Things we want for 1.2

[c4]

Yeah, honestly, the fact that multiple games still use this amazing engine in and of itself proves that we shouldn't change.  It makes the game accessible for people who don't have a fast enough computer to play MW2.

[CATAHA]

Ok, what computer you using? I can bet not so much peoples here playing on 'Pentium1 or lower' hardware, and it happens very rare. You never thinked about why web developers dont care about ie2 or netscape 3? they even dont care about ie6 mostly, and ie6 not so outdated software. =] I can even say that current engine of Trem is shit, caus its not running on my old x386 notebook, if im gonna use your logical base.
About models... How you can ask peoples make models for this engine, if engine not support them? Its like 'ok, make new models and we will think about enabling support for them in our code'? Sounds funny. Peoples too lazy and they wont do that, especially when they almost sure that Trem devs wont enable such support.
And, once more about xreal. Im not saying that it always faster than ioQ3. But in most cases it faster, just caus render engines getting beter with every day. If you too lazy for check links by yourself, i can re-post article here. Its from official xreal project site:

Renderer Speed Comparison IOQuake3 vs. XreaL
Now I want to show you the need of Vertex Buffer Object based rendering. Usually most open source quake based game engines have a similar renderer design to id Tech 3 nowadays. They usually use vertex arrays (VA) or compiled vertex arrays (CVA) that they create every frame by processing the entire scene geometry. You usually find this practice in IOQuake3 (of course), all IOQuake3 based games like Urban Terror or Tremulous, Q2E, QFusion, Overdose (at least an old version) and Alien Arena. This is a design that worked well in 1999 but it is a quite bad practice in 2009.

XreaL follows the advices of GDC papers and uses VBO based rendering for almost everything to avoid this horrible CPU bottleneck. The model geometry is uploaded like textures and don’t have to move over the PCIe bridge every frame.

There is a nice heavy benchmark map called tvy-bench.pk3 for Quake3 by hipshot that could be found at http://www.zfight.com. It is really nice and contains a test scene with really many polygons.

To make it short. Traditional renderers don’t give you more than 30 fps no matter what high-end GFX card you have. The XreaL renderer gives you more than stable 125 fps.

IOQuake3 with 22 fps in 1280×768:


XreaL renders it with 148 fps:

[David]

On a high end graphics card sure.

What do you think will happen on my old ultra-cheap pc?
It plays trem fine, somehow I doubt xreal will work very well.

And a benchmark map doesn't really mean anything for this discussion.  How does it handle normal maps that we care about?

[CATAHA]

xreal have normal map afaik. =) And it can be widely tuned for old videocards. Im not coder and cant say about details, but i know that engines evolving, devs making new optimal solutions, etc. Thats why i think that game engine upgrade is good idea. And, anyway, i think over 95 real (and possible) players nowadays have modern enough hardware to handle engine like xreal.

[Tremulant]

So people can't be arsed to show their ability to produce the superior models that we need xreal to use? great.

Modern engine gets higher framerates on modern GPUs vs. ioq3 with the cpu bottlenecks that it's known for, well, isn't that a shock... how does performance compare on older hardware, where it matters? don't quote a benchmark that lacks specifics(GPU CPU combo used?) and is performed with a torture map. If anyone's going to make the decision to move engines it's going to be the devs, and i don't imagine it's going to happen in a rushed way or just for the hell of it.

Maybe trem can move to xreal for a future release, but just forget about it for the time being, surely no one wants the extra work of porting it when they're already tied up with 1.2 and the community(well, some people have done nice work from time to time, but they're not the ones demanding the port now) have yet to demonstrate their credentials for improved asset production.

[jm82792]

I mean that it takes a lot of work to optimize/upgrade the current engine,
and since it's so old you runt he chance of breaking things.
Honestly what qualifications do you have regarding game engine and the ability to call people idiots?

So people can't be arsed to show their ability to produce the superior models that we need xreal to use? great.

They have in the past on a few occasions.
The thought is that they won't unless they have a reason,
take the third race as an example.
They made a fat portion of the assets when they had a reason.

Modern engine gets higher framerates on modern GPUs vs. ioq3 with the cpu bottlenecks that it's known for, well, isn't that a shock... how does performance compare on older hardware, where it matters? don't quote a benchmark that lacks specifics(GPU CPU combo used?) and is performed with a torture map. If anyone's going to make the decision to move engines it's going to be the devs, and i don't imagine it's going to happen in a rushed way or just for the hell of it.

I agree, selecting stuff based on initial impressions never goes well.

