PDA

View Full Version : Lightwave CORE and Game Development.



GandB
02-05-2009, 12:41 PM
Alright; so instead of all the Game-Dev comments/concerns getting buried in the numerous threads about CORE...I thought I'd start one soley for Game Dev thoughts and questions.

I see that NT has decided to go with Collada as a base format, which is a start for Game Development. What about attention not only to popular formats; but the whole pipeline in general? Other than a few glamour shots on the front page, of LW's involvement in shipped titles; I see no discussion of the pipelines and methodology involved?

Does NT want to be a part of the (mentioned it before) LARGE market of Game Development or not? What outside recognition has Newtek received for game development? I see the "not-so-popular around here" team at Lux have been paying attention to us:

http://www.luxology.com/press/release/010807_modo_Frontline.aspx

The slow to inactive activity in the Game-Dev section of the Forums here should tell you that not many people (read "Game Artists) view LW as a product they want in their pipeline. Again; Autodesk is leading in this area, because they realize the extensive market that exists in the Game-Dev world.

As I've said in one of the previous "CORE" threads; does NT intend to continue "glossing" over the potential client base in the Game Art world...or are they going to get more serious, and attempt to catch up in this area.

All views, thoughts and opinions here are welcome.:thumbsup:

-Keith

DonJMyers
02-05-2009, 04:19 PM
It will will take a rewrite and more interoperability for LW to get into gaming production.

It vaguely sounds like you could even use the core as a an opengl gaming engine if you really wanted to. I know somebody came up with an LScript that ran a 3d chess scene that was a functioning game. A robot arm next to the board would move move the LW pieces.

geothefaust
02-05-2009, 04:26 PM
It will will take a rewrite and more interoperability for LW to get into gaming production.

It vaguely sounds like you could even use the core as a an opengl gaming engine if you really wanted to. I know somebody came up with an LScript that ran a 3d chess scene that was a functioning game. A robot arm next to the board would move move the LW pieces.

Apparently you didn't see that it DID get a rewrite, lol!

But I do agree, I think that you could actually use LW CORE as a game engine, with enough programming know how. Pretty cool.

I was shocked to see that they are using collada as the native file format. Which is great, and I think points or hints at NT doing more in that field. Of course, in a few weeks when we have our hands on CORE, we'll know more about it and we can give our feedback too, as the HardCORE members will have direct input on the final product via the dev team. So that's sounding great to me.

GandB
02-05-2009, 06:59 PM
Here's a game being made with the FPSCreator (http://www.fpscreator.com/) game engine:

http://www.youtube.com/watch?v=0qk2770E9vg

Here's a thread on The Game Creator's forums, showing some of the custom media that was created for it:

http://forum.thegamecreators.com/?m=forum_view&t=137567&b=25

Talking with one of the developers; I'm told their art team used Lightwave primarily for all the modeling. The format that FPSC uses is .X.

-Keith

r10k
02-05-2009, 07:06 PM
But I do agree, I think that you could actually use LW CORE as a game engine, with enough programming know how. Pretty cool.

If you were insane, sure.


Here's a game being made with the FPSCreator game engine:

You realise that no one seriously into game development (either indie or otherwise) uses FPSCreator, right?

mav3rick
02-05-2009, 07:12 PM
ahahahahaahahh lol :) good one r10k :)

GandB
02-06-2009, 05:17 AM
You realise that no one seriously into game development (either indie or otherwise) uses FPSCreator, right?
Not too sure if you looked at all that I laid out there; but those guys seem pretty serious to me. Serious enough that they modded the engine in many ways on their own. Serious enough that they have contracts to market the game (as well as a second one they've been cooking up). Jake Cotton (Errant AI) uses FPSC to get his Game Art into play, with minimum fuss:

http://www.blackdracogames.com/forums/index.php?board=61.0

http://forum.thegamecreators.com/?m=forum_view&t=126691&b=24

It's not the tool; it's the developer(s) that make the difference.

As far as using CORE as a game engine; that remains to be seen. TrueSpace 7 has those same abilities; in fact there are a number of people who have made small games with it (though they are not stand-alone; you need TS to play them), such as a 3D minesweeper and pacman.

-Keith

StereoMike
02-06-2009, 05:50 AM
haha, in fact the old lw had it already (but not publically available):PIM :D
On starting the plugin it replaced lw's viewport with a blazing fast 3d engine and you could assign javascript to everything. Hey, I've seen a playable quakelevel inside my lw viewport (with realtime-dof).
So if third parties were able to do this to lw 8.5, imagine what's possible with a much more open platform.

mike

GandB
02-06-2009, 07:44 AM
I remember that; what happened to it? I had asked about it awhile back, but it seemed like the dev just dropped it?

GandB
02-06-2009, 09:58 AM
It was also pointed out to me that Unity has a pipeline for Lightwave (along with.....wait for it....Max). Have NT even looked into this?

Wickster
02-06-2009, 10:04 AM
I'm pretty sure the Game Industry was on NT's mind as well when re-writing LW. I mean it's hard to miss a multi-billion dollar industry that is slowly surpassing the movie industry. I guess all we could really do to be sure is wait and get an answer and an their gamedev outline from NT.

StereoMike
02-06-2009, 10:14 AM
@GandB: Actually you can develop stuff with it when you're a beta tester (Veggie is an example). I got a recent pimtools build some days ago, so there's development, but nothing public (and no indication it will go public). Based on pim is also Loco, a realtime mocap thing. I've seen stuff, where actors wear googles and mocap while they see the CGI scene in realtime (and not a green screen).
The dev has himself never dropped his baby, it just developed into something different (hollywood etc).

mike

GandB
02-06-2009, 10:35 AM
I mean it's hard to miss a multi-billion dollar industry that is slowly surpassing the movie industry.
Apparently it hasn't been so with NT, as I see very little concern in the Lightwave site. What they've presented (on the main site....games...) does little to overcome a game artist's desire to shore up his pipeline with industry standard tools...that will actually get him hired, or cut down on his development time (in an Indy's case).

But, as you said, we'll have to wait a bit to see if they actually address this issue with more than a passing paragraph.

Incidentally, the Indy Game scene is also fast becoming a multi-billion dollar industry as well. You've got teams of any where between 2 or 3 people, to a dozen or more; with an average time of 6 months to a year, before a title is shipped (with proper design docs and systems already in place). This is really where NT could pick up the slack with a pricepoint well below that of Autodesk. An Indy doesn't have to worry about having an industry accepted tool...just whatever gets it done for him (or her...gotta remember the girls :D) in a timely manner, the fewr the steps, the better.

Sounds like PIM will never see the light of day for the masses then.

Off the subject, but I've always wondered...is that you with the cucmber facial?:D

-Keith

geothefaust
02-06-2009, 10:38 AM
If you were insane, sure.

I didn't say otherwise. ;)

Adrian Lopez
02-06-2009, 11:05 AM
Is there any benefit at all to using Lightwave itself as a game engine, other than the novelty factor? After all, it's not as if you'll be allowed to redistribute Lightwave along with your game.

