Rf and Jet3d
-
- Posts: 51
- Joined: Fri Jul 08, 2005 1:31 pm
- Location: India
Rf and Jet3d
Jet3d the 'younger brother of genesis" has gone a long way from genesis 3d. It already features a directx 9 renderer, and shaders. I thinkk its time for an upgrade from the genesis 3d engine. I would love to see Rf integrated into irrhilcht or OGRE someday...
That's what gekido has done, he has moved away to his own engine, Beyond Virtual, there are a couple of post on this board regarding BV do a search and see what progress has been madeI thinkk its time for an upgrade from the genesis 3d engine.
Genesis is a bit old, but it still is a good engine, Wxb1 is working on Harware T&L wich will update Genesis somewhat, from there RF can grow into a more modern engine..
but it's not a matter of waving a magic wand and Hey presto, RF is now OGRE.....RF is a game-shell of Genesis3d, RF and genesis3d are one, everything we use (entities, drivers, dot3 stuff etc...) if we change engine's it would mean re-writing everything from scratch,
and i don't wanna see everything disappear,
Vexator, an old member of this board attempted to make a game-shell using OGRE, he made a fair bit off progress but i haven't seen him for years and are unsure if he is still continuing it or not
...there's actually a link on this board linking you to his website quest posted the link. do a search for vexator battlefeild.
In all due respect i think it would be an impossible task to undertake, be patient and see what gekido has up his slevee.
-
- Posts: 51
- Joined: Fri Jul 08, 2005 1:31 pm
- Location: India
umm..
in THEORY, the original idea for jet3d was that you could do a port from genesis fairly easily, however since the two projects have diverged in their own directions without a central planning source (ie the role that Eclipse originally played), i doubt that this would be an easy task.
there is a gameshell project under development using jet3d, it could be used as a starting point to begin development of an rf-style non-programmers shell, but i'm not sure how much code from rf would be useable...
there is a gameshell project under development using jet3d, it could be used as a starting point to begin development of an rf-style non-programmers shell, but i'm not sure how much code from rf would be useable...
-
- Posts: 51
- Joined: Fri Jul 08, 2005 1:31 pm
- Location: India
thats true, but how long has it taken us to get this far? by the time we finish, ogre would most likely be just as out of date as genesis3d is now. sad but from what im predicting true.
my thought is that we develop something similar to the video driver system of rf/genesis - instead of making direct calls to d3d or opengl, it makes calls to a driver, which automatically routes it to the correct api according to user-specified settings. it could be, very painstakingly, set up the same way for the rf shell, rf makes calls to a "wrapper", ie. moveobject, and the wrapper would handle the interface between rf and the engine used. although this would take a long time to develop, it would make rf basically futureproof and non-enginespecific; want to make a irrlicht rf? just write an "irrlicht driver". i believe this is how it would work.
this would take even longer than writing rf from scratch on a new engine, but at least you would know you would never have to re-write rf again.
my thought is that we develop something similar to the video driver system of rf/genesis - instead of making direct calls to d3d or opengl, it makes calls to a driver, which automatically routes it to the correct api according to user-specified settings. it could be, very painstakingly, set up the same way for the rf shell, rf makes calls to a "wrapper", ie. moveobject, and the wrapper would handle the interface between rf and the engine used. although this would take a long time to develop, it would make rf basically futureproof and non-enginespecific; want to make a irrlicht rf? just write an "irrlicht driver". i believe this is how it would work.
this would take even longer than writing rf from scratch on a new engine, but at least you would know you would never have to re-write rf again.
art != graphics engine
the ratio of programmers to 'designers & artists' in the rf community has always been about 1000 to 1 - and this doesn't seem to be changing anytime soon.'You guys know how to code...'
there are always a billion people looking for features to add to rf (or suggestions like this) but a VERY limited amount of programmers (all volunteers at that) to implement any changes.
