AR196 Project in AD2K

[smilo]

funny you should ask. great timing actually,
i was just thinking that i should post a project update.

after much frustration with the tail section,
i decided to move on to the mid-fuselage,
which is the aft gun-station/cockpit area.
it also includes the windscreen and canopy.
asper usual, i've encountered a few problems,
but, what's new?
we learn from our mistakes, right? right.

suffice to say, after days of working,
i decided that i didn't like
the way the fuselage looked
and i didn't like the way
i had the bulkhead templates laid out.
sooo, late yesterday, i tore it apart and redid it.
actually, i should clarify,
i started to redo it.
fortunately, so far, the canopy still fits.
although, i did take some artistic license.

i am still having problems with
the way cfs views parts,
which leads to visual overlapping,
or what are known as the dreaded bleeds.
what i am having trouble understanding
is just how this works.
i mean, i can build, say, the tail section.
i know that the sim draws the part
most distant first, then closer and closer to the POV.
basically, if one is out of sequence,
it bleeds through the others,
so, one must be very careful
to build in the proper sequence.
that's all well and good if you are viewing
say, from the forward/left to the aft/right,
but, what about looking from another point of view?
the sequence has changed,
and there are the damn bleeds!
aggravating, to say the least.
needles to say, i have pushed this problem aside
in hopes that i will come up with a solution later.
much later, i fear.

a few days ago, i did make an interesting discovery.
when compiling a model in ad2k,
an option is given to build in fs98/cfs or fs2k/cfs2.
since i am attempting a cfs model,
i've always chosen the fs98/cfs option.
just for the heck of it, i went with fs2k/cfs2,
just to see what would happen.
basically, not much that i can see.
the model now shows in both, cfs and fs2k.
and of course, the model has no bleeds in fs2k!

which got me to thinking,
what does fs2k have that cfs doesn't have?
mainly, i believe, it's z buffering.
whatever that is.

granted, by any stretch of the imagination,
i am not a programmer, but if i was,
i would try to figure out just what z buffering is
and find a way to make a cfs z buffering upgrade.
yes, i know, i'm a dreamer. so what?
and yes, i know there is a cfs z buffer upgrade.
it's called cfs2.
but, i don't want cfs2, damn it.
for the most part, i prefer cfs,
but i want cfs with z buffering to stop the bleeding.



so, how are things going with you?

[Ivan]

My understanding of the differences between CFS and FS2000 and later is that FS2000 and later use the relative locations of the objects to determine which should be in the foreground. In other words, it determines the viewing planes (Glue) on the fly, though it CAN be tricked by having objects for which relative locations don't indicate viewing priority. An example of this would be the Pilot and Canopy Frame. Both are similar distances from any viewpoint, but generally we prefer to NOT have the pilot show through the Canopy Frame.

With FS98 and CFS, WE determine where the viewing planes (Glue) are to some extent. Unfortunately most folks don't really understand how to use "Glue" to set the viewing planes. Also, there are application limitations such as serial listing of Pieces and Glue which don't allow for the flexibility of SCASM. There is also not the possibility of setting Pieces as ALWAYS being in the background.

Hope this makes sense.

As for how things are going, Anna Honey got back from Seville, Spain yesterday. She is watching kids today while I am at work.

I am working on repainting my P-40C to have the British Camouflage and AVG markings. (You really can't have a P-40 without a Shark Mouth!) There will also be an alternate "U.S. Army" paint scheme which is easy. I am still messing with the P-47D-23 to see what the rolling issue is all about. Veltro, Stuka, Dauntless, and Spitfire are all gathering at the door to the Paint Shop. The B-25C is in line for a minor rework for the Forward sections to use AF99 Template Parts to see if I can make the internal view of the Cockpit look better.

- Ivan.

[hubbabubba]

AF99, GLUE, SCASM, and Z-BUFFERING in a nutshell.

Let's imagine that we have a train with, from front to rear, has 1) a locomotive; 2) a tender; 3) a wagon, and; 4) a caboose.

