> > make sure you're viewing a voxel with ra2 normals first of all. i would pick one that has very rounded parts. like the kirov or the hovercrafts skirt or the flying saucer.
> Or the V-3 Missile. Also nice.

i think it's too small and cylindrical so you will only get one row or columns of normals.

> > use the ability of the paint program to take a snapshot of the screen while Olaf's voxel viewer is running in normals viewing mode.
> OK: here's the first prob:
> - When viewing the bottem of the saucer, you can see that XCC will display 10 sections, while the viewer in the SE will show 14! Am I having a problem with telling gray-shades apart, or is one of the editors wrong?

yes you could be having trouble telling grey shades apart. it's difficult to see. but doesn't the dropper tell you exactly what the number of the grey shade is without you clicking on certain spots. btw why are you looking at the bottom of the saucer? does the bottom stick out? it's not flat right? are you not looking at the top because it has little details that break up the smoothness? btw will's viewer should display ra2 normals completely wrong since it uses totally random colors instead of shades of one color. it may use the same random color for 2 adjacent normals making you think it's the same normal.

> > use the dropper option which is used to pick up a color from one part of an image to use somewhere else on the image just because this option will tell you the number of the grey color at a certain spot. write down those numbers around a sphere. remember that the normals are distributed like the faces of a sphere. so guess what horizontal & vertical angle the normal is trying to represent. in other words if the normal was on a voxel of a sphere then where would it be? i think you should remember that there are 15 horizontal angles and 16 vertical angles (think of this like columns and rows if you like) . and that i believe none of them represent absolutely up and absolutely down but near to them.
> No, there is a top, and a bottem. Looked at the bottem of the saucer, and there was one greyshade (016) for the bottem, and the same was on the cable-tram. I think that 232 is the code for the top. See below for additional comments.

yes i did notice one number was used quite a lot to represent up and i noticed it mostly by looking at the decks of the ships. however remember that none of the normals are supposed to point exactly up or down. there should actually be exactly 15 normals that point equally up but in various o'clock directions and another 15 for down. when up is picked out by a computer algorithm then one number would be picked consistently but really there are other numbers that could just as easily have been used. it's just the way they programmed it. don't let that fool you.

> one number probably also means fullbright. if i were you i would probably represent the sphere on a piece of paper as if it was squashed down from the top so it bursts like a star with 15 sides or appendages or spikes or whatever you want to call them where the center means up and the ends mean down. write down the numbers along those "arms" of the star.
> So there will be 16 numbers (the vertical) on each spike. BUT: are these 16 numbers divided on a 180-degree base (so back and front doesn't matter in vertical numbers) or do they come back, so divided on a 360-degree base?

there should be 16 different gradients for how much up or down a normal is pointing so that spans 180 degrees and goes from nearly all the way down to nearly all the way up. there should be 15 gradients for what o'clock direction a normal is pointing. that spans 360 degrees. i'm guessing at least one set should mean exactly forward. since it's an odd number that also means no set will mean exactly backward but 2 sets are equally close enought to straight back. draw the star with 15 arms and write 16 numbers in each arm. please confirm these numbers first though. i know another way it can be tested. i'll go into that in another post though. i can't test it myself now though since i don't have the game. Again those test slab voxels i released to Olaf would be handy.

> > you should also note that Olaf's viewer reverses left and right so for example the millenium falcon which has it's cockpit on the right side will appear to have it on it's left side. so after you take a snapshot you should flip that image left right to stop from confusing yourself. btw Will's viewer is the same.
> OK. Mirrored

> > btw the 15 16 thing could be wrong but i'm pretty certain it's that way. i guess you could confirm this by looking at the flying saucer and counting how many slices around there are since the saucer appears to be perfectly round.
> Yes.

> > now i want only one person to be working on this coz it would be a waste for multiple people to do it. so try it out for a short while and if you think you can complete it then tell me, otherwise just quit and let someone else volunteer. don't volunteer to do it and then not do it coz then you would be wasting our time coz nobody else would be working on it coz they would be counting on you to do it.
> I think I can complete it, but I when I have a couple of spikes (or layers, vertical) I want someone to look at them and try to figure out the algorythm behind it. Will post the numbers in a new post.

i don't think there is an algorithm behind the scrambling. without the scrambling there is a simple algorithm but the scrambling seems totally random. it seems it was done by hand actually; just some guy mixing up the numbers in his head and entering them into a look-up table. i'm afraid you will have to go through every single normal. all about 240 of them. if i remember correctly they go from 0-240 and then there's a gap and then number 244 is used. that's probably a flag, but then 240 must also be a flag since it should stop on 239 since you begin on 0. anyway good luck.

> > btw in Olaf's viewer, for the ra2 normals, the normals numbers are directly used to get the grey color, but for the tiberian sun normals they are multiplied first to get the grey color that's used on the screen.
> Won't look at TS normals.