It's amazing that RF has progressed as far as fast as it has - starting from scratch and rewriting everything is a major undertaking and not something that most of the rf programmers are interested in doing, particularly for something that they do as a hobby in their spare time.
RF has always been about 'providing the technology to the community', not selling out to epic etc and making billions of dollars from it.
If someone was able to sell a GAME made with RF to a publisher and make billions of dollars off it, then great - all the power to them. that's the reason RF exists.
Having an 'end goal' in mind, such as selling your company or selling your engine is completely opposite of the whole reason that most of us are in the RF community in the first place.
Take the Reality Engine & Artificial Studios as an example. Sure they made a cool engine, and sold their company to Epic. What do they get out of it in the end?
Not a single game was released using Reality Engine - how does this benefit the game development community at large - it doesn't whatsoever
How does this benefit the owners of Artificial Studios - they USED to own their own company, could choose how their company is run, etc - but now all they are is Epic employees...no matter how much they sold the company for (and i can tell you that it wasn't much), all they got in the end was a job at Epic.
Same thing with the creators of Counter-strike - they made the worlds most popular online game & what reward did they get? A Job at Valve. woo hoo
If you want to really make money in the game industry, it is ALL about creating a cool new game concept (the Intellectual Property) - this is what will make you a lot of money in the long run...
-
- Posts: 51
- Joined: Fri Jul 08, 2005 1:31 pm
- Location: India
My statement seems to have caused a lot of hue and cry
Gekido(and others), I just talked about selling rf to epic, to express that if rf was integrated into OGRE, our engine will be close to epic's, I really dont want rf to be sold to the like of them.....not in a million years.
Moreover, Ogre already hasw features like dot3, embm, shaders, volumetric lighting/shadows, per pixel lighting, physics, networking, etc. That took so long/are yet to be integrated into rf.The only thing left to add is scripting, an editor, a few entities, etc, and thats about it!
Gekido(and others), I just talked about selling rf to epic, to express that if rf was integrated into OGRE, our engine will be close to epic's, I really dont want rf to be sold to the like of them.....not in a million years.
Moreover, Ogre already hasw features like dot3, embm, shaders, volumetric lighting/shadows, per pixel lighting, physics, networking, etc. That took so long/are yet to be integrated into rf.The only thing left to add is scripting, an editor, a few entities, etc, and thats about it!
- QuestOfDreams
- Site Admin
- Posts: 1520
- Joined: Sun Jul 03, 2005 11:12 pm
- Location: Austria
- Contact:
kids... kids... kids...
I don't know what you think, but porting rf to another engine isn't as easy as just pressing a button and voila it's done... the RF source code consists of several 10.000 lines of code
just as an example, there are over 800 references to geActor or geActor_ functions in 40 different files, 900 to geWorld and almost 1400 to geEntity
you just have to think of an engine that doesn't support a system like the genesis entity system and you can forget half of the RF code
additionally I want to mention that you can't just start porting rf to any other engine, first you have to learn how this new engine works, and I can tell you that it will take at least several weeks to learn the basics (well you can probably learn how to setup the engine and render a simple room in 1 or 2 days)...
I don't know what you think, but porting rf to another engine isn't as easy as just pressing a button and voila it's done... the RF source code consists of several 10.000 lines of code
just as an example, there are over 800 references to geActor or geActor_ functions in 40 different files, 900 to geWorld and almost 1400 to geEntity
you just have to think of an engine that doesn't support a system like the genesis entity system and you can forget half of the RF code
additionally I want to mention that you can't just start porting rf to any other engine, first you have to learn how this new engine works, and I can tell you that it will take at least several weeks to learn the basics (well you can probably learn how to setup the engine and render a simple room in 1 or 2 days)...
I dl'd Jet3D and i was less than impressed by it. With it's supposedly DX9 renderer it had graphics crappier than RF (no offense) while running at a low framerate and even used some of the same default RF media.