Let's imagine that each section (locomotive / tender / wagon / caboose) is a "bleed less" section for whatever reason.

AF99 works a bit like a giant three-dimensional tick-tack-toe. It ask first if your POV is above, level, or under the train (remember that AF99 deals normally with aircraft, not train...), then if it is front or aft of the train, then if it is right, left, or center of the train. Depending on your POV position, it will refer you to a showing sequence that should, in theory, prevent bleedings. But in practice, it doesn't always work. For example, if you're POV is between the tender and the wagon and AF99 "decide" that it is a front position, it will send you to the showing order caboose / wagon / tender / locomotive, making the locomotive bleed through the tender. If it "decide" that you're aft, the showing order will be locomotive / tender / wagon / caboose, and the caboose will bleed through the wagon. The problem here with AF99 is that it has only front and aft choices, no "center" position.

You can "cheat" AF99 by creating glue parts. Glue parts are invisible "walls" that extend indefinitely in space; you are either on one side of the wall or on the other side of the wall. Depending on the side you're on, the showing sequence is toggled. For example, you could place a glue part between the locomotive and the tender - imagine an infinite wall standing vertically between them - that would show the locomotive / tender sequence when behind the wall and the tender / locomotive sequence when standing in front of the same wall. If you repeat the same procedure with the wagon and the caboose, you can now stand between the tender and the wagon without bleeds because the the glue parts are "forcing" the sequences to show correctly.

SCASM can do even better. In AF99, you can glue the locomotive to the tender, the tender to the wagon, and the wagon to the caboose, thus making the whole train bleed-less in an "Ivan's conga" fashion. But AF99 will still go through the tick-tack-toe charade, each time pointing to your "glued train", no matter where the POV is. SCASM can get rid of the tick-tack-toe and the gluing sequence can be done in various ways. In the case of the train, SCASM can do;

- locomotive / tender / wagon / caboose (AF99 too);
- caboose / wagon / tender / locomotive (AF99 too);
- locomotive / (wagon / tender) / caboose (AF99 can't);
- (wagon / tender / locomotive) / caboose (AF99 can't);
- (tender / wagon / caboose) / locomotive (AF99 can't);
- (tender / locomotive) / (caboose / wagon) (AF99 can't);
- (caboose / wagon) / (tender / locomotive) (AF99 can't);

And a few other combinations where "/" is a glue part and "()" a section that can be treated as a whole section. SCASM is therefore more flexible when it comes to gluing sequence. With the train, the advantage is academic, but it is quite useful for more complicated objects. I'm not sure, but I think that AD2000 works a bit like that; ordering within a group that is glued later to another group.

Now Z-buffering. From FS2K on (CFS2 included), the visual engine enters all visual polygons in a buffer and shows them in a sequence that goes from the farthest to the nearest, no matter glue parts or not. This works pretty well, but not always. For example, a roundel on the top of a wing may flicker if it shares the same plane as the upper wing surface itself; they're both at the same distance. A bended polygon can also cause problem, like the "pioneers' tools" on the left side of my jeep. In CFS1, the shovel is intentionally bended like any decent shovel should be, but CFS2 considers that part of the handle is behind the side wall of the jeep and makes it disappear...

This was, in a nutshell, what I know about the subject. I hope it helps you understand what you're doing and why it works... or not!

[smilo]

thanks gents, that was informative.
what is bugging me is the pov.
i can make it work from one pov,
but it's all screwed up from the others.
i can see that i will need a lot more trial
and error before i figure it out...mostly error.

another thing i need to do
is translate AF99 terms to ad2k terms.
there is no "glue", there is a technique
that seems to do the same thing.
i can't remember its name right now
and i don't want to fire up
my other machine just to refresh my memory.

also, as i recall, i can build an assembly,
say, for the wings, floats or whatever,
save it, then add it to another assembly.
there are certain limitations and rules for doing it,
but i'm sure it can be done.
i plan on using it later on,
but for now, i just want to build assemblies.

in ad2k a master project is called a master assembly.
a master assembly is made up of sub-assemblies
which contain say, the tail cone, aft fuselage,
vertical stabilizer, horizontal stabilizer, rudder, etc.

i just remembered the term!!!
seal planes are used in ad2k fs98/cfs models
as, i believe, "glue", is used in af99.
the trick is figuring out the proper way to use them.

