This is the very reason I am testing like this. Another very quick fix that just came to mind is to stitch 4 1x1 tiles into a 4x4 texture.

[Josh]

Why would you use an odd number like 192? Why not use a pow2 number like 256 or 128? if you do this then it becomes very easy to line up geometry with textures details, like cutting a window out in the right place inside the outline of a window frame drawn on the wall.

[Daemoc]

I certainly did not go out of my way to come up with that number. Not in the way that you are thinking anyway. The idea was to come up with a standard grid size for my content. As you can see, all of the content there has a grid like aesthetic. There is a reason for that. Originally this grid was built around a meter. One grid = one meter. That obviously caused issues with the editor power of 2 grid once I started importing content over.

So the compromise I came up with was 96cm which was fairly close to a meter and worked with the grid snap in the editor. This does in fact work except for the power of 2 to 96x96 UV tile scale issue. And in all honesty, that does not bother me as it is an easy work around.

I do see other issues that might creep in later with the grid coordinates though. Objects centered on a game grid will not center on an editor grid. This would not matter for the little real time adventure type thing I have going on right now, but it might for a more turn based “tactical” design. I don’t know. I don’t know enough about the programming end of it to say, but it looks like that could be an issue from a common sense stand point.

The issue with going full blown power of 2 for the grid size is 64 is too small and 128 is too large for a game grid in my opinion. From a strictly aesthetic standpoint anyway. That said, I may have to make another compromise here as well.

In fact, as I have been typing this I have decided to try an make a 64cm game grid work.

[Josh]

It looks like in the above shot each floor tile could be 1.28 meters wide (the stairs are 2.56 meters wide). That means two tiles high would be enough for the character to walk under. It looks very close to that right now.

But if you are more particular about the size then I can understand how that would be difficult to settle.

[Daemoc]

Maybe I am approaching this the wrong way. I guess at the end of the day the actual scale, or unit of scale rather, does not really matter to me. A unit of measure is completely arbitrary in the real world. Relative might be a better word. I guess for the digital world a power of 2 is a constant.

Maybe the thing to do here is use power of 2 dimensions eg: 128x128 grid and just treat my content as a 128 = meter. I can do that on export and it will be completely transparent.

Either way, I will figure something out. This entire thread is just myself thinking out loud anyway.

[Daemoc]

I have put the content on pause for a bit while I am deciding on grid sizes. In the interim we have worked on some other scripting ideas. It took two days, but we now have a basic point and click character control as well as basic AI...

Content Update 5

We also added camera zoom and rotation into the mix.

Neither of us are coders and are pretty much starting from zero so it is taking a while to get anything done. Although, slow progress is better than no progress I guess. We still have some major issues to tackle. Most important being a "grid" snap system for the reticule and therefore player movement. The pick collision is doing strange things, need to get that figured out as well.

There seems to be a hard coded deceleration in the point to point motion that I just can not get rid of. Acceleration works as advertised, but the slow deceleration is always there. I am faking a quicker stop by stopping the character short of actually reaching the destination point at the moment. It looks a little better this way, but really it is just a hack fix.

[carlb]

that looks good and well done

am sure you sort the deceleration