What you really want is an easy way to host your game engine within Lightwave. If CORE is as flexible as Newtek claims it to be, I expect Lightwave CORE to make a great tool for manipulating objects in game.

warmiak
02-06-2009, 11:51 AM
Is there any benefit at all to using Lightwave itself as a game engine, other than the novelty factor? After all, it's not as if you'll be allowed to redistribute Lightwave along with your game.

What you really want is an easy way to host your game engine within Lightwave. If CORE is as flexible as Newtek claims it to be, I expect Lightwave CORE to make a great tool for manipulating objects in game.

Precisely, what matters for most game developers is how easy is to inject and redirect functionality built into a 3d package into whatever output/format is needed for a particular engine/library.

In other words, nobody will use CORE as an game engine but people would gladly incorporate CORE into their pipelines.

Oedo 808
02-06-2009, 12:07 PM
Those who are interested in game development are almost certain to get the basics that LW has so far been lacking, the interest coupled with the suggested accessibility of the Core engine almost guarantees that someone will bring this to the table sooner or later.

What my concern is, is that NewTek are aware of this and will trust to that happening. The thing is that I don't want game development in Core to be a third-party add-on, not even if it is at a low cost, or even free.

I understand that it is impossible to cater for all tastes and that some people are going to be disappointed that Core will not bring what they are looking for specifically, but as has already been alluded to, I wouldn't call game development a niche market. For enthusiasts like me, and I'm sure the same goes for small indie devs, that fact that LightWave has already featured as part of a larger pipeline brings little comfort.

Well, there's still plenty to come from NewTek, let's hope we see something to get our juices going before March is out, there's no need to be calling out that the end of the world is nigh... not just yet anyway.

*edit*

Speaking for myself, I've no interest whatsoever in seeing Core being used to host other engines, I'm more interested in if it can extrapolate and export the animation data to be used in a proprietary format when you've used a myriad of modifiers to get the motion you're looking for, that would be a groovy. I guess what warmiak was saying.

GandB
02-06-2009, 12:17 PM
....yes...not just yet. ;)

warmiak
02-06-2009, 12:22 PM
The thing is that I don't want game development in Core to be a third-party add-on, not even if it is at a low cost, or even free.



Well, there is no such thing as a "standard game development environment".

The potential for CORE lies precisely in ability to extend and incorporate its features into whatever environment people are working in.

Every game engine is different and will require slightly different pipeline and thus people will write their own plugins and extensions - furthermore 90% of these extensions won't even be made public.

GandB
02-06-2009, 12:27 PM
The potential for CORE lies precisely in ability to extend and incorporate its features into whatever environment people are working in.

Yep; just like Max.


Every game engine is different and will require slightly different pipeline
Except that nearly every game engine goes out of it's way to support Max.

jkir
02-06-2009, 12:44 PM
I am also interested in games specific side of things. I am generally don't need all those things 3d software developers prefer to demonstrate as their achievements. I need just a reliable export and several games specific features. Editable vertex normals for example. Whatever revolutionary a soft is it have no sense for me if nevertheless I need to buy 3ds max or Maya or XSI to edit my normals.

Oedo 808
02-06-2009, 02:40 PM
Well, there is no such thing as a "standard game development environment".

But there are features such as solid viewport GLSL shader and normal map support that anyone interested in game development would hope to see. I'm not trying to be a whiney old git, but if I don't mention it and it doesn't come then I've got no one to blame but myself and others of a similar ilk for not making it heard that this is what we want to see from Core.


The potential for CORE lies precisely in ability to extend and incorporate its features into whatever environment people are working in.

Every game engine is different and will require slightly different pipeline and thus people will write their own plugins and extensions - furthermore 90% of these extensions won't even be made public.

If I ever come to do anything decent, I won't be handling plugin development, but I'm sure the guy who is would like to see me upgrade to Core, I'm not hugely interested it what other companies do with their formats, though it would be nice to see more releasing a LW plugin with their SDK, and Core makes this a real possibility. I just hope that the animation tools in Core allow for a simplifed parsing of the data for exporting to a proprietary format. If you think that much is a given then that is good news for me. It's difficult trying to gauge whether a software platform will fulfill your future needs.

Myagi
02-06-2009, 03:27 PM
it would be nice to see more releasing a LW plugin with their SDK, and Core makes this a real possibility. I just hope that the animation tools in Core allow for a simplifed parsing of the data for exporting to a proprietary format. If you think that much is a given then that is good news for me. It's difficult trying to gauge whether a software platform will fulfill your future needs.

I think the reason why the other packages might be more popular is that they had better character animation and a few other things (like the ability to conviniently add meta-objects and properties useful for their game exporters), so they could buy one app and be done with it. As a result of being more popular they also released exporter plugins etc. for what they used. So it's not so much a lack of game support by NT itself. If Core has CA improvements to make it an alternative, more devs will start using it, and plugins will be released, without NT necessrily doing anything specifically for it. (The new node system and extensability through the SDK takes care of the meta-data stuff).

That said, I've always found LW to be great for games, by being a really great modeling tool at a great price. With an easy to use, and more importantly, always open LWO format, making writing an importer very painless and quick. When I had the choice, I worked with LWO, but I've worked on projects using MAX and Maya too, and had to write exporter/tool plugins for both. It's really not that big of a deal, and wouldn't be a problem for anyone serious about it.

What will weigh heavy in the decision for what software to use will be its total abilities/usability (and pricing too) which aren't directly game related, like the mentioned CA tools. "Can we use this app to efficiently create all* the 3D content we need, so we don't have to buy several apps?", if core answers that with "yes" I'm sure devs will come, and with them plugins/support.


(* all that needs to be done in a third party app)

Oedo 808
02-06-2009, 05:11 PM
Thanks for the info Myagi, I know that LightWave would not have been the main choice for development for the reasons you mentioned, but even if the other main packages were not used in development they often still got a plugin that could be used if an SDK were released. I had thought that this was because it was a bit awkward exporting the animation data out of LightWave, I guess I was wrong, which is actually good news for me.

Although if it's true that it was no more difficult to provide a plugin for LightWave than it was for Max or Maya when XSI was the main development platform, then I'm not sure that Core will convince studios to start adding LightWave plugins to their SDK. As you say I guess we'll only see support if Core gets used as the primary package.

Given that LightWave seems to be fairly MoCap friendly, I had thought it was the reasons already mentioned above that made LightWave a not so popular choice for game studios, not really because it's CA tools weren't all that.

It's like being in a bloody maze. :stumped:

Myagi
02-06-2009, 06:08 PM
I had thought that this was because it was a bit awkward exporting the animation data out of LightWave, I guess I was wrong, which is actually good news for me.

For my own project, I have an animation exporter too, so I can export rigs and character anis to a format the game tools can use, and it wasn't rocket science to code either. So when the app choice is made, it's gonna be about which the animator can work better with. I don't have any data but I wouldn't be surprised if Character Studio was a big reason for MAX gaining popularity. Later and still(?) Maya has the rep of being the best animation app among the main players, and I beleive it is or was cheaper than MAX which made that an appealing solution.



