Topic: Descent 3 Advanced Tutorials...

[Babylon]

(rats, there's no silver text option like in old times...)

Just wondering that, with D3 modding and level building seeming to be at an all time low nowadays, if it would be worth my time to draw up advanced tutorials explaining the various nuinces of things (like extensive .gam tablefile documentation), as well as explaining modding and special effect tecniques.

Because I don't want to waste my time if it won't be useful to anyone.

[-<WillyP>-]

Yeah, I'd love to see some more documentation on the gam file.

[Kaiaatzl]

I know some of us (me for example) would really like to build levels for D3 if we can find better or at least faster ways to do it (and make more than bad or  mediocre levels).  I want to do some singleplayer levels (I tried the scripting with a few *very* simplistic levels and the scripting was actually fun).

I'd really like to see some new tutorials, most of the ones I've seen so far were incredibly hard to understand (I haven't been able to use any of the lathe tutorials properly because they didn't explain how you had to place vertices to make the structure you wanted).  Except for the scripting tutorials, most of those were actually pretty useful (but you need to have a level in order to script it ).

[-<WillyP>-]

Here's a good place to get started: DescentiaPedia: Descent 3 Level Editing

[Scyphi]

I've long had ideas for creating D3 levels, but had all just about given up because of D3Edit's little nuisances that drove me over the wall, that and the time and effort it took, which was more than I could dedicate to it for any given length of time. For me, it was just easier to make D1-D2 levels with something like DLE-XP, settling upon D2X-XL as the next best thing to a D3 level.

So I suppose if the tutorials made things simpler for one to use D3Edit, or at the very least get me to be more patient with the at-times very stubborn program, then I suppose I'd be for it.

Assuming I can find time for it, of course.

[Babylon]

What were your issues with it?  Was it just learning the interface controls, or understanding how everything related?

If it had to do with stability or interface, I can't help with either; although Atan has made great strides in the improvement of both areas.  There is a drop down menu that he added that shows all the user controls at any time, as well as auto-save and recover features that help in the (now) rare moments that it does crash.

If it's understanding how to model geometry (it's not too much different than modeling in Blender, except that you're forced to manipulate faces and vertices in the grid view rather than the main-view screen), or the relations of rooms/portals/terrain, then maybe it was never explained well for you.

[Kaiaatzl]

I think the problem is more the time that it takes, both to learn the editor and to build even a simple level.  Like I said earlier, learning basic Dallas scripting for me was actually faster than learning to build a basic room in D3Edit.  Good tutorials could really help with teaching the editor faster (and maybe even help speed up level building times).

[Matthew]

D3 isn't designed with the simple, blocky levels of D1 and 2 in mind. However, that's not say it takes a long time to make simply blocky levels if you just use the import room and extrude face functions.

[SaladBadger]

I'll be completely honest: Coming from a Radiant (and a little bit of UnrealEd) background, I find D3Edit very very irritating to use. This is probably me being way too used to defining and scaling things using the mouse, which I still have not figured out how to do. Atan's work has helped a ton though, but it still isn't enough to salvage the editor in my eyes.

However, the lathe tool is awesome.

[Kaiaatzl]

I cannot figure out how to use lathe.  None of the tutorials helped, ever, as they just showed how to make specific structures within specific rooms... they didn't say anything about how it worked (I'm one of those people who needs to have some kind of background knowledge about how the tool works in order to use unless it's very intuitive).  I can't see how the pattern of verts placed has anything to do with the structure created (the tutorials say to do the math, but they don't say what math to do).

@Matthew, Half-Life 2 wasn't made to be simple and blocky either, but I've been using Valve's Source modding tools for a while and I found the map building with Hammer to be very fast and intuitive (even for complex rooms and areas).  And the Source engine is quite a lot more advanced than the Fusion engine.

[-<WillyP>-]

The theory behind lathe is really pretty simple. There is no math to do, unless you really like math and want to do some, in which case pm me and I send you some copies of my kid's homework worksheets.