Maybe trem can move to xreal for a future release, but just forget about it for the time being, surely no one wants the extra work of porting it when they're already tied up with 1.2 and the community(well, some people have done nice work from time to time, but they're not the ones demanding the port now) have yet to demonstrate their credentials for improved asset production.

So the biggest reason we have no engine port is that we have no better assets(well we have a few that could work with some tweaking), and the port isn't attractive enough without better assets? Honestly the assets are nothing in (my humble opinion) compared to getting the coding done.
The devs have some "power" and they could say we need assets for the eventual engine port,
and you get the idea.

[freezway]

My point is, higher poly looks nicer and is easier... high res looks nicer, I run at 1080p and low res is very very noticable, and my system runs at ~200fps, so ALLOW, NOT FORCE, people to use higher poly models.

[Asvarox]

That's cool if ioq3 runs faster than xreal on older machines, but you gotta admit that there are less low-end computers in use and more modern ones everyday. Soon you will wake up and notice that there are few people left playing trem because it doesn't attract new players (as most of them tend to rate thing by how they look first, not to mention 4 years without update), and old ones will just get bored.

As for "there are no assets so there's no point to implement new engine features/port" - that's kind of wrong, because I'm 100% sure that nobody will create something for the game that can't even use it. as Amanieu said quite long time ago "Technology attracts artists". And port was done at least once, by TremFusion guys (= Amanieu ).

So people can't be arsed to show their ability to produce the superior models that we need xreal to use? great.

TRaK showed some of his work for a game based on xreal in random dev shots, really neat stuff.

I'm not asking to port Trem to xreal for 1.2 - I just want 1.2 to be released.

[c4]

My point is, higher poly looks nicer and is easier... high res looks nicer, I run at 1080p and low res is very very noticable, and my system runs at ~200fps, so ALLOW, NOT FORCE, people to use higher poly models.

Make some, put them in a local pk3, and release them here.  Then people have a choice...

[Pazuzu]

Oh, by the way, here's a more practical feature most people can probably agree on: Antialiasing and anisotropic filtering. Those jagged edges need to go, and I shouldn't have to use the NVidia control panel as a workaround to force it. This is 2010, these should probably be standard features.

[Plague Bringer]

On a high end graphics card sure.

What do you think will happen on my old ultra-cheap pc?
It plays trem fine, somehow I doubt xreal will work very well.

And a benchmark map doesn't really mean anything for this discussion.  How does it handle normal maps that we care about?

David, come on, you're intelligent. This benchmark is arguably more CPU intensive than ATCS or Niveus or Arachnid2.

And who cares about your dino-PC? In an advancing videogame market (which is the current status of the videogame market) things get better. They get better because it's a good idea to appeal to the majority rather than the minority. You are part of the minority.

[c4]

Oh, by the way, here's a more practical feature most people can probably agree on: Antialiasing and anisotropic filtering. Those jagged edges need to go, and I shouldn't have to use the NVidia control panel as a workaround to force it. This is 2010, these should probably be standard features.

What's wrong with using your control panel?  It's like what? 10 seconds more work?

[Pazuzu]

It's just the sort of feature that should be part of the actual program. You can't access the control panel in-game... Besides, most modern engines have it.

[freezway]

Yeah, antialising should be standard.

[Thorn]

Errr, r_ext_anistropy and r_ext_multisampling aren't in the stock client?

[c4]

Errr, r_ext_anistropy and r_ext_multisampling aren't in the stock client?

They are, just the people complaining about stuff don't know anything at all.

[gimhael]

The 1.2 client has /r_ext_multisample, /r_ext_texture_filter_anisotropic and /r_ext_max_anisotropy to be precise.

[Pazuzu]

In that case, I guess what I want are controls for them in the Options, maybe with bloom controls too. The average newbie isn't going to look up those variables (we know this already), or even know that there are variables for them, because... well, seriously, we want people to look up variables to change bloom?

[c4]

If you have all that shit, you end up like the nexuiz settings window which is almost impossible to use.  If one cares so much about bloom and anti aliasing, then they will look up the cvar.

[A Spork]

It's just 2 little lines dude....
Antiliasing: No/2x/4x/whatever.
Bloom: Slider(?)
And those are the kind of things newbies should be able to find....

[SlackerLinux]

default client doesnt have bloom

also porting to xreal fine im sure it could happen later but for 1.2 no way they are nearly finished after years and your saying hey lets start again on another engine no just no wait for 1.2 is released then for 1.3(or should it be 2.0 would be a big change) then the devs could consider it

[jm82792]

Slacker I agree with you.
Finishing something isn't easy.
I've been working on a lighthouse, cloud, water, rocks, etc. scene for a few months now.
Still not photo-realistic(I'm aiming for it though),
if I lived and breathed Blende it would have been done in a few weeks.
The devs aren't paid so things don't happen quickly.