[Ivan]

Hey Smilo,

In my mind there is no question that AD2000 can do more than AF99. It is fairly easy to tell by just looking at the end results. Since you are leading the way with AD2000, I don't think you can really expect any help from those of us who have not played with this tool. I may play with it eventually, but got too much stuff going on and too little time to experiment at the moment.

- Ivan.

[smilo]

as i see it,
there are basic similarities in the programs.
mainly, in the way pieces are drawn
and the way they are viewed by the sim.

if nothing else, over the years,
i have learned to expect nothing.
while it is true that i would gladly accept assistance
from anyone who has experience with ad2k,
i will be the first to admit,
that i have as yet,
not met that person.

so i will continue to muddle along.
hopefully, learning something as i go.

[No Dice]

You are going to get it done Smilo, and you are going to have the only worthy AR196 on the net.
Well worth waiting for.

Dave

www.thefreeflightsite.com

[hubbabubba]

thanks gents, that was informative.
what is bugging me is the pov.
i can make it work from one pov,
but it's all screwed up from the others.
i can see that i will need a lot more trial
and error before i figure it out...mostly error.

(...)

i just remembered the term!!!
seal planes are used in ad2k fs98/cfs models
as, i believe, "glue", is used in af99.
the trick is figuring out the proper way to use them.

I would bet that these "seal planes" are the same as the gluing parts in AF99. I would eat my mustache if they're not!
(I don't mind; it will grow up again anyway...)

P.S.- Just went and check the English help file;

The following only applies to FS98™ and CFS™. FS2000™ uses a new viewer that makes useless the seals and the associated JUMP PLANE/PLANE instruction pair. When you design for FS2000™, you just need to build the sub-assies, assign them a color or a texture and make them move. That's it !

What is a seal...
A seal is a chain of vertices defining a plane that is used to sort the different parts of the 3D model during the viewing process. Parts of the 3D model can be either sub-assies, sets of sub-assies, or sets of polygons.

Let us assume a given seal interfaces the sub-assy "A" with the sub-assy "B". Depending upon the position of the virtual observer with respect to the seal, he will see "A" in front of "B", or behind it. Thus, the viewer must draw "A" after "B" or before it... The choice of appropriate seals is very important in the design of a model.

The seals are used in the Model Editor (instruction JUMP PLANE) but they must be declared in the Graphic Editor. To do that, you just have to select an existing chain or a template and add it to the seal list which is maintained by Aircraft Designer 2000. The plane of the seal is defined by three points of the supporting chain or template.

The seals may be displayed in the Graphic window or hidden. The show seals option of the Utilities menu controls the display of the seals.

I think my mustache is safe for now

[smilo]

save your mustache, hubba.
i won't take that bet.
i may be dumb, but i'm not stupid.

one thing i just noticed reading your quote,
templates can be used a seal planes.
i didn't know that.

the thing that's odd,
is that i was able to do it
with the Do17 double rudder,
but am having fits with the 196.
go figure.
maybe, i'm looking too close.

thanks for the encouragement, Dave.
it make take forever,
but i'll do my best.

[Ivan]

Hello Smilo,

Your project can't take any longer than the ones I have worked on. Now THOSE take forever. It has been about a year since I last worked on the Dornier 17Z though I can honestly tell you I have been far from idle.

- Ivan.