Although if it's true that it was no more difficult to provide a plugin for LightWave than it was for Max or Maya when XSI was the main development platform, then I'm not sure that Core will convince studios to start adding LightWave plugins to their SDK.

I was a bit unclear there, but my reasoning was if you get something that can compete with the others, BUT at such a good price/terms, it should be a good incentive to switch to it. Especially devs normally will have more than one seet, the savings aren't trivial. So even if it doesn't surpass them, but it's a viable alternative, then the price tag gives it an edge.



Given that LightWave seems to be fairly MoCap friendly, I had thought it was the reasons already mentioned above that made LightWave a not so popular choice for game studios, not really because it's CA tools weren't all that.

I honestly don't know much about mocap in LW or otherwise, only one of the projects I've worked on used mocap. That was almost 9 years ago, and the animator used XSI, don't rememer much more. Even with mocap, the rig still has to be built, and the mesh skinned/weighted, so you don't get around those tools even when mocaping.


edit: or was it even called XSI back then? It might have been the last version before XSI :)

GandB
02-06-2009, 07:54 PM
I wouldn't be surprised if Character Studio was a big reason for MAX gaining popularity
That is a huge reason for MAX use today.

I believe XSI was just called "Softimage", wasn't it?

Myagi
02-06-2009, 08:04 PM
I believe XSI was just called "Softimage", wasn't it?

yeah, might have been SI 3.9

GandB
02-06-2009, 08:14 PM
Didn't MS buy it up, try to mess with it....give up and toss it out?

Myagi
02-06-2009, 08:20 PM
forgot about that, but a quick look in the wiki confims it. Bought it in 94 and sold it in 98.

Oedo 808
02-06-2009, 08:43 PM
Alright, thanks for the feedback Myagi, I don't want to take this topic too far off beam, so if I've anything else to ask that's a bit peripheral I'll send it in a PM.


For my own project, I have an animation exporter too, so I can export rigs and character anis to a format the game tools can use, and it wasn't rocket science to code either.

This is good to hear, I'm pants at coding and I do have someone to help me but they're not a LW specialist, at least now I know who to call when things go tits up. ;)

I may well end up using your normal map plugin Myagi, but what I am saying is that I am not upgrading to Core just to end up using the same plugin that's been updated for Core, nor will I upgrade if I can get better shader reproduction in Blender, and though I'm not a great Blender fan it is pretty impressive in many areas if you are interested in game development.

However, I thought that there were more areas of concern than the aforementioned that are currently beyond my scope of understanding that hinder LightWave's use in game development, but you make it sound like adding some CA tools to 9.6 is all that is missing to make LightWave the complete package.

Now there are features with Core like the history stack that look great for modeler, but supposing there is no apparent breakthrough with CA, let me ask you this, instead of upgrading to Core, would it not be a better investment for those interested in game development to buy into something like Messiah instead?

What really has me confused is that this thread started off in the direction of "Now with the rebuild LightWave can be a real asset to game development" but seems to have fizzled to "The only reason it's not viable is because it lacks proper CA".

I'm quite confused.

Oh, Softimage was the primary package used by Monolith in Aliens vs Predator 2 and I think also No One Lives Forever, mind you, nine years ago you'd have had to have been on the development team for those.

Myagi
02-06-2009, 10:26 PM
However, I thought that there were more areas of concern than the aforementioned that are currently beyond my scope of understanding that hinder LightWave's use in game development, but you make it sound like adding some CA tools to 9.6 is all that is missing to make LightWave the complete package.

It isn't the only thing, but it is an important one. There's also other smaller things like for example the meta data stuff I briefly mentioned. With meta data I meant like the object user properties thing in MAX (Maya has something similar) where you can write any text you want, can be helpful to add special game properties on an object for the game exporter. A real world example of that is Trespasser (from '98), where they wrote the object scripts in those properties, so they used MAX as a level editor too. With LW it would have been a bit trickier to use it as a level editor too. (With Core those issues are gone, due to the whole node stuff, it's up to par with the other apps there)

All the talk of what a dev would pick is in general, you have a dev that has budgets and deadlines, so they'll look at what app is the most complete one and can do what needs to be done. But in some cases they mix it up, like ID that used LW for plenty of modeling, and Maya for animation in Doom3. When saying that good CA in LW is missing etc., I think that's in terms of that it would help give it mass appeal to be a big player, not that it's useless in its current state.

With regard to the thread title having fizzled out to CA. I'll try to sum up the value of Core in game developement. Assuming it has CA that can compete, you'll have CA, plus the flexibility of the new architecture that allows you to do the little things you can in the others, like for example use it as a level editor depending on the game engine of course. It's not going to be a replacement for UnrealEd if you're working with that engine. Though I remember the now defunct(?) gMax had a demo of using it as a level editor for Quake3 (or whatever version), so if you have a really flexible base to work with there's a lot you can do.

So if it is flexible like it sounds, it makes it easier to integrate more complex game tools in a seamless manor. Maybe interface to the game's material library and shaders for example, so you get WYSIWYG in LW and so on. But again, games were successfully made without all that hooplah, so you don't need it, but makes things easier, nicer, faster, with bigger market appeal.

Dammit, I really tried to keep it short, but it still ends up a whole page ;)

The bottom line is, if LW currently does everything you need then that's great. Heck I still use 8.5 and am happy with it, granted I've only done a handful simpler animations so far, so maybe I'll be pining over character studio and hope Core can come to the rescue at some point. (Allthough I'm planning on getting the 9 upgrade because I liked some of the new features, and I'd get Core thrown into that deal, so it wouldn't be a big decision to jump to Core if I like it)



Oh, Softimage was the primary package used by Monolith in Aliens vs Predator 2 and I think also No One Lives Forever, mind you, nine years ago you'd have had to have been on the development team for those.

I'm afraid I can't take any credit on those, only worked at smaller local devs on the other side of the pond :)

Oedo 808
02-07-2009, 12:11 AM
Dammit, I really tried to keep it short, but it still ends up a whole page ;)

Hey, you don't have to keep it short on my account! Seriously, thanks for all the info. :thumbsup:

I really hope that things like hooking in a material library will prove to be possible, and I don't mean in the crazy convoluted way that some things work with 9.6. With the amount of possibilities that Core currently promises to provide, I'll have to give this HardCORE business some serious consideration, as they say, you have to be in it, to win it.

Anyway, it's 07:10 here and about time I went to bed. :sleeping:

Thanks again for your time. :thumbsup:

Myagi
02-07-2009, 08:59 AM
Of course we still have to wait and see how the CA thing plays out in Core, but for someone that was going to upgrade to 9 either way, this a golden opportunity to get your hands on Core for $100 extra. (I stand corrected, Core comes with the upgrade for free, as the upgrade pricing has been adjusted)

Just so I haven't give the wrong impression above, I'm first and foremost a coder, I just happen to enjoy doing 3D stuff too which helps for my own one man projects or to test/fix things without bothering an artist. The flexibility thing also means that some things need to be coded (like the material thing), I didn't mean to imply it can all be done without a line of code. The way it sounds, a lot can be done with python in Core, so that gives easier access to this flexibility without going hardcore with C/C++.

