(November 2018) Monthly WIP & Screenshot thread


This is very nice looking… But maybe tell us what we’re looking at? I’m ok with even a title “Yellow dude contemplating his oneness with nature during an existential crisis”.

But yes. Looks very nice :slight_smile:


its a WIP screenshot =P

I can tell you what it’s not :

  • it’s not a distracting side project
  • it’s not a basic voxel world
  • it will not use an Isometric ish view
  • the yellow people are not using mocap data and definitely don’t fluidly traverse the world
  • doesn’t use PBR and I haven’t found a cool new tool to create pbr textures from an image
  • this is not very similar to the early days of Makers Tale

ooo! Can’t wait to hear/see more then! :smiley:

Looks really nice!

Make it sway in the wind and you end up with something like this which is my favorite grass :wink:

it already does :wink: but needs tweaking, I haven’t really played with this shader for months, just slapped it in.

also it LOD’s, which can be seen in this old post.


Attempting to create some kind of atmosphere in a completely generated world. The amount of time I’ve spent trying to perfect world generation is ridiculous. Some of the views in the video are quite pleasing to the eye, though. I guess it can only get better after I put some trees and more of my weather system into it. It’s a nice moody day, though. Can’t complain so far… Framerate is capped by fraps.


Tessellation was one of the last pieces of the puzzle for me :

jamie model, with a simple fur like 2k height map (used as diffuse map), tessellated to buggery… 10 year old GTX 470 pushing 18m triangles @ 60fps 4xMSAA… pretty nice.

for some perspective, the same model subdivided 3 times in blender with no tessellation = 2m triangles @ 60fps 4xMSAA :

TL;DR; : tessellation is yielding a 900% performance boost to vert count ?

Not a heap to see really, its a bit of a mess, this is a proof of concept can I tessellate a 3d model using a heightmap for displacement, and what sort of performance does it yeild?

zoomed right into the neck with wireframe on :

Interesting, very interesting. Now I need to find a real use case.


LOL …fail…


Ok, confirmation this is black magic…

A basic Quad + Vector Displacement Map (an image) + Tessellation + [BLACK BOX OF UNKNOWN MAGIC] =

my head is exploding, much like the fail image above, with ideas on how to use and abuse this crazy simple yet powerful concept.


I guess it’s good for adding detail. Nvidia talks of an n-patch algorithm to make meshes more rounded. Would be nice to see it in action in Jme.

That was a bit more complicated. What’s next ?


Nice. Throw it at Jaime and see what happens? Make a Minecraft esque model and turn it into one of your yellow zen dude(tte)s? Do they even have a gender? People need to know. These are the real questions.

smoothes him out

could do, the model would have to be rounder than a typical minecraft model since the smoothing wont make a cube a sphere.

That project is shelved due to lack of originality. They would have reproduced by fission.

Some marathon sessions of Rimworld last weekend got me in the mood to work on NPC tasking systems. This was long on the Mythruna to-do list and I had jokingly told my son I was “doing research” when playing Rimworld but that wasn’t far off the mark.

Friday night I threw together a simple 2D prototyping base app with some borrowed graphics from kenney.nl and then I was off.

Doing the prototype on an ES has thrown a few wrinkles in the mix that may have taken a little more time away but I think in the end what I have will be more useful for it.

Anyway, here is a video of it at the “stupid” stage as of a few minutes ago:

…best viewed embiggened at 1080p else it will be hard to see what’s happening.

From the youtube description:
In the vein of “sometimes sucking at something is the first step to being sort of good at something”, we see a dirt-stupid AI scheduler prototype. It needed to get to the stupid stage before I could start making it smarter… ie: I needed to have actual use-cases of it being stupid. Fortunately, the scenario is a lot less complicated than I thought it would have to be. Maybe 25% of my original plan for a scenario it would take to get ‘meaty’.

Mobs will haul any available wood to the stockpile and chop any trees when there is no more wood to haul. Each mob can only carry five units but trees may produce up to three logs of varying units. It’s this last bit that makes it interesting from a scheduling perspective because once one NPC has ear-marked picking up a log it’s not available for anyone else… even if that NPC will not take all of it. It leads to a nice bit of complexity that will let me play with ‘real’ job scheduling earlier than I thought.

So next step is to coordinate tasking so that NPCs don’t do stupid things like walk all the way across the map to pick up a log next to the person that just chopped it. Then add the next levels of the prototype scenario and iterate.

This is sort of the minimal environment to testing such things but the application of the ideas will be directly usable in Mythruna’s town management.

“So like Dwarf Fortress/RimWorld/Etc?” Yes, sort of like that. We’ll see how it goes.

End goal is to build a self managing town that repairs itself when disasters/players strike… and is predictable enough to simulate in broad steps when the player is not there. Reach for the stars, eh?

Music by me.
Placeholder graphics from https://kenney.nl/


I created a use case for tessellation, which took all day learning WTF is going on in zbrush, the turning a hi res obj into a low res one for re-projection… I have a very low poly mesh that tesselates to a detailed mesh. (link to gallery with details rather than flood this thread more). Mucking around in zbrush with no idea what I’m doing sucks, I just need a good “multires” model =/

Hi, it was a long time since my last post…

I was busy, had lot of things to do - In addition to writing Skullstone, I had to arrange matters related to my day job - I gave up boring IT and now I work in gamedev.

And here is some (I hope) interesting thing, a fragment of beta gameplay. (the walls are in beta too :slight_smile: )


Nice :slight_smile: and also “witam rodaka” :slight_smile:

1 Like

Not nearly as beautiful but just to participate in this thread every few months:

What you see is a badly cropped (blame open shot for that) simple Weapon HUD with “Reload Animation”.
Background is that I need to “polish” my game in it’s current state so that one has a bit of feedback, else it feels like a plain prototype


working on a simple diorama to showcase jme graphics, very early on, I’m reworking a screen space distortion filter, added dispersion, to make a better glass shader, modifiying PBRLighting to better support transparent materials.

No effects yet, low poly. Models support tesselation. Rendering an additional light probe from inside the tank will make a huge difference down the line.