Hard Light Productions Forums

Modding, Mission Design, and Coding => The Modding Workshop => Topic started by: Omniscaper on January 08, 2004, 12:09:04 pm

Title: How feasible are PLANET models?
Post by: Omniscaper on January 08, 2004, 12:09:04 pm
I've been searching alternative ways to handle planetary backdrops that are not so static. Animated textures seems like the way to go as well as large textured planes (B5-Project Earth backdrop). I'm currently modeling a ringed planet for an animated render for texture use, but I was tempted to plop the model right in the game!

What are the size limitations and draw distance does Freespace work with? If I could find out those specs, I could at least make planetoids or localized nebulae.
Title: How feasible are PLANET models?
Post by: Lightspeed on January 08, 2004, 12:13:12 pm
Draw distance used to be 150 km, with HT&L it's 30 km.
Title: How feasible are PLANET models?
Post by: phreak on January 08, 2004, 12:17:29 pm
its extended out to 10,000,000km now.
Title: How feasible are PLANET models?
Post by: kasperl on January 08, 2004, 12:19:24 pm
there might be a problem with poly usage though, to get a convincing sphere and all.....
Title: How feasible are PLANET models?
Post by: Omniscaper on January 08, 2004, 12:20:11 pm
Whoa. So much for progress, in that arena. That is strange that it shrunk. Weren't there levels in Retail FS2 that were larger than 30km. Wouldn't HT&L builds muckk up it up? Anyways...

Thanks for the tip. I guess I'll make Mini Me's homeworld.
Title: How feasible are PLANET models?
Post by: Omniscaper on January 08, 2004, 12:21:32 pm
Whoa, insta-posts, I was responding to lightpeedwhen I finished typing that. Time to brush up on those typing skills.
Title: How feasible are PLANET models?
Post by: Omniscaper on January 08, 2004, 12:25:42 pm
10,000,000 km!!? I hope thats not an exageration. I suspected a sphere issue would arise. You suppose a 15k sphere would be enough for a convincing textured planet?

The point of this endeavor is to avoid a looped animated texture of clouds and to replace it with a rotating sphere utilizing the transparent alpha channels.



My bad Mr Kasperl
Title: How feasible are PLANET models?
Post by: kasperl on January 08, 2004, 12:58:23 pm
uhm, Omni, no offence, but look to the bottom right of your post, a little button next to the "quote" one
(http://dynamic4.gamespy.com/~freespace/forums/images/edit.gif) (http://dynamic4.gamespy.com/~freespace/forums/editpost.php?s=&action=editpost&postid=392429)
Title: How feasible are PLANET models?
Post by: Nico on January 08, 2004, 01:09:13 pm
Quote
Originally posted by Omniscaper
10,000,000 km!!? I hope thats not an exageration. I suspected a sphere issue would arise. You suppose a 15k sphere would be enough for a convincing textured planet?

The point of this endeavor is to avoid a looped animated texture of clouds and to replace it with a rotating sphere utilizing the transparent alpha channels.



My bad Mr Kasperl


15K? that's way enough. With 2K, you'll need to be quite close ( so close it doesn't fit completly in the screen ) to begin to distinguish the polys. That's 2K triangulated btw ( just tried in Max, I don't like giving numbers out of my ass :p ).
Title: How feasible are PLANET models?
Post by: Omniscaper on January 08, 2004, 01:15:46 pm
Ok then, kool. 15k it stays. Is 60,000 meters or kilometer sphere within drawspace?

I tried a Truespace measurement of 6 meters and hit the multiply function in PCS by a factor of 10000 and I see nothing but the targeting reticle for the object. Fred shows that I'm outside its radius.
Title: How feasible are PLANET models?
Post by: redsniper on January 08, 2004, 08:19:28 pm
just so you know, each mission takes place in a 1200 kilometer cube, so don't make any 60,000 km planets (60k m is okay tho)
Title: How feasible are PLANET models?
Post by: Liberator on January 08, 2004, 08:28:56 pm
Inferno already has a planet model.  And it looks damn good!
Title: How feasible are PLANET models?
Post by: Nico on January 08, 2004, 08:30:02 pm
heh? The planet doesn't HAVE to be in the playable zone :p
Well phreak says it's 10,000,000 so you should be able to put a real sized planet in there ( I have a couple like that actually, oe of them is a ringed planet, and I'm pretty usre that's the largest pof ever :D )
Btw, about the polys, if 2K is enough, don't go and put 15K! If you want a margin so you can be very close, put 3K et voila
Title: How feasible are PLANET models?
Post by: Omniscaper on January 08, 2004, 10:47:12 pm
How do i place a planet outside playing zone, without being ignored by the drawing engine. Ugh, could you folks give me a link to existing planets, besides earth (joke). Need a reference.