Oedo 808
02-07-2009, 10:24 AM
No worries, I've done a couple of ultra-basic programs, I know what work goes into the more advanced stuff because too much went into those pants DOS based Q&A. I don't need to be a coding master myself, but in the same way you might not want to bother the animator, I'd like to be able to add unique keyframe events for things like PV animations, and not have to bother the coder everytime I want to try something out.

I wouldn't worry about giving me the wrong idea, I'm quite sure that hooking in a library will be possible and not too difficult for those experienced in that field. One of my concerns was how it will render once it's hooked in, that's all. But the more I read about Core the more it sounds like proper pixel shader reproduction should be viable.

If you plan to sign up to Core that's good news, it's someone else to bat for the needs of game devs.

Cheers.

Frank_Geppert
02-07-2009, 10:49 AM
At Dexsoft (http://www.dexsoft-games.com) we are offering contents for game developers, we already made thousands of models and textures. I do all my work with Lightwave. The reason was the price and the good feature set. Max was too expensive and Truespace not relieable. C4D missed a few features at that time (texture baking, creating normal maps).

I really love to work in LW. But often I feel behind the Max users. This is especially very hard and sad since I meet Max warez users regularly (using illegal copies). I am against this kind of behavior.

But what does Max better in terms of game dev support? They have options to influence vertex normals, smoothing groups. We have to cut the gometry to get hard edges or we have to create additional geometry. Both is not always good in game engines (shadow techniques or polygon budgets in mind).
Max always supported more export formats. But now I am happy with FBX and Collada. But it really worked only in 9.6. All versions before could not export to game engines.
Max has some shader tools, real-time previews.

Another sad thing is the project website at Newtek. At the bottom of this side you will find games. And there is no game for 2008 and only 1 for 2007, 3 for 2006. It became less and less every year:
http://www.newtek.com/lightwave/projects.php
Is there a reason for that?

I really don't want to move. I want to go on with my LW license and I probably have to do so except I pour couple of thousands to Autodesk. So I really hope that the future of LW tools will also aim toward game devs.
Thanks!

jasonwestmas
02-07-2009, 10:57 AM
Well, there is no such thing as a "standard game development environment".

The potential for CORE lies precisely in ability to extend and incorporate its features into whatever environment people are working in.

Every game engine is different and will require slightly different pipeline and thus people will write their own plugins and extensions - furthermore 90% of these extensions won't even be made public.

Now why do many game engine software companies cater to Max more than any other, package? Most likely there is something simple about the code that makes the conversion of assets into the engine easy. I don't think it is simply because it is an old app that's been tweaked and has a history in game dev, there is a deeper reason than that. So if CORE taps into that simplicity or clean path-making for conversion of assets then we will have Core moving more in that Game Tool direction.

For example. In the past .lwo and .mdd format have always had something simple and convenient about it so those formats were used in other applications. .lwo especially in the game industry.

Frank_Geppert
02-07-2009, 01:12 PM
Now why do many game engine software companies cater to Max more than any other, package? Most likely there is something simple about the code that makes the conversion of assets into the engine easy. I don't think it is simply because it is an old app that's been tweaked and has a history in game dev, there is a deeper reason than that. So if CORE taps into that simplicity or clean path-making for conversion of assets then we will have Core moving more in that Game Tool direction.

For example. In the past .lwo and .mdd format have always had something simple and convenient about it so those formats were used in other applications. .lwo especially in the game industry.

Again, some simple features like editing vertex normal are missing in LW and are big show stoppers for many game devs today. So you have little or no control over smoothing / hard edges.

Myagi
02-07-2009, 01:27 PM
So you have little or no control over smoothing / hard edges.

It's not as simple/convinient, but I wouldn't say no control. Leaving verts unmerged where you want hard edges works, the importer or exporter can auto-merge them after having calculated smoothed normals (so you still have a solid/water-tight mesh if needed). In the end the verts that don't share smoothing will be split somewhere in the pipeline, because one vert would have multiple normals.

Frank_Geppert
02-07-2009, 01:38 PM
I understand what you wrote but I don't have such importer available for all these engines we are providing models to.

I also think it is better to have a good model right from the beginning, right out of the 3d-application instead of creating tons of adapted importers with workarounds to the mentioned problems.

I only want to work, I dont have enough resources to code my own tools, I lose money then.

GandB
02-07-2009, 07:45 PM
Now you see my conundrum with the LW vs. Max debate, Frank. Time will tell, I suppose.

Frank_Geppert
02-08-2009, 12:16 AM
It's not as simple/convinient, but I wouldn't say no control. Leaving verts unmerged where you want hard edges works, the importer or exporter can auto-merge them after having calculated smoothed normals (so you still have a solid/water-tight mesh if needed). In the end the verts that don't share smoothing will be split somewhere in the pipeline, because one vert would have multiple normals.

I thought about this over night. I think there might be a more flexible solution compared to an importer. My idea is like this:

We create the mesh as usual in Lightwave. We break edges to control smoothing and hard edges. Then we export to Collada.
A new tool reads the Collada file (Collada is simple XML-like text format), calculates the vertex normals and merges the vertices sharing the same position, stores all this info in the Collada file and the import should be fine with most modern engine (C4 like an example).

What do you think. Can you program this? I am willing to pay. Maybe some others are also interested. We could share the costs.

Regards,
Frank

particledecay
02-08-2009, 02:11 AM
Vertex optimization happens as part of Unity's import process so I don't need tri-stripping and can use LW -> FBX happily. Does anyone's pipeline still require manual optimization?

doimus
02-08-2009, 03:20 AM
Good thing is you're having a LW vs Max debate at all, because just some time ago (before v9) that discussion would be pointless. Max would win in any possible way. Now, the fact that debate exists, proves we're getting somewhere with LW.:)

I look at it all from the ArchViz perspective, which was the main reason I got LW in the first place. Back in the LW7.5 days we didn't have proper antialiasing, GI, not to mention GI caching, but then came FPrime, new render engine, LWCAD, and now, Lightwave 9.6 is one hell of archviz beast.

And that was all on the old, "less-open" core. Imagine the possibilities of the new rewrite.
I think that CORE development cycle will be all about modelling, CA and cooperation with other packages. Just what the gamedev community needs.

Frank_Geppert
02-08-2009, 05:52 AM
Vertex optimization happens as part of Unity's import process so I don't need tri-stripping and can use LW -> FBX happily. Does anyone's pipeline still require manual optimization?

