DeleD Community Edition

[Basiror]

Hi,
after playing around with deled the whole afternoon there are some featues that might be useful for future editing

a) Implementation of a paint tool to draw blend weights for texture layers on "Grid" based surfaces
One could extend the grid surfaces to Bezier, Bspline ... surfaces easily
This is particularily useful if you want to use the Grid objects as terrain mesh.
I know there were debates about heightmap terrain vs mesh based terrain, coming from COD editing I prefer mesh based terrain since it s a lot more flexible and powerful and its yet easy to implement based on the current feature base.
One currently can only blend 3 layers right?

b) I don t have the PRO version yet so I don t know if its already supported, but editing grid meshes currently requires you to switch between 3d view and 2d view

would be nice if one could do all this in the 3d view itself.

Select an object, draw the 3 axis and check for mouse<-->axis intersection to handle input (Drag & Drop)

This could also be implemented for any other basic modification (translate, rotate, scale)

[Basiror]

another thing I realized, when you create a brush the brushes are always created at the ground with default (height,width,depth)

would be great if you would store the bounding box of the last selection and use this as a reference where to place the new brush and how large it shall be in the 3rd direction

ala radiant editors

another thing, it seems like your using a 16 depth buffer or maybe you choose 32 bit, in the latter case the choosepixelformat function tends to fall back to 16 instead of 24 bit, just use a 24 bit depth buffer

[elementrix]

[Paul-Jan]

Good suggestions, thanks!

Two questions:

[Basiror]

CODRadiant does have a painting tool that allows you to paint color and alpha values onto terrain patches, however COD does not feature texture layers so you had to place several patches onto of each other and alpha blend them together


Combining this with texture layers would be awesome, btw would it be possible to add a 4th layer, rgb values are converted to rgba anyways, so why not make use of the 4th component to blend a 4th layer


As for the implementation, when you paint blendweights you should initialize the weights like this
(1,0,0,0), now whenever you paint you increase or decrease one weight by deltaX, you should increase/decrease the other weights by deltaX/3
and take care of the case when one weight becomes zero
r=r-deltaX/3 < 0
deltaY = |r|/2
apply deltaY to the other 2 components left to adjust
r=0

so the weights are convex(e.g.: ||weights|| = 1)

As for the depth buffer issue, I am working with 24 bit depth buffers since geforce2, I noticed the issue with 16bit while placing several brushes next to each other. With 24 bit you don t see that zfighting anymore

I can send you some pseudo code for blendweights painting, did implement this for gtkradiant earlier this year

[Paul-Jan]

Thanks for the info!

Quick note: there is a 4th texture layer, it is reserved for the lightmap so you cannot set it explicitly through the interface.