What follows is entirely speculation.

If anybody else has any other opinions, please let me know.

Examining some .3db files converted to directly to .3dzs, or text dumps of those files, I came up with the following:

3db Length Reality Ratio

havoc 4264p 17.01m 250.68p/m
hiphi 4427p 18.2m 243.24p/m
tiger 3591p 14.08m 255.04p/m
uh60 3902p 15.6m 250.13p/m
ch53 5783p 22.35m 258.74p/m
hind 4010p 17.5m 229.14p/m***

Here's the same data in a more readable format:


Most of the ratios are suspiciously close to 254 points/metre, which would equal 10 times the resolution of the accepted decimal foot value for .3dz files.

The Hind is an anomaly, but I can't look at the model directly, since it has too many points and elements to convert to a .3dz.

There's a problem sometimes when converting stuff was originally made in .3db format to .3dz. If a model has in excess of 256 nodes or elements, it can't be converted to a .3dz. The only thing to do is to look at a text dump of the .3db. But then, you never know if there's extraneous nodes in the model or not. You also can't tell if the model is shorter (or longer) than the accepted fuselage length on account of FLIRs and the like being added via hardpoints.

For a couple of the available models converted to .3dz, I had to manually remove extraneous rotor nodes.

What all this crap boils down to is that I suspect 1 3dz point = .0039737 METRES.

If ya' wanna' build a model in 3dz Studio, however, you can't just look at the length of the model in the project’s 3-View Data box and divide by ten. You have to look at the 3dz point numbers in the first column of that particular box (wingspan, length, and height) to determine the size of your model.

For instance, if you convert the CH53 from .3db to .3dz, the fuselage is 5783 3dz points in length. 3dz Studio will show the fuselage length in metres as 259.12. If ya' divide that length by 10, you get 25.912 metres. Often, the length of the CH53 fuselage is given as 22.35 metres. That's a discrepancy of more than 3.5 metres. But if you take the 5783 3dz point length and multiply by 0.003937, you end up with a length of 22.95851 metres. That's much closer to the generally accepted CH53 fuselage length of 22.35 metres.

I can't imagine that anyone actually cares about this sorta' junk, though.

Smartypants.



That's pretty slick research, Samovar. The scale thing always kinda puzzled me- it didn't seem consistent, but I guess it is.

When WB et al (sorry I can't recall the modeller at the moment. Dang whiskey...) added the Chinook, I was a bit offput by how out of proportion it looked alongside the Apache. Until I looked it up and learned the real Apache is actually LONGER than a CH-47.

Anyway. well done.

...

Are we going to see more MODs here in the near future SAMOVAR? ... your research looks like a stepping stone towards more models. Daggon' ... my GUNSHIP! is gone ... ;D

.

Anyone know what happened to development of the "The Times of Sand" mod for Gunship!? Looked really promising. The beta exe really improved the sim aspects of the game, and it always did look pretty good. Anyone know whether the Mod M8 add on that depended upon an older Adobe reader will still work? It's pretty ambitious and adds a lot to Gunship!

I think you and I talked about this when I was adding some features to Gurney's 3dz Studio. I added a field in 3dz Studio, under 'Tools', 'Options', where you can specify how many 3dz points there are to a meter. Gurney had hardcoded the program to use these constants: '1 3dz point = .0448 meter', and '1 3dz point = 1.76 inches', which work ok for EAW models, but not Gunship!. For Gunship! models, I'm using '1 3dz point = .0042 meter'. Then, to convert to inches, I use this calculation: 'inches = Meters_per_point! * 3.28 * 12', which, using .0042 meters/point, results in .165 inches per point. Again, this is a user specified option I added to the program.

Then, those factors are used in the 'Centre Mover/Rescaler' and 'Project' dialogs.

Of course, you can readily see that, since Gunship! uses many more 3dz points/meter than EAW (around 254 vs. 22), that we can create smoother contours on our models, which we also talked about. (1/0.003937 = 254)

I just realized I should add to the Update readme.txt that, for Gunship! modelers, they should change that field to something like what you said above, .003937 or something else close to that. In fact, I'll add a link to this thread. Also, I'll add that that calculation is necessary for the Centre Mover/Rescaler and Project dialogs.

Thanks for the heads up.

to fng2k: I likewise was a little surprised that the Chinook is smaller than the Apache (Sea Stallion too, I think) and also, I spent a *lot* of time trying to scale that model correctly, particularly getting the width, length, and height proportionally correct, and the contours also. In the next few days I'll analyze the model and see if I should rescale it.

lol, I hope you all will forgive me. I had to re-edit my above post a few times to correct some inaccuracies.

Nothing at all to forgive, WB.
The scale business had often confused me.
Shortly after I started this thread, way back when the world was young, and alternating current was only one of Tesla's pipe-dreams, I'd begun to think that the "real" scale in GS was 256 3dz points/metre. There seemed to be ample evidence in the .exe to support that assumption.
Later on, I ran into some inconsistencies while mucking about with object placements on terrain tiles, but I don't wanna' talk about it, though. I've been traumatized ever since.

I would think that it's somewhat safe to say that the scale is somewhere between 250 3dz points/metre and 256 3dz points/metre. That range ain't a big deal, as far as I'm concerned.
Most of my later models were built on 256 points/metre.

Not meaning to presume anything on your part, but why do you think the scale seems to vary between models?

Here's what I suspect:

• 1) I suspect there *was* a standard scaling factor for the Microprose artists to use.
• 2) When each artist would start a new model (and remember, that they *were* artists, and not programmers or military experts), they created the first few nodes and elements to their standard scale.
• 3) But, as they added, deleted, shrank, expanded, reshaped, etc. the elements to make the model more lifelike, the model became slightly out of scale.
• 4) Then, when they finished their model, either they didn't have a rescaler to rescale the entire model back to specs, as we have...and/or, they decided the finished size was 'close enough'. (Also remember that rescaling might introduce new inaccuracies due to rounding, etc., which is why I'm leery of rescaling in the first place.)
• 5) Then, they gave their finished model to the programmers, who, after inserting and viewing it in the game, *also* decided the scale was close enough.

Well, I'd assume you came to the same conclusion I did, not sure though.

Edit: Or, another possibility is that the artists had some threshold scale value(s) such that, if their finished model *strayed* far enough from those, that *then* they would rescale the entire model, but not otherwise. Impossible to say at this point, of course.