Many to Few - What is best for performance?

[BigDaz]

The crash.log's are located in the "GameCore 2.0.4/Editor" folder.

[Steely Dane]

Have you tried other formats instead of lwo? I use obj and 3ds for almost everything and never had a crash like that. Also, does your computer have on board video or a graphics card?

My computer is lower end (AGP video and single core processor) when compared to a typical gaming computer and I can still load about 1/2 million poly scenes and 100,000+ poly objects.

No, I haven't, but I'll give it a whirl. It has dual Nvidia 6800 ultra's in SLI format. Intel 6600 Quadcore cpu.

[Steely Dane]

One more note... the lwo has two formats. I do not know alot about it, but I guess Lightwave 1 is the old format and Lightwave 2 is for "Huge" objects. I use a utility called Accutrans (http://www.micromouse.ca) and it allows you to save / convert various 3d formats / triangles to quads,...., best $20 I every spent.

I've been using Lightwave for a couple years and haven't come across that. When I save out an object, it only allows you to save as one thing, and that's LWO.
Unless you are referring to exporting, in which case I can export an object as LWO 5, and I've only ever done that to get an object into poser, or vice versa. I imagine that would be Lightwave 1, whereas I am using LW 9.6 which would be LW 2?
Are you saying I should try and export using the older version? (might work, never thought of that before either)

I've heard of accutrans. It would be handy, but up till now, I didn't really need it.....

[Steely Dane]

The crash.log's are located in the "GameCore 2.0.4/Editor" folder.

My editor folder :data (fonts, gui, localizer,gameshell,images,objects), docs, scripts, shaders, templates, tutorials, config, gamecore (launcher), gameeditor, keys, splash.

no crash.logs.

[JimB]

Then it does'nt look like it made one,here is a pic of mine 

[Steely Dane]

Then it does'nt look like it made one,here is a pic of mine 

Hmm, yeah looks that way. I wonder wtf is going on there....
I reinstalled  GC, maybe it was broke from Mr. Previous Owner.
I'll let you know....

[Steely Dane]

All right, I tried exporting as an obj. file. Didn't work. Got this instead:

[Steely Dane]

Also tried exporting it as a 3DS file, but you can't export as a 3DS object with more than 65K vertices from lightwave apparently, and that certainly leaves out an object with a half million polys, so I'm calling bull$h!t on loading anything from LW that's over 45k polys, and when I get the CRASH.LOG error it's still not turning up in the editor, so I can't even send it in for a support ticket...LW5 objects are low poly exports so that isn't going to help either.....

ok, I tried one more time, and managed to get an exported obj. file (522 341 poly's), but parts of the object were missing, and lightwave export to obj. file, looks messed up, like the smoothing is over done, wether it's on or off.....

The whole thing is giving me a raging headache so I'm going to bed and maybe try some of the other export options in the morning....I still haven't tried DXF or COLLADA yet.


Right, I give up....it must be a LW issue, so I'm gonna go with Ron's suggestion of Accutrans for stuff bigger than 50 000 poly....it will come in handy anyway since I plan to sell a few models on turbo squid, and want to have more than LWO files for download anyway. Still can't figure out why the crash log isn't showing up though.....

[gekido]

Worst case simply export as multiple subobjects - you probably don't want to have a single mesh with ridiculously large numbers of poly's anyways.

For larger worlds, even ones that I've created as a single project file in max, you'll want to load any duplicated geometry as instances, split objects up into 'zones' or areas of the world so that they can be culled or occluded etc.  If your entire world is one massive million-poly object, the engine can't do any kind of rendering optimizations on the object whatsoever, so whether 10 polys or the whole object are in view, the rendering is going to take the same amount of time.

[Steely Dane]

Yeah, that's what I was doing already. I don't really need to be putting in anything much larger than 100 000 poly anyway, and that doesn't even come up often enough to be such a hassle that I can't load it in as subobjects like you've suggested. Just thought there was something wrong with what I was doing, if you were able to get larger objects in , and I couldn't......

Still don't have a clue as to why my CRASH.LOG doesn't show up where it's supposed to....

Thanks, G.

[pixel_legolas]

are you using normal maps? There is no reason to use high-poly models just because you can Optimising should always be done if possible.

But I guess you already have done alot to the models

[Steely Dane]

I will be, but only to super detail out the models...
If you can use high poly, beautiful models that border on photo realistic, why wouldn't you?
Although I have no beef with someone using maps to detail models, it makes little sense if the engine can handle the poly's.
I really dislike the look of blocky models, with fake detailing, just looks cheap to me...if something is round, it shouldn't have sides, lol (like a gun barrel for instance. Nothing more ridiculous than a 5 sided shotgun barrel).

