RAD2 shell 129

[Marcello]

Just to let you all know.... with RAD2 over (and a very satisfied 'me') work on Shell 129 is not finished. I'm probably halfway there or not even that. I have been working and solving little problems in the past week. Hopefully I can start working on additions this weekend. Some things I have done:

1. New texture for big waterfall
2. second smaller waterfall on rock
3. added first animated foam below waterfall
4. cleaned up some stuff
5. trying to solve alpha/waveset rendering problems
6. experimented with bench alcScript (working!!)
7. got me some nice waterfall sounds

The foam below the waterfall is just like it is below Giras waterfalls. It looks really cool and adds a lot. Alpha is rendering screaming bad though. Can't wait to see the end result when that is solved.

To do:

1. solve alpha/waveset rendering problems (big 1st priority)
2. add foam below all waterfalls
3. new more real texture on lower waterfalls
4. add sound to waterfalls
5. design and model benches
6. experiment with models and animated textures for fountainwater
7. oh yeah and experiment with textured object shadows

[Lontahv]

I actually find it interesting that the waterfall is xraying the waveset... waveset's usually seem to have a bad habit of rendering on top.

All I can say is do a little experimentation with the passIndex for the various objects.

Little Plasma lesson:

Most geometry is stored in a few big classes called "plDrawableSpan" now these are the actual things that the engine comes around to read and the names are VERY important. So, if you have some spans named something like:

SomeAge_District_000000_1
SomeAge_District_000000_2

The geometry contained the first span would render first (sorted by name).

What PassIndex does currently is change the number at the end of the span IIRC. So, if you set an object to passindex 7 it'll create spans that look something like this:

SomeAge_District_000000_0 <-main data in here
SomeAge_District_000000_7 <- your object with PassIndex set to 7 in here

I _think_ this is how it works... or it works for the most part.

So, if you put the passindex for the fountain and droplets higher the problem should be fixed.

Note: Some of the span-names listed here may be whacky.

[Aloys]

Personally I think it would look just as good if you got rid of the wavesets for the waterfalls and go with something more like Gira\'s waterfall. Just my opinion.
That aside you shell is really cool.

[Lontahv]

O.o waveset waterfalls.

[Marcello]

Actually the waterfalls are like Gira. Although not visible in the contest shell they actually have a Gira like texture now. I have never used wavesets for the waterfalls, although I do admit, the texture in Pahts 8E does look waveset-like.

I have problems:

1. The alpha rendered over a waveset doesn't really alpha, but shows a black jagged border around the alpha parts
2. The 100% alpha parts are alpha, but those inbetween don't show the waveset, but what's behind it.

The third one is solved. That's the one where the alpha textured object automatically got an evnmap. Thnx to Trylon's suggestion by PM last night I managed to set the correct flags by hand.

So what's left to fix is this:



I'm using Alpha PNG's btw. Everything not rendered over a waveset look good.

I will dig into your suggestion tonight or tomorrow Lontahv!

[Aloys]

Wooops, I totally hadn\'t noticed they were actual textures.. shame on me.
Forget anything I said then. (well except for the \" your shell is really cool\" part )

[Kato]

Hey, waveset waterfalls would be really cool.

Unfortunately, that would not only require an engine modification, but would require a waveset overhall not gonna happen.

Anyway, I'm assuming you've fiddled with relevant PassIndex settings? Meh, wavesets are wonky

-Kato

[Aloys]

Why would that require any change? wavesets only work on flat surfaces?
[edit]Don\'t answer, I just took a peek at the wiki.. That\'s really too bad that it can\'t be used on arbitrary geometry.. It even could be used for things other than waterfalls.. (Stargates anyone?)

[Marcello]

Yes I played around with Passindex and set Drawable flags to get rid of the envmap on the waterfalls. Just send my .blend to Trylon who will have a look at it too.

Not sure if I still have to do something with z-index and ztrans. Haven't done so now.

[Marcello]

Ok guys, thnx to all of you and some in depth help by Trylon my alpha-problem-days are over.

Passindex works like a breeze, but not over wavesets. Multiple overlapping objects that also overlap a waveset are sorted ok in relation to each other, but not in relation to the waveset. The waveset is invisible and through the alpha the object behind the waveset are seen. Moreover the edges of the alphachannel are jagged and show up black. One simple addition to my AlcScript fixed this:

Code: Select all

<object>:
    type: object
    visual:
        drawable: 0x40000004


Adding this code to all objects overlapping a waveset and katsjing: looks great

Btw there were some other not waveset related things I had to fix to make it work. For instance the big waterfall was duplicated and always showed the wrong one on top. Etc..