(September 2016) Monthly WIP screenshot thread

No I’m still on 3.0 using JBullet which is about twice slower at the moment. I just can’t get myself to trust 3.1 yet when I see a “random thing not working [3.1]” on the top of the forums every other day. My code is buggy enough without a questionable base.

Fast when colliding with what? Primitives? Possibly. Other hulls? No way, not from what I’ve tested. The reason why I switched to using a bajillion primitives is because the hulls were so dissapointing.

And that was using one hull per ship, can’t imagine how bad would it be if I split it so that everything would be convex without gaps.

Hull vs Hull

Yeah, jbullet is a lot slower that native, and it’s also a porting of an older version.
I think you can use the build i linked with 3.0 as well, all depends on if the java interface has changed.

I don’t know about 3.1, but i use 3.2 (ie master branch) and it works fine


Just try using the updated build the native build linked above by Riccardo. Bullet is a great engine but the bridge between it and jme is actually unstable and needs some precautions to be used at 100% of its potential.

Also,me trying to develop Unreal Tournament 5 with more missing textures.


It sounds a bit too much for me - will go and test it later. I used convex hulls for super early prototype of my game before switching to UE4 and performance was acceptable. Don’t remember what Bullet setup I had back then, though.

Just to note, for best performance single convex hull (not “convex triangle mesh”, have seen such collision type somewhere, maybe not even in bullet) should have from 4 to 32 points in it, the less the better. so two convex hulls 32 points each are better than single hull with 64 points (don’t necessarily apply to very small hulls, like 8 hulls with 4 points each vs single hull with 32 points).

But if boxes are good for you - go with boxes. It’s just me who is sick with boxes already :chimpanzee_lobotized:

Okay I’ll take a look :slight_smile: About the only thing I’d need 3.2 for is the physics so that would hopefully save a lot of adapting on my part.

Hmm, looking at some old screenshots I counted about 30 vertices and possibly more. Still, splitting into more shapes sounds kind of counter intuitive, unless there’s some distance check optimization done before. In that case it’d make a lot of sense.

zooming right out shows how I ballsed up the shockwave scaling, but at least it shows the effect before it gets muted



the shockwave looks really awesome! how did you accomplish that?

Cool tank’s don’t look at explosion x)

Last time I tried using that (it was like a year ago though) it kind of rendered over everything like the last post mentions. I kind of didn’t know anything about shaders then so it’s possible that I messed something up. Does it work as it should for you in all cases?

Some photos of today’s games conference I exhibited at. Not as fancy as PAX but in my home town so the flight is 12 hours shorter :stuck_out_tongue: .

I made sure to tell everyone interested about jme and surprisingly some actually knew it exists :smile:

Also, thanks to @grizeldi for showcasing the app there too.


yeah, but i change it as required, overall it’s a pretty simple shader.


Very Cool
Every thing is eye catching in the scene
Really how far you want go with JME ??? :stuck_out_tongue:

More testing of animations and hit box line up.



I see potential :smile:

1 Like

Isn’t some models from WoW?

No , https://shop.bitgem3d.com/

The skeleton and imp where freebie models. I bought several of the environment packs but I’m still looking for some other character models to go along with them.

Is it compatible with the AbstractHeightMap terrain? And if yes would you mind sharing it? :stuck_out_tongue: (talking about the shader base grass)

You’re not fooling anyone that is Unreal or Unity :slight_smile:

1 Like