Affordable hardware has made the issue a moot point as well.
Never really understood why one would make a game that resembles something akin to Unreal Tournament or Doom (model wise) when they first came out.
My machine is almost 2 years old and can handle anything out there, and it's practically obsolete.
If the technology has advanced, so should the art.
If the engine had all of the texturing capability my 3d app has including procedurals, it would be a personal wet dream to me  .....

I suppose if you are marketing to the lowest common denominator, it makes sense.
An argument can be made however, that the majority of avid gamers have decent enough machines that can handle some pretty hardcore stuff.
It seems to me that an "avid gamer" would have something pretty up to date (the ones I know anyway, they scoff at my dual sli cards), and in a practical sense, I'm building a game for gamers 2 years from now, because that's how long it's gonna take, and Crom only knows what kind of hardware is gonna be around by then.

Of course this is all personal preference. I'm capable of building my own highly detailed and accurate models so I do.
When I actually get the normal maps on it, it will be damn near photo realistic, and that's what I'm shooting for.
Unfortunately for the many that have to rely on free models or are more proficient at coding than modelling, this isn't the case, so mapping is gonna be the way to go.

I understand the purpose behind it, but it's just not a personal artistic philosphy of mine.
To be cutting edge you need to push the limits of yourself, your tools, and the recipients of your efforts.
Another consideration is that some little twit is more likely to warez your work than buy it, if they can play it on their POS eMachine...

[Ron]

I do not remember where I originally saw someone post this, but its a monthly survey of gamers hardware. Gives you a idea of whats out there:
http://store.steampowered.com/hwsurvey/

[Steely Dane]

Case in point:
The top three resolutions for monitors are 1280 x1024, 1024 x 768, and 1680 x 1050.
Yet the most common size for uploading images is the completely useless 800 x 600 which ranks second from last out of 16, as far as resolutions go.
I realize that some people use that for uploading pictures because their connection is pathetically slow, but my feeling is that it has more to do with not letting go of old habits.
There is nothing more frustrating than trying to find decent reference pictures on google images, and all I can find is thumbnail sized pixelated jpegs that look like a 3 year old took them .

Another issue is that a great many indie developers are coders and not graphic artists. They tend to use largely outdated or underpowered machines because they don't need that for what they do (I know one who swears up and down that his windows 98 machine is all "he'll" ever need, and he's probably right), and fail to realize that most gamers want a healthy portion of eye candy, and could really care less that you managed to get a sweet traffic system running with only a few hundred lines of code.
That's how big companies get away with releasing virtually the same game every six months.
Sweet new graphics + Same tired old concept = Big money, Vanna!
Sad but true, humans like shiny things...
Who wants to play a game with graphics that date from the late 90's on a $1500-$2000 machine?
You can't make games for an Alienware Area 51 x58 with a 6 year old Dell laptop ?(well I suppose you could if you had a better machine to actually test the game on).
That is just the facts.
It's not really fair to limit your prospective audience to your hardware choices, or lack thereof, is it?
It may ruffle a few feathers, with me saying that, but it's true.

There is no reason to use high-poly models just because you can

On the contrary, sir, there is every reason.
Michealangelo's David wasn't made out of limestone. It could have been, but he chose marble.
Why? I imagine because of it's beauty, it's longevity, and because he could.
All the same reasons I'm making the choice to go with pushed to the limit detail modelling.
Another example closer to the subject would be Shrek. With your line of thinking, beautiful movies like that would never have gotten made, had they said, "Why, when we can just draw it?"......do you see my point?
Advances are made by pushing the envelope, not because you should, but because you can.

According to the list Ron posted, the most popular specs are Windows xp 32 bit, with duo core intel cpus (with quad core slowly rising), 2 GB ram, with Nvidia DX 10 gpus, and a primary display of 1280 x 1024.
That's who I'm building for. Actually I'm building for those that are a little farther up the chain, but like I said, my game won't be ready for at least 18 months, and by then they will have caught up.
The argument that low poly models are the way to go for games is not exactly true anymore.
If you get no signifigant frame rate loss from high poly, why would you go low and sacrifice the opportunity to make a gorgeous looking game?
I'm not trying to be a jerk here, as I know quite a few of the users on here are of opposing thought on this with me, either by philosophy or necessity.
I hope I've not offended anyone, with my opinions on this.