This is not a problem of optimization of normals. Most engines do this. They can calculate new normals if not existent in your Colldada / FBX files, smooth them, whatever. But this exactly is the problem. Sometimes you dont want to have smoothed normals. Sometimes you have a rim, a hard edge, you want to have special shading, lighting. Then you need control to not get smoothed normals.
And this control can only happen in your modeling software. You can do that in Max and Maya but not in LW. I dont know about the other packs in this special matter.

Myagi
02-08-2009, 08:21 AM
I thought about this over night. I think there might be a more flexible solution compared to an importer. My idea is like this:

We create the mesh as usual in Lightwave. We break edges to control smoothing and hard edges. Then we export to Collada.
A new tool reads the Collada file (Collada is simple XML-like text format), calculates the vertex normals and merges the vertices sharing the same position, stores all this info in the Collada file and the import should be fine with most modern engine (C4 like an example).

What do you think. Can you program this? I am willing to pay. Maybe some others are also interested. We could share the costs.

I'll have a look at how the exported data and the tools look, if it's something I can fix somewhat quickly I'll just put up some free tool or plugin.

Sande
02-08-2009, 08:36 AM
Regarding game development, the Core already looks promising in many ways.

Fully open SDK, using open industry standards in scripting, gui and fileformats and the whole concept of Core as a development platform is very important for us working in the game industry. Well, actually it is important to everyone, no matter what field of CG you may represent. :)

While those may not, at the first glance, be that useful for hobbyists out there, this is often vital for bigger devteams and projects. Hopefully LightWave will now start to regain game industry support and when it does, it is sure that popular game engines which may currently be lacking in their LW-support are to follow. This may not happen overnight, but at least the very basic building blocks for popularity amongst developers seem to be there.

Other important things are:

1. Solid import/export regarding other packages.
LightWave will not become a industry standard game development platform instantly, maybe never, so easy and seamless integration to new and already existing pipelines should be the main priority. This looks promising from what we already know and Collada is a huge step to right direction.
2. Proper editing of vertex normals.
Like for example in Max (the industry standard in gamedev) - hacking the current smoothing-as-a-surface-property-implementation probably just doesn't cut it.
3. Normalmaps in viewports.
Simply a must have - solid, fast, viewport rendering in general is very important for everyone. Naturally the basic things like alphatextures should also work without messing the drawing order (like they currently do in 9.6)...
4. Manipulators.
Again like in Max, Maya, Modo - well, basicly everything else out there. Those are not just useful in modeling - those also help people who have previous experience with other 3D-software to learn LightWave. There is no reason why manipulators couldn't be an option, so you could turn those off if you prefer to work without.

Of course, besides those already mentioned and CA, Core should also be very, very good at modeling AND TEXTURING. Currently Modo does a lot of things very nicely in that department - I wouldn't mind looking for some reference from that direction (no need to reinvent the wheel everytime, but instead of blatantly copying, we should also try to improve.)

UV mapping is quite embarrassing currently in LightWave - which tells a lot how much effort has previously been put to support game development with LightWave.

Hopefully that will now change, and like said before, it does look promising... :)

Frank_Geppert
02-08-2009, 08:44 AM
I'll have a look at how the exported data and the tools look, if it's something I can fix somewhat quickly I'll just put up some free tool or plugin.

Great. Thanks! I hope you will be successful.

GandB
02-08-2009, 06:57 PM
I second that one. :)

Frank_Geppert
02-09-2009, 12:32 AM
I just would like to underline again how great such a plugin will be. As a kind of Collada post processing tool it can help every user in need of Collada transfers, not only Lightwave, Modo, Truespace whatever.

I think calculating the normals is not the biggest issue. Melting the vertices will also be quite simple.
But I have no idea how to find all the adjacent polygons in the data set.
I wish I could help.

Myagi
02-09-2009, 08:11 AM
I did some research yesterday. There are a few options and issues with this.

The biggest issue is that there is no information exported about the smoothing angle in the materials. I would've assumed that since no normals are exported, they would save some meta-information in the materials so you know if it's smoothed and what angle.

That leaves 2 or 3 options
- assume all materials to be smoothed, and assume a fixed angle 89.53 or whatever
- rewrite an exporter plugin from scratch that exports normals too
(- not so user-friendly option to have a conversion tool read the dae and lwo file so it can access material properties to get smoothing settings. will probably be less appropriate for collada files from layout which I assume can contain multiple lwo's?)

If we, for a moment, assume we somehow know what smoothing angle to use, I would suggest that the free "Collada Rifenery" tool (found on collada.org) would be the best choice. It's a "node" based tool where you can add any combination of "filters" to post process collada files (everything from gpu cache optimization to tri stripping). I could add a generate normals plugin for that.

edit: I should add one detail, I used 9.5. Don't know if they improved it it 9.6. 9.5 is horribly broken, I can only export a single collada file without restarting LW, if I export twice the second time contains no geometry at all.

warmiak
02-09-2009, 09:17 AM
I am curious , why wouldn't you want to export normals in the first place ?


Isn't an application like Lightwave , with all its power in terms of geometry manipulation , the best tool for the job , much better than custom hacked engine tools ?

Myagi
02-09-2009, 09:20 AM
it's not about not wanting to, it's about that lighwtave's exporter doesn't do it (unless it changed for 9.6)

IMHO the biggest mistake by NT here is not the lack of support per se, it's that they don't understand that the best way to support is to allow people to help them selves, which in this instance should be to include the collada exporter source they have in the SDK. If something is missing, or you need some special feature, you can within minutes fix it yourself and be done with it.

GandB
02-09-2009, 10:48 PM
I just hope NT is listening and actually comes on this thread to answer some of the questions and concerns here....or at least let us know whether or not they support game development using Lightwave??

Frank_Geppert
02-10-2009, 02:03 AM
Myagi, thanks for looking into this!

The Collada exporter of LW 9.6 is much better than 9.5. It works fine now with animation data and is more reliable. But I have no idea about the details.

I am not sure about your issue with smoothing angle. I think most engines do not use several smoothing angles. They have some inbuilt shading algorithm and that is about it. So maybe it is more important to just have some well distributed normals on top of the geometry. I just mean, they should be smoothed in a certain area where all polygons are connected to each other but not taking disconnected into account. After melting the vertices it still will have the area-smoothing.

Maybe my idea is a bit more amateurish. But this makes a bit sense to me right now.

Frank_Geppert
02-10-2009, 03:18 AM
I just got another idea: If it is too hard to recognize the smoothing areas via broken edges - maybe it is more comfortable to program it with the help of weight maps? We could use weight maps to create certain smoothing areas then.

But the first version would produce a better workflow. We could see the shading better in the LW editor, if we brake the edges.

Chrizto
02-10-2009, 03:32 AM
I'm pretty sure the Game Industry was on NT's mind as well when re-writing LW. I mean it's hard to miss a multi-billion dollar industry that is slowly surpassing the movie industry. I guess all we could really do to be sure is wait and get an answer and an their gamedev outline from NT.

slowly surpassing?
In regards to what we're talking about in here (3D) the game industry has left Hollywood behind for some time ago...

