March (2014) Monthly WIP Screenshot Thread

The Desktop version is here: Kartex Desktop

Click the image to download the Android version.

3 Likes

Nice work. At some point, from jumping a bump, the kart got underneath the track :smiley:

Added custom Material support to the JFX bridge

2 Likes

I sometimes get a hankering to do more with particles than god intended. Here is a static point cloud of sprites doing bizarre things:
[video]Wavy Particles - YouTube

I’m not sure why the video stutters so badly here. I runs at several thousand frames per second when FRAPs isn’t running.

I probably won’t do anything with it. I was playing around with some ideas before trying to generalize a quad-based point sprite alternative. I just needed something to test with and then got carried away. It’s fun what you can do when you animate some parameters.

4 Likes

@pspeed for some reason I find that undulation… unsettling.

@thecyberbob said: @pspeed for some reason I find that undulation... unsettling.

Heheh… likewise. :slight_smile:

I’m playing with marching cubes style iso-surface stuff… and I decided to see what it looked like with my glass material:

Not really useable for anything but it does look cool. :slight_smile:

7 Likes
@pspeed said: I'm playing with marching cubes style iso-surface stuff... and I decided to see what it looked like with my glass material:

Not really useable for anything but it does look cool. :slight_smile:

Actually I see a usecase for space nebulars generation here with a bit adjusted terrain generation, and a fog effect if inside volume.

@Empire Phoenix said: Actually I see a usecase for space nebulars generation here with a bit adjusted terrain generation, and a fog effect if inside volume.

Yeah, it does look space-like.

The reason it’s not really useful for any of my purposes is because I have to turn off depth to make it render properly. But for some distance nebula or something it would be fine.

While I’m here, this is what it looks like with triplanar textures:

Edit to add a couple more:

8 Likes

Nice pics !!
The last image reminds me of the floating mountains of Avatar.

Couldn’t resist posting a pic after playing with post filters. :slight_smile:

3 Likes

Look quite fantastic. It’s totally random, but it would fit fantasy worlds perfectly IMO. Good job :smiley:

Wouaw Really nice.

Is it all procedural or editable?

Nice, it’s just that all those marching algos seems to create the same patters everwhere on low res voxel data. At least my brain thinks to see a pattern. But voxel based terrain is probably the future. Finally getting rid of the heightmap limitations and if you trust in papers, cheap indirect illumination

@haze said: Wouaw Really nice.

Is it all procedural or editable?

It’s procedural and editable… well the algorithm that generates the mesh doesn’t care… but I haven’t added editability to the app. That’s what you’d call a minor detail.

@zzuegg said: Nice, it's just that all those marching algos seems to create the same patters everwhere on low res voxel data. At least my brain thinks to see a pattern. But voxel based terrain is probably the future. Finally getting rid of the heightmap limitations and if you trust in papers, cheap indirect illumination

The net result is greatly influenced by the density function. A lot of people are just working from some number of fractal noise octaves (this being no different) and so the results will look kind of similar. But regardless of how you surface the density function (marching cubes or otherwise) it’s the density function that is controlling what the surface looks like. You may just more or less get more accurate surfaces from one algorithm over another.

For example, this density function (worst case scenario for testing):
[java]
float density = 1.0f;
if( (loc.x % 2) == 0 )
density *= -1;
if( (loc.y % 2) == 0 )
density *= -1;
if( (loc.z % 2) == 0 )
density *= -1;

        return density;

[/java]

Produces something quite different. :slight_smile:

Though it is interesting that the grass and light seems to be on the wrong side… I may have to look into that for this edge case.

So I tried animating the density field every frame and update the mesh. It’s really kind of inefficient because it generates a new mesh every time but it’s also hard not to. (It does free the old one, though.)

Here is a video of the animated density field:
[video]Animated Marching Cubes - YouTube

2 Likes

Rivers, i see rivers :wink:

I was not fully referring to the densitiy functions, more a scenario like this:

If you want a straight upward ramp, you basically have only possible angle/slope. But it can really be that i just think to see a pattern here because i always tought there has to be one since all mesh is based on the same vertices from the lookup table.

@zzuegg said: Rivers, i see rivers ;)

I was not fully referring to the densitiy functions, more a scenario like this:

If you want a straight upward ramp, you basically have only possible angle/slope. But it can really be that i just think to see a pattern here because i always tought there has to be one since all mesh is based on the same vertices from the lookup table.

Well, you can see in my undulating sine waves that there are lots of slopes. It just depends on where the 0 boundary is in the density field. The lookup table only controls what triangles are drawn for a given ‘cube’ but the corners of those triangles are 100% dependent on the sampled density field.

In fact, I just had to rewrap my head around all of that to do the animated demo in the video. Modifying the density field is not intuitive when you are used to block worlds. :slight_smile:

Hello monthly screenshot champs! Working on anything cool these days? Whether it’s the same thing as it was back when this thread started or something brand new, however big or small, please consider sharing it for the 2015 jMonkeyEngine showcase video:

Keep the WIPs coming!