I DL'd it too, the content is the same as it was in 1999, but all both (that's 2 in numbers) of its developers keep posting about how wonderful its features are, maybe they should post less and
produce something. My favorite 'what to replace RF with' engine is zeropointgameplay.com. Open source, multiplayer, RPG. Right now there's no way to make
content unless you have Maya.
I DL'd it too, the content is the same as it was in 1999, but all both (that's 2 in numbers) of its developers keep posting about how wonderful its features are, maybe they should post less and
produce something. My favorite 'what to replace RF with' engine is zeropointgameplay.com. Open source, multiplayer, RPG. Right now there's no way to make
content unless you have Maya.
No one was lead to believe that Genesis3D 2.0 (later named Jet3D), even when it was released pre-alpha as ‘Genedit’, that there would be any easy port from Genesis3D.in THEORY, the original idea for jet3d was that you could do a port from genesis fairly easily, however since the two projects have diverged in their own directions without a central planning source (ie the role that Eclipse originally played), i doubt that this would be an easy task.
It was clear to us that too the new engine would handle the BSP very differently, scene managment included portal rendering, and actors would would not directly convert. (Required a rebuild using the base NFO/KEY files.)
Part of the reason understood is that Jet3D levels are Subtractive, as Unreal, while G3D= Add. G3D compiles a BSP, using a PVS logic for optimization. Jet3D Doesn't compile, and Uses VIS portals.
We were all salivating at the rednering quality also. So anyone who was around then knows Jet3D is a much better looking render. Sorry to break the news, this must come as quite a shock.
Huge difference between G3D and Jet3D.
Compile times w/ default level design= G3D: 4 hours Full - Jet3D: 3 Min.
Rendering Quality= G3D: Per-Quake2 - Jet3D: Equal / better than Unreal 1.0
Editor = G3D: BSP full-bright, static need to compile to get full preview
- Jet3D: Full WYSIWYG, move and adjust light and see changes dynamically in real-time. Sound, particle effects, lens flares, all effect can be more easily set in a “full” real-time editor.
Optimizations= Unlike G3D, Jet also uses ‘portals’ for scene management. John Pollard, wrote that properly set VIS portals can be more powerful than a PVS that takes hours to calculate automatically. Let’s face it often times ‘manual’ is the best way.
Test, with the SAME level template used to show the rendering quality, benchmarked better performance statistics than G3D back in 99. Of course back then it was more of a stability issue with the core engine and editor. Which, has been fixed on so many level, especially major fixes.
As with other features it is a better compare with Unreal than it is to Genesis3D, at least as far as game content creation goes.
Another example, is skyboxes. Skyboxes are rendered using a portal, as with Unreal. So your skybox can be this small 128x128x128 cube, you place a portalobject(camera) in it, and projects the infinite sky appearance. After we get layers for one, this will look great with rolling clouds over the sun or moon.
If rf is integrated into OGRE someday, we could really sell RF to Epic(Rf=Unreal4)
I felt like that Reality Engine acquisition was more buying out the competition. Then they will pillage the feature they like, and leave the carcass
I dl'd Jet3D and i was less than impressed by it. With it's supposedly DX9 rendered it had graphics crappier than RF (no offense) while running at a low frame-rate and even used some of the same default RF media.
That and other statements made are based on [a lack of knowledge].
First if you are getting bad frame rates, it is either due to you video card setting. Mainly Anti-aliasing (check the default cube, if it shows, 6-9fps, it is an anti-aliasing issue)
The other reason, and apparently none in this community have ever really dug deep into Unreal. Is “VIS portals” With set VIS portals, frame-rates are actually equal or often better than G3D has to offer.
Rendering: Back in 99, the rendered looked at least as good as Unreal did back then. Being ‘blind’ as opposed to just plain ignorant as to this may be the reason here.
If you poked around Jet3D University, you would see the compare in two screenshots and read about the frame-rate issues as well.
Jet3D = Reality Factory Wannabe, IMO.