The question is how many horses do we need to drag this chariot?

I think the game industry is pretty much bought by Autodesk. Especially after the XSI acquisition. Many big players have switched to XSI.
Add a little ZBrush to the mixture, and you have something close to an "industry standard" these days.

Why compete? Be creative, find new uses for the Core, or give it a go for gamedev, its up to you.

Personally I don't like dead fish, so I tend to swim in the opposite direction.

Frank_Geppert
02-10-2009, 03:47 AM
I think the game industry is pretty much bought by Autodesk. Especially after the XSI acquisition. Many big players have switched to XSI.
Add a little ZBrush to the mixture, and you have something close to an "industry standard" these days.

Why compete? Be creative, find new uses for the Core, or give it a go for gamedev, its up to you.

Personally I don't like dead fish, so I tend to swim in the opposite direction.

Yes, swimming in the opposite direction makes sense sometimes. But I can give you reasons why supporting game devs might help in the future:
More and more people are interested in making games, small studios, indie studios and hobbyist. Even if not all of them finish their work, they need a modeling tool with some render options (baking light maps, normal maps) plus animation.
Autodesk delivers to big studios. The huge amount of smaller developers cannot afford it. The kids use free warez copies because of this.

An affordable tool might change this situation.

If you watch the game engine market then you see what I am talking about. There are more and more tools and they become more and more easy to use (Unity3d, Shiva, DX-Studio, Esperient) just like the 3d apps you know. Many artists just want to see their assets in real-time, want to walk through their own scenes.

This is a growing market and it is just the beginning. And Autodesk tools are just too expensive to them.

But on the other hand there are free tools like Wings3d, Blender and Truespace. So LW gamers edition has to be better. I personally think that LW 9.6 is better, that is why I am using it but I still miss some important features like you can see above.

Myagi
02-10-2009, 08:21 AM
Myagi, thanks for looking into this!

The Collada exporter of LW 9.6 is much better than 9.5. It works fine now with animation data and is more reliable. But I have no idea about the details.

I am not sure about your issue with smoothing angle. I think most engines do not use several smoothing angles. They have some inbuilt shading algorithm and that is about it. So maybe it is more important to just have some well distributed normals on top of the geometry. I just mean, they should be smoothed in a certain area where all polygons are connected to each other but not taking disconnected into account. After melting the vertices it still will have the area-smoothing.

Maybe my idea is a bit more amateurish. But this makes a bit sense to me right now.


The actual smoothing/normal calculation is not much of a problem :) The reason I'm concerned about the material smoothing settings is that I assume you want WYSIWYG results. I mean if you look at the object in your LW viewport, you'd want it to be smoothed the same way in the exported data.

For example you have an object, using 3 materials. One material has no smoothing at all, one has an angle of 30 and the last one an angle of 89. In order for the exported file to look identically smoothed, I need to know the smoothing setting of all 3 materials.

I must say I'm perplexed what NT thought when they made the collada exporter. They skip normals (I tried 9.6 yesterday too), but also skip all smoothing information, so there's absolutely no way to get your exported object to look like it does in LW.

The only simlpe option was the one I mentioned previously, that is to enable smoothing on _all_ materials used (with the default 89.53 angle). Then I could write the normal generator and apply that smoothing angle on everything. An ugly solution.

GandB
02-10-2009, 08:48 AM
small studios, indie studios and hobbyist
That's what I'm pressing for here. Indy Game Dev is a niche that LW could dominate in. Couple that with the ability to make breathtaking cutscenes with high-poly work and great camera action...and you've got a winner.

Everytime one of us realizes our game creation and it goes to market...it would be something else NT could showcase on the main site here.:thumbsup:

Chrizto
02-10-2009, 09:04 AM
Totally agree to that. It could be a niche to consider investing in.

The framework is there, so it could be developed within Core.
As for 9.6, I don't think it is possible.

Frank_Geppert
02-11-2009, 12:31 AM
The framework is there, so it could be developed within Core.
As for 9.6, I don't think it is possible.

Why do you think so? LW 9.6 is fine. I already use it as my main tool to create game contents. The only thing I am really missing is the vertex-normal export we are talking about. If I need sculpting and 3d-painting I can use external tools (Modo, Bodypaint, ZBrush or 3d-Coat).



The only simlpe option was the one I mentioned previously, that is to enable smoothing on _all_ materials used (with the default 89.53 angle). Then I could write the normal generator and apply that smoothing angle on everything. An ugly solution.

I can understand your concern. If this would be possible then we had a perfect replacement for Max's smoothing groups.
But really, this is not ugly for me. It is prefectly fine. Up to now I dont need several smoothing angles. Most models have only one material, one texture and one uv-map. So I really need only the re-alignment of normals. And I am sure most game devs just agree with me here.

Thanks again for your great help!

jasonwestmas
02-11-2009, 08:44 AM
Why do you think so? LW 9.6 is fine. I already use it as my main tool to create game contents. The only thing I am really missing is the vertex-normal export we are talking about. If I need sculpting and 3d-painting I can use external tools (Modo, Bodypaint, ZBrush or 3d-Coat).




I can understand your concern. If this would be possible then we had a perfect replacement for Max's smoothing groups.
But really, this is not ugly for me. It is prefectly fine. Up to now I dont need several smoothing angles. Most models have only one material, one texture and one uv-map. So I really need only the re-alignment of normals. And I am sure most game devs just agree with me here.



LW works great for a lot of static game asset stuff. It is also possible to get great character stuff out of LW too, it just may take a little longer with characters because of several reasons. Smoothing angles for polys can be transfered over pretty easily from what I hear.

Frank_Geppert
02-11-2009, 12:46 PM
Smoothing angles for polys can be transfered over pretty easily from what I hear.

So what did you hear exactly? I would like to know how this can be done.

jasonwestmas
02-11-2009, 02:38 PM
If my memory serves me correctly, Deep Exploration might preserve smoothing angles from .LWO objects and you can then export to DX or other animation formats. Various engines, I've heared, import .lwo objects with the smoothing angles too. But yeah this isn't always guaranteed. There are are also game engine tool kits similar to Granny 3D that are able to properly deal with LW files, but they aren't cheap!

jasonwestmas
02-11-2009, 02:49 PM
Something else came to mind. Did any of you try to export to collada and import that? I would think there should be a script or something that would be able to export smoothing angles to collada format. Collada is know for being very flexible.

Myagi
02-11-2009, 03:09 PM
The above post(s) just made me think. Frank have you tried importing an LWO to blender and then export to collada? Maybe that will export good normals?

I would've tried but everytime I have to touch blender's UI I get an urge to throw my computer out the window so I don't have it installed ;)


@jason: The issue is lightwave's collada exporter. If you open a collada file it exports, you'll see it's completely devoid of normals and smoothing information of any kind. So if I wanted to code a tool that take such an exported file, to generate normals, I would not be able to re-create the normals exactly like in LW. (Unless I did some assumption that smoothing always has the same setting). The other option would be to write a new collada exporter from scratch, which is more work.

jasonwestmas
02-11-2009, 04:45 PM
The above post(s) just made me think. Frank have you tried importing an LWO to blender and then export to collada? Maybe that will export good normals?