[smilo]

WOW!!! a year, huh?
has it been that long?
time flies when you're having fun.
or, in my case,
time flies when you're getting old.

funny you should mention the Do17.
just last night, i was thinking
about opening it back up,
but decided against it.
i don't want to get any more confused
than i already am.

even though it's the project
i set down to start this one,
my weak brain can't have too much going on.

i swear, i must have adhd or something like that.
(there was no such thing when i was a kid)
i have a b*tch of a time staying on task.
it seems like it's a constant struggle.

for example, just a few minutes ago,
i was working on the 196 canopy.
been at it for several hours.
all of a sudden, for no apparent reason,
i decided to check my email,
there was a cfs forum post notification,
and vua la, here i am.
(hubba, please correct my French, vua la)

[hubbabubba]

WOW!!! a year, huh?
has it been that long?
time flies when you're having fun.
or, in my case,
time flies when you're getting old.

funny you should mention the Do17.
just last night, i was thinking
about opening it back up,
but decided against it.
i don't want to get any more confused
than i already am.

even though it's the project
i set down to start this one,
my weak brain can't have too much going on.

i swear, i must have adhd or something like that.
(there was no such thing when i was a kid)
i have a b*tch of a time staying on task.
it seems like it's a constant struggle.

for example, just a few minutes ago,
i was working on the 196 canopy.
been at it for several hours.
all of a sudden, for no apparent reason,
i decided to check my email,
there was a cfs forum post notification,
and vua la, here i am.
(hubba, please correct my French, vua la)

Actually, smilo, it is "voilà", but your phonetic deformation of it may be more appropriate; I would try "vua-lah" to emphasize the "flat" a at the end.

Leonardo Da Vinci is renowned for his tendency for multi-tasking and leaving things unfinished. For every painting, sculpture or invention he completed, 90% was left mid-way. So, you're in good company.

The Taifun took me 2-3 years of research, 2 years of modeling, and then I came to this forum for would-be testers and, after more than a year, I finally released it. But who's counting...

[Ivan]

Voila!

Pardon my French.

If you think THAT is ADHD, you should see what I have been working on in the last few weeks:

In looking over the Stuka's Canopy, I realised that I spent way too much time trying to make mirrored Left and Right Components. I edited one of my earlier programs to take a AF99 .AFC Component file and flip the bit that indicated a mirrored or non-mirrored part. The result was to make a process that was very error prone and would take 20-30 minutes into a program execution that would take about 5 seconds with no possibility of error.

You already know about the fix for Tails bleeding through the Cockpit from an interior view. After figuring out what to do, I executed the process with about a half dozen of my projects.

There are also a few more items that are not worth mentioning in a discussion that is publicly viewable.

- Ivan.

[Ivan]

I started typing the last message a couple hours ago. My son has had my computer for school work and games until a little while ago.... Missed Hubbabubba's response before I submitted mine.

Hubbabubba, I had the impression the Taifun was a much quicker project than that. I guess we never see what others are researching. Regarding research, I am doing a fair amount of that myself. So far since yesterday, I have gone through manuals for a P-47 Thunderbolt, F6F Hellcat, P-40B/C Tomahawk, and a Hawk 75 / P-36.

It would be really cool if something actually resulted from all the reading I have done.

- Ivan.

[smilo]

VOILA!!!
thanks fellas.
as soon as i saw it,
the memory banks kicked in.

as i've said in the past,
i am continuously amazed
by how many french words and phrases
i use in my everyday conversation.

it's to bad that i didn't pay closer attention
in that french class i failed back in the ninth grade.
so it goes.

i hope i am not kidding myself,
but i feel like yesterday was a very good 196 progress day.
even though i can't say that i have it down pat,
i am starting to figure out the viewing order process.
trial and error, trial and error,
and then some more trial and error.

i have been trying to put together
the gunstation/cockpit area which includes
the fuselage exterior and interior,
the exterior and interior canopy framework,
panel area, and later, the tailgun needs to be added.

one thing that i've noticed is that
pieces must be added within a sub assembly
in the proper sequence, or else
the texture shading does not look right.
for example, if i cover 5 bulkhead templates
from a forward bulkhead to the aft,
and then later, decide to add 1 more template
just forward of the 5 i previously did,
there is a shading issue with the new
far forward group of pieces.
i have tried changing the sequence in the model editor,
by moving the new pieces to the top of the list,
but still have the issue.
i went to bed with this problem
and might have come up with a solution.
i need to go test the theory.

maybe, a few screen shots would help clarify.
next post, after i fire up the working machine.