On a side note, are alpha maps in DDS or TGA's strictly opaque and invisible? I was hoping that the greyscale would give different levels of transparency. But i am having trouble getting semi transparent surfaces. I made an earth textured sphere within a larger cloud textures one with alphas.
Title: How feasible are PLANET models?
Post by: Nico on January 09, 2004, 03:51:29 am
How do you place it? well, you put it in fred, far away from the player :p
Title: How feasible are PLANET models?
Post by: KARMA on January 09, 2004, 05:21:07 am
IIRC it's transparency level, but it is not supposed to be used on models, it's just for backgrounds. From what said the coders, if you put transparency levels on models you just can't be sure of what will be the result, and the result may vary on different systems
Title: How feasible are PLANET models?
Post by: Carl on January 09, 2004, 05:26:38 am
for a game a resonable resolution for a planet model should stay at 25 latitude by 50 longitude, or 1250 polys. no need to go higher than that.
Title: How feasible are PLANET models?
Post by: Nico on January 09, 2004, 06:48:25 am
Well, seeing his ships are already unresonable, I don't think a  planet will change much to the total :p
Title: How feasible are PLANET models?
Post by: aldo_14 on January 09, 2004, 07:51:19 am
At long distance you could maybe use a mix of transparent and normal pixels for that sort effect.........  if the planet is far away enough it should be impossible to distinguish between the transparent & opaque regions.
Title: How feasible are PLANET models?
Post by: Nico on January 09, 2004, 09:21:37 am
In theory. But in fact, due to the screen lining stuff, camera rotation, etc, it just makes a big ugly thing, and the mipmaping makes it even worse.
Title: How feasible are PLANET models?
Post by: Woolie Wool on January 09, 2004, 09:53:38 am
I love planet models. The way they interact with the light and really look 3D makes them ten times better than background images. If I can get a modeler on the Starforce team (hint  hint), I'll have full-sized 3D models of the planets Savara, Tyrian, Deliani, and maybe Erigone made (I already have Earth from Inferno).
Title: How feasible are PLANET models?
Post by: RandomTiger on January 09, 2004, 10:11:46 am
Quote
Originally posted by PhReAk
its extended out to 10,000,000km now.


If its set to high it can degrade the effectiveness of the z-buffer I believe.
Title: my experiemtns
Post by: Star Dragon on January 09, 2004, 12:30:48 pm
I failed to make a viable deathstar (the battle versus size and siccessful interactivity). If I made it total size, Fred crashes, half size fred works but in game it's crap and you fall through the planes. Making it small eough to work with means you ruin the illusion of size...

Experiemnt number two was a major success though. I designated a sphere to be  a moon and used a lunar texture on it 360degree wrap. next I scaled it to be about 25km around. next I had Zentreadi warships atttack it. success! I even made the planet warp out before destruction (seeing that huge jump point form and the planet get sucked in was great, especially as teh zentreadi was STIL firing at it even though it was gone, until teh warp point closed completely. heeh Better luuck next time boys!

  Now my question. we CAN make planets (even in retail) BUT so what? you would need to alter teh sphere to have destructable/targetable points:  like cities, bases, resouces, planetary defenses... right?  

   BTW I scaled down my mooon and called it "Moon Base Alpha" it's about 4km across and looks cool when I made 11 star destroyers pummel it until they killed it 5 mins later (even a small moon has lots of Hit points!

l8tr!
Title: How feasible are PLANET models?
Post by: Lightspeed on January 09, 2004, 12:45:22 pm
Quote
Originally posted by PhReAk
its extended out to 10,000,000km now.


Since when? Weren't there problems with the z-buffer? Does the 12_05 exe have this implemented?
Title: How feasible are PLANET models?
Post by: StratComm on January 09, 2004, 12:51:32 pm
I was under the impression that skyboxes were designed for things like planets, which should always be at the same heading no matter where in the battle zone one is flying.  A far-clipping plane at 200k-300k meters would have sufficed, but if it's been bumped all the way up then that's fine too.  The HT&L limit was too low, but only by about one order of magnitude.

As for targeting cities and such on a planet, one must ask why.  You'd never get that close in a space sim, and if you did you'd throw any attempt to remain in reality right out the window.
Title: How feasible are PLANET models?
Post by: Nico on January 09, 2004, 02:18:29 pm
skyboxes, to me, are meant for nebulas, noty planets.
as for targeting cities, well, they do so in starwars, no? :p
Of course, starwars and realism...
Title: How feasible are PLANET models?
Post by: phreak on January 09, 2004, 02:24:07 pm
Quote
Originally posted by RandomTiger


If its set to high it can degrade the effectiveness of the z-buffer I believe.


only for 16-bit ones.  i put the near plane at 1m to help improve z-accuracy far away.
Title: How feasible are PLANET models?
Post by: Omniscaper on January 11, 2004, 04:25:18 am
Whoa, z-buffers and stuff. I lost you folks a couple of threads back. How many meters should I make the planet model. I use PCS and Truespace. Should I make the goal size in Trusepace or should I multiply it via PCS.

Where should I place the planet in Fred? Because the 12/05 build I'm using is not displaying the planet at all. I'm quite sure I placed myself outside the planet''s radious when creating the scene in Fred.
Title: How feasible are PLANET models?
Post by: karajorma on January 11, 2004, 07:02:15 am
AFAIK the 12-05 build won't show the planet at all. I think Phreak has fixed the problem but no one has released a build with the fix in it yet.
Title: How feasible are PLANET models?
Post by: Gregster2k on January 11, 2004, 01:37:59 pm
:wtf: Phreak HAS released it. SCROLL UP AND LOOK AT HIS SIGNATURE! :wtf:
http://www.swooh.com/peon/phreak/

The "Inferno build" on Phreak's site has the fix. I made a custom test mission with a Planet Earth in it with the FRED included with the "inferno build". IT WORKS --- FRED does show the planet, and its visible ingame using the game exe inferno build as well.

the problem is that the planet from Inferno is so big you have to manually input the coordinates for the planet and set it to at LEAST 10,100,000 units away from the player ship, give or take. (That's for the earth model though, your mileage may vary!)
Title: How feasible are PLANET models?
Post by: karajorma on January 11, 2004, 01:53:15 pm
Preaks sig has been pretty much the same for the last couple of months so there's no way to tell except by clicking on it every couple of hours.

It used to point at an older version that hadn't fixed the problem. Stick around for a few weeks and you'll see why I didn't notice it. If it's not announced most people miss it.

Oh and there is no need to shout. You'll get a bad rep pretty quickly doing that .
Title: How feasible are PLANET models?
Post by: phreak on January 11, 2004, 07:40:01 pm
yea i know i don't update my sig. i should do it more often