Well to put it simply I want to get some new stuff into the game engine to give it some life,
besides I enjoy stuff like it and in my spare time after this large project I am working on is finished.
My friend is working on a 1K poly OM, he is also drafting the plans for the house we are going to build,
and is doing a plethora of things but should finish it soon. The mesh is around 70% done, and I will texture it.
I think making the antiailising settings and such GUI would be a VERY VERY sweet thing,
we need more players and having things more newbie friendly is always nice.

Regarding the people who whine about how their old computer won't run a modern client.
Soon i7 PCs will be $500, chances are that your grandma probably owns a computer with a half gig of ram,
a couple megahertz CPU, and a decent integrated graphics card. Eventually your computer will fry,
and you will buy one. If it's a NEW (as of a few years) low end laptop or desktop it WILL play trem on a PORTED client.. A droid phone( I think) can run the current trem with some work.

[Tremulant]

Regarding the people who whine about how their old computer won't run a modern client.
Soon i7 PCs will be $500, chances are that your grandma probably owns a computer with a half gig of ram,
a couple megahertz CPU, and a decent integrated graphics card. Eventually your computer will fry,
and you will buy one. If it's a NEW (as of a few years) low end laptop or desktop it WILL play trem on a PORTED client.. A droid phone( I think) can run the current trem with some work.

How do you know all this, is it not possible for a low-end laptop to use intel graphics? does xreal outperform ioq3 on such hardware(if it does that's great, tell me)? As for this braindead statement about phones, someone was pulling a blistering 4FPS or similar with an android port, i predict that we'll all be playing the xreal port on our phones by this time next year...
I have a new computer now, i have more processing power than i know what to do with, i'm not concerned about my ability to play the game, what does concern me is that an important part of the market for free games consists of  those people who are stuck with less than entirely modern hardware, it would be a shame if these players were to miss out because modders want to cram their sloppy(i _do not_ model myself, so however sloppy these are better than i can manage, but still) models into the game...

Hey, it's not that i don't want to see trem progress, if the devs feel that a port to xreal at some stage in the future is worthwhile i'll be pleased to see it happen, but it makes sod-all sense to rush into such an engine shift when everyone's already deeply upset that 1.2 isn't out yet, do you lot think a switch to xreal would hasten 1.2's release?

Adding options in the gui for AA, bloom, etc doesn't seem unreasonable, as long as we stop short of getting sauerbraten's graphical settings dialogue.

[freezway]

yeah, no new engine for 1.2, sounds are the main thing we need for that. bloom and anti* settings would be nice.

[Pazuzu]

Adding options in the gui for AA, bloom, etc doesn't seem unreasonable, as long as we stop short of getting sauerbraten's graphical settings dialogue.

Oh god no.
"Do you want high-detail shadows? Or slightly-less-high-detail shadows?"

[jm82792]

Regarding the people who whine about how their old computer won't run a modern client.
Soon i7 PCs will be $500, chances are that your grandma probably owns a computer with a half gig of ram,
a couple megahertz CPU, and a decent integrated graphics card. Eventually your computer will fry,
and you will buy one. If it's a NEW (as of a few years) low end laptop or desktop it WILL play trem on a PORTED client.. A droid phone( I think) can run the current trem with some work.

How do you know all this, is it not possible for a low-end laptop to use intel graphics? does xreal outperform ioq3 on such hardware(if it does that's great, tell me)? As for this braindead statement about phones, someone was pulling a blistering 4FPS or similar with an android port, i predict that we'll all be playing the xreal port on our phones by this time next year...
I have a new computer now, i have more processing power than i know what to do with, i'm not concerned about my ability to play the game, what does concern me is that an important part of the market for free games consists of  those people who are stuck with less than entirely modern hardware, it would be a shame if these players were to miss out because modders want to cram their sloppy(i _do not_ model myself, so however sloppy these are better than i can manage, but still) models into the game...

Hey, it's not that i don't want to see trem progress, if the devs feel that a port to xreal at some stage in the future is worthwhile i'll be pleased to see it happen, but it makes sod-all sense to rush into such an engine shift when everyone's already deeply upset that 1.2 isn't out yet, do you lot think a switch to xreal would hasten 1.2's release?

Adding options in the gui for AA, bloom, etc doesn't seem unreasonable, as long as we stop short of getting sauerbraten's graphical settings dialogue.