I would've tried but everytime I have to touch blender's UI I get an urge to throw my computer out the window so I don't have it installed ;)


@jason: The issue is lightwave's collada exporter. If you open a collada file it exports, you'll see it's completely devoid of normals and smoothing information of any kind. So if I wanted to code a tool that take such an exported file, to generate normals, I would not be able to re-create the normals exactly like in LW. (Unless I did some assumption that smoothing always has the same setting). The other option would be to write a new collada exporter from scratch, which is more work.


Hmm, that's something Deep Purple could do I would imagine, he was the dev that was working on that exporter. Other than that, it would take another plugin or another program to get the normals just right.

Myagi
02-11-2009, 04:56 PM
yeah, this "other program" is what I offered to do, if it wouldn't take too much time (since I don't use collada myself, and have no intent to do so anytime soon). A completely new collada exporter would be a little outside the scope, even though that would be the best solution. I don't see the source of the current one being released.

It hasn't really sunk in yet that they have an exporter that doesn't export normals nor information to accurately reconstruct them. Sorta like selling a car without wheels ;)

Tzan
02-11-2009, 05:07 PM
Vrooom!

Years ago I had to model/UV in LW >export to VMRL >import to Max >merge points, make smoothing groups. I guess things havent changed much.

jasonwestmas
02-11-2009, 05:13 PM
It always comes to using max for games, don't it :P

Myagi
02-11-2009, 05:17 PM
To be fair, it's the collada exporter in particular. The OBJ exporter for example does generate normals.

jasonwestmas
02-11-2009, 05:24 PM
To be fair, it's the collada exporter in particular. The OBJ exporter for example does generate normals.

True, most game engines are application specific anyway and create their own exporters that create the assets from within the animation package (Usually Max or Maya). Sooo .dae and .obj are not really an issue anyway in my experience seeing as the exporter is pre-made. Even little ol Torque does this. So it really depends on the tools we are using.

Oedo 808
02-11-2009, 06:34 PM
Hi chaps, I've only taken static models through Blender before, I wonder could someone advise me on a good method of getting animations over to Blender?

Cheers.

jasonwestmas
02-11-2009, 06:57 PM
Is there a way to use .mdd in blender? What about DirectX or collada? I can't really think of any other formats that might work atm.

Frank_Geppert
02-12-2009, 12:43 AM
The LW-blender-game pipeline is not what I would prefer. It might work with obj, lwo but then only with static models. So maybe it is a workaround for a few models then.

Collada is the right format now. I checked it. The latest version exports animations well into game engines (e.g. C4, Shiva) and the other option (FBX) transfers to some more engines (e.g. Unity, GameStudio).
Lightwave is a great tool and offers all what most game devs need. So why should I give it up and switch to another tool only because of an export issue? It would be really sad.

Some of you mentioned Max and to be honest: I could not explain me and my family why I have to buy a software for something around $5,000 (yes, it is that expensive in Europe). I even have problems to explain why it could make sense to invest in Core, but Max? Then I probably better go with Maya. But I would do everything to not become an Autodesk customer. I am even willing to pay for such an exporter.
I lately thought about becoming a HardCore member only because of this, only to push Newtek to improve their Collada assets for game devs and to give feedback to the tool-set. But I am still not sure about it. It will take time, patience and money. I indeed could just get another tool then.

Myagi: I really think we can live without the smoothing angles. If you absolutely need them then I would like to type this info in an edit field, so your tool knows it. Then everything gets the same smoothing angle, but I can decide if it will be 40 or 60 degrees.
The smart re-alignment of the normals and the following melting of the vertices is everything we need. And as I told, this will be good for other Collada files as well.

doimus
02-12-2009, 04:44 AM
I lately thought about becoming a HardCore member only because of this, only to push Newtek to improve their Collada assets for game devs and to give feedback to the tool-set.


Well, with collada replacing lwo as the MAIN file format for LW Core, I guess they are on the right track.:)

MentalFish
02-12-2009, 04:58 AM
As a LightWave and Unity user I have high hopes for CORE being a step in the right direction for a better workflow when making games.

Frank_Geppert
02-12-2009, 05:20 AM
Well, with collada replacing lwo as the MAIN file format for LW Core, I guess they are on the right track.:)

Yes. But Collada is a very open format, simple text information, you can use some of the data structures, you can also leave them empty. Newtek left a lot of the slots empty like we realized here (vertex normals, smoothing angles as an example).
Collada can save animations, shaders or even physics nodes. All this can help to transfer informations to game engines. But it just has to be supported.

Having Collada does not automatically make LW a game creation tool but a well-thought feature rich implementation of the Collada format will be fantastic. We have to judge this later on.

Oedo 808
02-12-2009, 05:46 AM
Is there a way to use .mdd in blender? What about DirectX or collada? I can't really think of any other formats that might work atm.

I'm not too sure on those, I have played about with bit Blender, although it has been under a bit off duress (how could it not be with that UI :D). But the more time I spend with it the higher my opinion of it becomes, I'll pass those formats on and see if they'll be viable, thanks.


The LW-blender-game pipeline is not what I would prefer.

I can understand that, but while my coding skills still leave a lot to be desired, I'll have to go with what those who I'm working with prefer. Though I can't really comment on the abilities of Blender, I can say that the short time I've spent with it in recent weeks have made me decide that it is a tool I want to become more familiar with.

Myagi
02-12-2009, 10:12 AM
Myagi: I really think we can live without the smoothing angles. If you absolutely need them then I would like to type this info in an edit field, so your tool knows it. Then everything gets the same smoothing angle, but I can decide if it will be 40 or 60 degrees.
The smart re-alignment of the normals and the following melting of the vertices is everything we need. And as I told, this will be good for other Collada files as well.

I'll try to make a plugin for Collada Refinery, using a fixed smoothing angle, and see how that goes.

I could maybe add the option to specify smoothing in the material names. For example if you don't want default smoothing on a material, the name could end with a special tag like [email protected] The tool could strip away those tags during conversion.

Frank_Geppert
02-13-2009, 01:12 AM
Great, thanks!

doimus
02-13-2009, 06:10 AM
Yes. But Collada is a very open format, simple text information, you can use some of the data structures, you can also leave them empty. Newtek left a lot of the slots empty like we realized here (vertex normals, smoothing angles as an example).

