(May 2018) Monthly WIP & Screenshot thread

This is not properly a project, but yesterday I made a talk about Shader Nodes at the Google I/O Extended 18 event organized by GDG Catania :slight_smile: Here’s a photo with all staffers, speakers and some partecipants (I’m in the background).


Here’s the presentation material: Speech shader nodes – Google Drive
And here’s the project I used as example: GitHub - aegroto/JMEShaderNodesExample: jME shader nodes example, made for Google I/O Extended Catania 2018


I think that adding more colors makes an image less boring to an eye.


Not sure if this sort of creative falls into what’s topic is up to (strictly speaking it’s not programming, nor especially Java), but I’ll dare to post, since significant part of it was indeed influenced by you guys, and JME in general. Sorry for being off for so long, can’t make promises right now (RL is tough), but hope to be back with you one day.

p.s. one more small excuse is that AFAIR Paul has a guitar somewhere in his place too :stuck_out_tongue:
p.p.s. there’s a DarkMonkey theme pictured, not to be completely off topic here on hub

Sounds like a very very soft version of Samael :wink:

Point taken) Still, Samael is a band, not a single person, nor living in a heavy-pressed dictator’s country. But point taken, thanks :slight_smile:

p.s. u’ll barely believe me, but I was in a band that warmed up Samael when they visited us btw.

1 Like

I have a project in mind, that may need a lot of GUI. When i was bored during my job I wrote a concept (last 3-4 years) now I start to touch code. My aim is to not loose brain during my job and finish the project within the next 5 - 10 years. Im not an IT guy, never learned or studied it - so im not sure if it is spagetthi code in the end :slight_smile:

After checking and trying with lemur library I am currently enhancing it.
I try to make the basic example scrollable (the demo let you have lines out of the textbox in Y direction) and I will also add things, I believe I may need. Of course this is otional, so the old functionality is not gone.

What is "working (98 %) right now is the “scrolling” in X and Y direction.

Once I finish and I believe it is working I will open a separate topic in the lemur categorie.


I’ve been super busy with spoxel the last few months since I have had a huge influx of beta testers (3k or so?). I’ve received a ton of great feedback and there have been lots of changes. Here are some of the dungeon changes i’ve been working on.


these purple/pink “explosions” looks very cool

1 Like

Picked up where I left off on those metaballs I was working on a while ago:

Now rendered to texture, with each color group rendered separately and up to 1020 points each. Now it’s just a matter of adding it into the game itself…

Edit: Weeeelll, shit.

That didn’t turn out half as good as I had imagined. Seems like the great band of prosciutto is metaballing over internal systems that aren’t set to it. I guess I either have to reduce the connection thickness or come up with some clever way of forcing metaballs to avoid areas they aren’t supposed to make connections through.

Maybe I could just change the function that’s usually additive to subtract for the avoided points?


Today I have been playing around with modeling and getting some sort of proper animating and modeling pipeline going.
I modeled the scene in blender3D with the Sprytile plugin.
The model was free and I used Mixamo to generate animations and combined them in blender.

Here is a result of what I am busy with:


Latest SiO2 bullet ES integration updates.

Hooked up contact entity publishing and a debug app state to view it. Also hooked up collisions to the control drivers so they can make decisions based on contact information.

Here is a video of one of my wanderer guys with full debug view on… randomly wandering around and avoiding bumping into walls. Watching these guys run around is very Zen.

Edit: I think the only thing left to add is ghost types… ie: the ability for ghosts to not collide with everything. (For the wanderers, we want the perception ghost to collide with everything but a ghost you use for sensors probably should only collide with non-static stuff… for example.) Then I have to port my prototype over to use this new physics code to be sure I’m not missing something.


I know that music :wink:

1 Like

It should’ve switched to “you spin me right round” when it got stuck in that corner.

Heheh. I’m impressed with how often they find their way back out again.

I watched this guy wander around various configurations for like 15 minutes while I was trying to figure out why fraps wasn’t working. He finally got stuck once the first time I got fraps working (of course) and it was a case where he was wedged directly into a corner where left/right steering away from the obstacles was just taking him right back in.

Obstacle avoidance in this demo is 100% prioritized over the wander influence and since the block was nudged up against a bunch of others, it wouldn’t slide. He was going to be there forever.

Well I would assume he’d get consumed by the sun in a few billion years.

Anyhow this is what my map ended up looking as:

I removed the prosciutto fat borders, scaled connectivity coefficient with distance from the galaxy center and fed everything into a single pass blur shader. Looks pretty good now imho, the only question now is what to do with those randomly generating nebulas.


I added the ability to configure collision masks for ghost objects. So you can have ghosts that only collide with mobs, only static, static + mobs, only ghosts, no ghosts, etc…

In the process of playing around, I turned up the mob count. As I sat there watching the red-peg-cyclops mobs wandering around it was very mesmerizing. So I thought I would share.

Enjoy your Zen experience… lol:


I’m just sort of wondering what would happen if you added a second larger “view” capsule collider (say twice or 3 times as large) with a lower influence of say 10% or 5%. Would they make less abrupt turns?

I suppose the next step after that would be having a sort of angle of a like a dozen rays going out and scaling the rotation force by distance at which an obstacle is detected.

The capsule size is only partially to blame for the abrupt turning.

There is a lot of tuning in things like this. How fast do you scale the turning force in relation to the obstacle and so on. Ideally, I’d sit and fire these guys at a wall at various angles until the turning influence in all cases was just enough to have them follow the wall.

…but this is a demo and I don’t have time for that (especially considering random wandering alone sends them right back into their obstacle half the time). I’m 99% sure it’s possible with just what’s there, though. The only thing the capsule size would affect in that case is how far from the wall they would walk. (And of course how quickly they have to react when hitting it face on.)

There is nothing in the bullet integration that prevents adding multiple capsules so someone could always try it. There is nothing in the framework preventing it, anyway.

I will spend lots of time tweaking steering in the real game, though. With a real environment with actual walk-goals, etc. it will be easier to see what will/won’t look good.


I tried Mixamo with mine and their default model. After I import the fbx file in blender I converted the blend in j3o file but animation is not playing same as the one in blend file. Can you tell me what magic you did to run the mixamo animations in jme.

edit: probably I need root bone?

Hi there

  1. Well First of I did not export the model as FBX from Mixamo, I used Collada(.dae).
  2. After that I imported the .Dae file in blender.
  3. I then had to fix the materials in blender.
  4. Then I saved the model in my jME project assets.
  5. Converted the model to .j3o.
  6. An started using it in code.

That is the short of it.