You do not seem to grasp the jest of what I am saying.
I am not trying to treat what I am saying as an absolute fact but more of an example of where computing power is and where it is going. People are crying over how their slow computer plays trem, I am saying soon enough the cheapest computer you can buy should be able to play a ported version of trem. Yes I am right trem will play on a droid, 4 FPS yes but still my point is still valid and I did say "I think". That means I did not state that as a factual statement!!

I agree wait for 1.2, twiddle our thumbs and see if anything gains momentum.
A GUI would be nice as you said.

Regarding modeling, you've gotten a bad taste in your mouth because a decent portion of the talent here isn't good.
There are a few gems but I got back into 3D(year or so hiatus) because I saw what is being put out and it made me sort of gasp in how discombobulated it is/was. My friend who is the amateur architect/draftsmen. who can model really well was also laughing and getting sick at the same time.


Tremulant try to cut the crap and utilize your intelligence to explain why I'm being an "idiot".

[Tremulant]

You do not seem to grasp the jest of what I am saying.

Ah, you jest, that's alright then.

I am not trying to treat what I am saying as an absolute fact but more of an example of where computing power is and where it is going. People are crying over how their slow computer plays trem, I am saying soon enough the cheapest computer you can buy should be able to play a ported version of trem. Yes I am right trem will play on a droid, 4 FPS yes but still my point is still valid and I did say "I think". That means I did not state that as a factual statement!!

I think moore put it best when he said "Yo dood computers dey gets fasters".

I agree wait for 1.2, twiddle our thumbs and see if anything gains momentum.
A GUI would be nice as you said.

Regarding modeling, you've gotten a bad taste in your mouth because a decent portion of the talent here isn't good.
There are a few gems but I got back into 3D(year or so hiatus) because I saw what is being put out and it made me sort of gasp in how discombobulated it is/was. My friend who is the amateur architect/draftsmen. who can model really well was also laughing and getting sick at the same time.

I have a bad taste in my mouth for various reasons, I don't care about your architect friend, is he modelling for trem?
Basically we have a bunch of enthusiastic but inexperienced modellers going "ZOMG EVIL DEVS U NOT RAISE POLY LIMIT, U STOPS US CONTRIBUTIN"
 

Tremulant try to cut the crap and utilize your intelligence to explain why I'm being an "idiot".

I was generalising about those who seriously believe it's impossible for trem's appearance to improve without a new engine and that a new engine is a great direction for the DNF of open source game releases to take, but since you ask, yes, i suppose you're an idiot too, sadly the reasons are becoming too numerous to list...

Why don't you all just agree that number one on your list of things you want for 1.2 is:

• A RELEASE DATE

Then you can go off and start a "what we want for trem 1.4" thread, where you can speculate wildly to your heart's content, sound good?

[CATAHA]

Why don't you all just agree that number one on your list of things you want for 1.2 is:

• A RELEASE DATE

May be because most peoples here understanding that release date depends from amount of features devs gonna release in 1.2?
Untill 1.2 not released yet and no one can say when this event gonna be... all can talk about anything they want see in 1.2 version. May be devs gonna look and say: 'omg, we seriously need this exactly feature before we gonna release 1.2' and shift release date. =)
So dont blame peoples for talks.
P.S. We all know about things and features in 1.2, even unreleased models wont be surprise to us commonly. Why you suprising that peoples like talk about things that was lost sight of?

[UniqPhoeniX]

No artist cares about making a high quality model just to show that they can if they can instead make a model for another game that is likely to use it. Expecting a model to get the devs (the same ones who have taken 4.5years for a new version... so far) to take on and finish the huge task of porting to XReaL before hell freezes over is kind of unreasonable at this point.

I agree that 1.2 should be finished first (someone PLEASE make some sounds ), but you won't be getting any skilled artists making assets for a 10+ year old engine.
Also, so far many large Trem related projects (TF, AMP, Unv,) have died due to lack/fail of coders, so even when you are half way done, gl convincing any artists that you are going to finish it.

If anyone has concrete information on the CURRENT min requirements of XReaL, please post/link. (I'm guessing GPU from ~2006 should work?)
And the reason such a large percent of the free game market has crap hardware is that almost everyone else has moved on to a commercial/F2P game/engine, which is because there are barely any decent free & open games with a good engine. Porting to a newer engine has a good chance of attracting lots of players from commercial games.

Also Tremulant, stop trolling, Trem is already the DNF of open source games. (Actually, DNF could come out sooner than 1.2 for all we know.) There are pretty much only inexperienced modelers (and Stannum and a handful of semi-experienced modelers) BECAUSE of the old engine.

[Thorn]

Nvidia 6200+, won't run on intel, no idea of an ATI comparison.