I think, (The level editor app)’RF edit pro’ was a failed attempt at a true WYSIWYG editor like Unreal, and Jet3D, so you could say that was a “wanna-be” like Jet3D editor from it’s original design. Others have tried to surgically remove Jedit from Jet3D and add it to another engine to no avail. Quite simply, it would be easier to write one from scratch then attempt to make any use of the Jet3D editor with another engine. Especially Quake clones. Jet3D goal is to be like ‘Unreal’. Although the RF team has made many enhancements, it's still an outdated engine core. Bottom line: Not much to “wannabe”
Jet3D was always an "Unreal Wanna-be", as I was told by it’s designers. The Jet3D Dev team will carry on that vision.
2 Developers?I DL'd it too, the content is the same as it was in 1999, but all both (that's 2 in numbers) of its developers keep posting about how wonderful its features are, maybe they should post less and
produce something. My favorite 'what to replace RF with' engine is zeropointgameplay.com. Open source, multiplayer, RPG. Right now there's no way to make
content unless you have Maya.
Let me see:
1. Tom Morris.
2. Jeff Mossenette
3. Krouer
4. Paradoxnj
5. Seven
6. SuperPig
7. Myself.
Yep, that’s two.
It’s the quality and the time devoted of each, not the quantity.
Same in 1999?
Let me rattle off a few of the many changes.
First all of the bugs have been fixed since 99, bugs that drove me and others away.
Of course, you would have to have been there or read the old forums more to know and understand the real issues.
What was originally called ‘Jedit’ had an overhaul twice, first for the CVS, Which gave us Vertex editing. Which opens up more options for shaping the primitives Then Tom Morris’s Jstudio.
Recently he added enhancements to the Texture browser. As now you can organize you textures into a tree, by category, and refresh the contents without having to reload the level editor.
Actor studio Texture size limitation was removed, Wireframe view and Drawface statistics added, which is essential is setting up VIS portals...
('getting tired of telling what all should already know here)
Before making statements absolutely untrue, research and ask, there is enough info, in Jet3d.com and Jet3d U and their subsequent pages. All of this info is on here.
Heck the start up page for Jet3DU shows a clear rendering difference advantage of Jet3D over G3D. Dare-to-compare, I bet any money Jet3D is better looking than G3D.
There are more new features now done lately then was in the last release.
poke around the wesbites, if you ahve the time it is all spelled out...
(more to come...)
i appologize. i admit i didn't spend much time with it so i shouldn't have been so quick to judge.
I was impressed by the ability to quickly preview a level w/out compile. and i admit i didn't DL the gameshell app, so im sure i missed a lot of stuff. what i disliked the most is that even though it was billed to have DX9 and shaders i couldn't even tell if there were integrated into the build i had, per lack of current documentation, an issue RF had for a long time as well. to tell the truth i wasn't impressed with RF at first either, so dont feel bad. I know you guys put a lot of hard work into it and i commend you for your efforts. Im sure one day J3D will be a top engine
I know the work it takes to write an engine, i've been a part of several engine writing projects and am currently involved in one now for an entry into the gamedev.net 4E4 contest. So dont take my comments wrong, you guys have done a great job. I wish the best to you and your project!
I was impressed by the ability to quickly preview a level w/out compile. and i admit i didn't DL the gameshell app, so im sure i missed a lot of stuff. what i disliked the most is that even though it was billed to have DX9 and shaders i couldn't even tell if there were integrated into the build i had, per lack of current documentation, an issue RF had for a long time as well. to tell the truth i wasn't impressed with RF at first either, so dont feel bad. I know you guys put a lot of hard work into it and i commend you for your efforts. Im sure one day J3D will be a top engine
I know the work it takes to write an engine, i've been a part of several engine writing projects and am currently involved in one now for an entry into the gamedev.net 4E4 contest. So dont take my comments wrong, you guys have done a great job. I wish the best to you and your project!