Then we just have to bugger them into implementing those features into LW. They'll do that since they're really nice guys, and they don't have other investors or shareholders to please other than their users. Which makes them even nicer.:thumbsup:

Frank_Geppert
02-13-2009, 06:17 AM
doimus: I absolutely agree with you. But how can I do that? Do I have to pay $400 now to become a hardCore member only to give feedback to the new development direction? Is there a possibility to suggest features as a 9.6 user? I am a bit confused about this topic.

shrox
02-13-2009, 06:22 AM
I have used Lightwave for game for 10 years now. The first I did was "Streets of SimCity" for Maxis. The last was "10six" from Sega. Now I am doing movies, so there are those that do use Lightwave for literially everything.

doimus
02-13-2009, 07:12 AM
doimus: I absolutely agree with you. But how can I do that? Do I have to pay $400 now to become a hardCore member only to give feedback to the new development direction? Is there a possibility to suggest features as a 9.6 user? I am a bit confused about this topic.

This forum is a good place to suggest features, there's even a subforum for that.
Just look at the "LW on Linux" thread (http://www.newtek.com/forums/showthread.php?t=56975) - it seemed as a rather pointless idea when it was started, and we have linux version today. Same could be said for instancing, GI caching, all stuff that was requested here and brought into LW sooner or later.
For example, you don't see many people complaining that LW lacks in the architectural visualization department these days. But three years ago there were nothing but grumpy old arch visualizers here, ranting about render engine, slow GI and whatnot.

As I said, we are the "shareholders" of Newtek. It's in their interest to keep us happy. We just have to show some good will and persistence at the same time.

In the meantime, as shrox said, everything can be done, you just have to find a proper shortcut. That's the erm... fun part of using LW.:D

Frank_Geppert
02-13-2009, 07:16 AM
Thanks doimus. I just thought posting a feature request here gets lost. Look how much irrelevant information grows in these threads, sometimes almost spam. I cant imagine that Newtek members are reading it. It would cost too much time. One thread (about Core) has about 600 posts, probably 99 percent do not contain any useful information (my posts included).

There must be another option, maybe I better find an email. I will search for something like that. Thanks again.

GandB
02-13-2009, 08:11 AM
That's why I started this thread soley about LW and Game Development. We haven't heard anything from the LW guys yet; hopefully this issue doesn't get glossed over again.

The Dommo
02-13-2009, 08:41 AM
I hope there will be some tools.

If not for anything commercial, I would like to be able to make some games with my brother who is a coder at Frontier here in the UK, coding XBox306 and PS3 games. There are things we would like to do together, and maybe sell commercially after that.

Black Tengu
02-16-2009, 12:59 AM
I'm a big noob, so bear with me. I read on this thread that the Lightwave Collada exporter wouldn't export normals, but when I export a Collada file from Lightwave and open it up in Blender I do see the normals? What am I missing?

Frank_Geppert
02-16-2009, 01:15 AM
What kind of normals are you seeing in Blender? Normals in the middle of the polygons or vertex normals?

I am not sure bout Blender but I know from some game engines that they can calculate missing data like tangents, normals and more.

Black Tengu
02-16-2009, 11:10 PM
What kind of normals are you seeing in Blender? Normals in the middle of the polygons or vertex normals?

I am not sure bout Blender but I know from some game engines that they can calculate missing data like tangents, normals and more.

They are surface normals.

Silkrooster
02-17-2009, 12:12 AM
I have used Lightwave for game for 10 years now. The first I did was "Streets of SimCity" for Maxis. The last was "10six" from Sega. Now I am doing movies, so there are those that do use Lightwave for literially everything.

Hey thats cool. I still have [that game] here, though it doesn't play well on an XP. Funny part is I didn't have a computer with enough horsepower until I had XP.

Edit: [that game] equals streets of simcity. :D

Frank_Geppert
02-17-2009, 01:16 AM
They are surface normals.

Surface and vertex normals are different things. Vertex normals influence shading (soft and hard edges). Surface normals tell in what direction the polygon faces (Remember these flipped polygons?).

Myagi
02-17-2009, 08:34 AM
To add some technical detail. You calculate the vertex normals by averaging adjacent surface normals together. But which surface normals are averaged with which is controlled by the smoothing properties. In Max the smooth groups control that, in LW it's the smoothing angle. So the exporter needs to calculate the normals or at the very least export the smoothing information needed to reconstruct the vertex normals (identically to the app the object was made) yourself.

GandB
02-17-2009, 09:15 AM
Great News!

Torque is working towards making import of Collada meshes a reality; thereby eliminating the archaic .DTS format as a neccessity! Take a look here:

http://www.garagegames.com/community/blogs/view/16385

warmiak
02-18-2009, 09:35 AM
The LW-blender-game pipeline is not what I would prefer. It might work with obj, lwo but then only with static models. So maybe it is a workaround for a few models then.

Collada is the right format now. I checked it. The latest version exports animations well into game engines (e.g. C4, Shiva) and the other option (FBX) transfers to some more engines (e.g. Unity, GameStudio).
Lightwave is a great tool and offers all what most game devs need. So why should I give it up and switch to another tool only because of an export issue? It would be really sad.

Some of you mentioned Max and to be honest: I could not explain me and my family why I have to buy a software for something around $5,000 (yes, it is that expensive in Europe). I even have problems to explain why it could make sense to invest in Core, but Max? Then I probably better go with Maya. But I would do everything to not become an Autodesk customer. I am even willing to pay for such an exporter.
I lately thought about becoming a HardCore member only because of this, only to push Newtek to improve their Collada assets for game devs and to give feedback to the tool-set. But I am still not sure about it. It will take time, patience and money. I indeed could just get another tool then.

Myagi: I really think we can live without the smoothing angles. If you absolutely need them then I would like to type this info in an edit field, so your tool knows it. Then everything gets the same smoothing angle, but I can decide if it will be 40 or 60 degrees.
The smart re-alignment of the normals and the following melting of the vertices is everything we need. And as I told, this will be good for other Collada files as well.


In all my exporters I simply create additional duplicated vertices for "hard" edges - there is no other way, that's how 3d hardware works.

Myagi
02-18-2009, 10:21 AM
Sure, in the end they have to be split, but I could imagine that engines who need/prefer closed manifolds, for example edge data for stencil shadows or when same geometry is used for poly based collision. They might want the unsplit geometry data for preprocessing those things, before they create the split data used by the GPU.

Not a big problem if you make your own exporters or importers. But if you make generic output not aimed at a specific target environment, or you're working with third party toolchains (modding and whatnot), it's probably nicer to have the "best possible" data.

warmiak
02-18-2009, 12:23 PM
Sure, in the end they have to be split, but I could imagine that engines who need/prefer closed manifolds, for example edge data for stencil shadows or when same geometry is used for poly based collision. They might want the unsplit geometry data for preprocessing those things, before they create the split data used by the GPU.

Not a big problem if you make your own exporters or importers. But if you make generic output not aimed at a specific target environment, or you're working with third party toolchains (modding and whatnot), it's probably nicer to have the "best possible" data.

I guess .... on the other hand, it would be just as easy to create unsplit geometry out of vertices/indices as it would be to split the data :-)

Myagi
02-18-2009, 02:25 PM
Yeah no doubt :) , but if you don't have control over the importer (like in the mentioned examples) the only thing you can affect is the data you generate and attempt to make it as compatible as possible.