Irrlicht Engine

The long waiting is finally over - Irrlicht 1.8 is released!

You can get it here: http://irrlicht.sourceforge.net/downloads/

A short list of feature can be found on the homepage, a way longer list can be found here: http://irrlicht.sourceforge.net/changes.txt

Thanks a lot to everyone involved and to our fantastic community!

Cool! Downloading now! Thanks!

Holly Molly! Awesome! a release! great!

That's a very good new , but the shadows of the demo and specialFX example don't work very well...
http://jl-chrono.olympe.in/sc.jpg

Awesome, the wait is finally over
Thanks guys

Woo! Awesome!

JLouisB wrote:That's a very good new , but the shadows of the demo and specialFX example don't work very well...
http://jl-chrono.olympe.in/sc.jpg

It's cause by bugged model (sydney.md2) which we use in the demo. Code works properly Similar problems are related to a model (dwarf.exe) used in 08.SpecialFX.

Hurray!

Very goooooooood

Ok nadro

Great news, many thanks!

What's the problem with the dwarf? It has a few minor holes, yes, but has no streaks!?

In dwarf.x case I said about holes (some peoples may think that shadow algoritm is broken, but a problem is related to a model, eg. JLouisB probably say about these issues).

Hi.

For the dwarf the shapes are enclosed properly (properly closed volumes), but the head is not connected to the torso with polygons, that could project holes when projecting the shadow.

For the Sydney model, the modeler has badly enclosed volumes (arms are capped and he modeled the hands over it). So instead of 1 properly enclosed volume, there are 3 and they are using the sames vertexes and edges. I have corrected the model, but perhaps something that could be done in 1.9 in code to find out when it's going to streak/leak and not draw the other shape (would surely leave a hole on the projection but will not draw as it doing now.

From what I seem to understand of the current shadow algorithm is that it need the shape to be closed inside one single volume, having the volume separated seem to be unsupported. So if modeler are enclosing they models properly, the shadow will not give any problem.

It remind me a lot in school when we were making mistakes on our HL maps and they were "leaking" ..

christianclavet wrote:From what I seem to understand of the current shadow algorithm is that it need the shape to be closed inside one single volume, having the volume separated seem to be unsupported

Yes, this is one of the most important requirement for a stencil shadows algorithm.