Picture a simple cylinder. stand a can of pea soup on your desk if it helps. Or better yet, open D3Edit and make a new cylinder.  Now picture an axle through the center of the diameter of the top and bottom of the can.  Let's say this axle is the y axis, in the cylinder in D3Edit, it is the vertical axis in the two left view-ports, the top showing [XY] and the bottom [ZY].  In the [XZ] view, the axle is shown as a single dot in the center of a disc.

Now, if we had created this cylinder in lathe, we would place verts in a vertical row a certain distance from the Y axis in either the [XY] or the [ZY] view.  The distance from the Y axis to each vert determines the radius of that section of the finished lathed model.

So if we started from a blank screen, we would have a Y axis in the [XY] view port, represented by a green line, and a vertical row of verts, represented by blue squares, some distance form the Y axis, and parallel to it.  Lathe will copy the verts, rotate the copy a specified distance around the axis, and paste the copy. So, if you choose to make an eight sided model, lathe will rotate the copy 45 degrees, then ninety, etc., until there are eight sets of verts, arranged radially around the axis chosen.  And of course make faces, too.

Now, the nice thing is that the verts, which in this example are all the same distance from the axis, don't have to be all the same distance from the axis.  So, in the [XY] view, lets make the top vert ten units from the Y axis, the next twenty, then thirty, etc..  In the [ZY] view, they will still be all in a straight line, superimposed over the green line representing the Y axis.  In the [XZ] view, we will see them as a row of verts, extending in a straight row starting ten units from the intersection of the two green lines.  The Y axis, in the [XZ] view, is represented as a single point at the intersection of the two green lines, and extends perpendicular to the two green lines in both directions. Lathing this example will give us a model of a cone with the pointed top cut off.

[Foil]

Exactly.

In simple terms, the lathe tool just takes your points and rotates them around an axis, creating faces along the way.

[Edit: Attached a quick example (whipped up in my cad application) of a couple of points near an axis.  Lathe them all the way around (using eight facets here), and you get a cylinder.]

[Edit: One more quick example, with a couple of extra points to be rotated, and the points highlighted on the result.]

[Scyphi]

As for me, I never did quite determine what exactly I didn't like about D3Edit, but it always frustrated me. I think it was simply the way it insisted some things had to be done, which to me seemed more complicated.

It wasn't about not understanding the program, because eventually I slowly figured out how it was supposed to work. I just viewed it as...inefficent. Other programs, such as Blender, can do the same thing a heck of a lot easier and simpler. Plus, with Blender, it's easier to further tweak things in ways that D3Edit just about outlaws.

Since the subject came up, though, I will also agree that Atan's modifications to the program have made it a heck of a lot more tolerable to use, and I wish I could say I had the time to really mess around with some of those new improvements, but I can't.

And yes, time is also an issue, but that's not D3Edit's fault.

[Babylon]

With complicated geometry comes complicated options and tools.  And greatly complicated responsibilities.

You know, if you truly insisted on it, you could build everything in Blender and just puzzle it together in D3 Edit; in fact, that's how the Descent 3 single-player was originally made (except Outrage used 3DS Max).  Although D3 Edit can create geometry, it was made by Outrage really as a puzzle-piecer-togetherer.

@Alter-Fox: Source had better be way better than the Fusion Engine - it's at least six years older (1998 - 2004) and a generation ahead.  It's easy to forget though that the Fusion Engine, although unfinished, was years ahead of its time, and easily provided the best graphics available until Soldier of Fortune 2 came out.  (Yeah, that's my opinion and I'm sticking to it.  ;P)  At any rate, it certainly beat the Source engine of the first Half-Life...

[Scyphi]

Yeah, but I really don't think D3Edit really needed to be that complicated, and I know there are people out there who agree, or at the very least can relate. Like I said, Blender can do a lot of the same things and more a heck of a lot easier.

And the thought has already crossed my mind to just build it all in Blender then piece it together in D3Edit, as that would be much easier for me. Unfortunately, there is the issue of converting files so that both programs could read them, and the sad truth is that I don't know many means of doing it, at least not with Blender. Granted, I once found a converter program that could do it, but it'd cost me fifty bucks to get the full version just so I could use it, and I wasn't willing to pay it. 

Maybe that can be Atan's next project, figure out some way to get D3Edit to work around that issue...whatever the case, if it can be I would be forever indebted to that person.