The actual UT2003 demo

[demalion]

Hmm...nice piece of work. Some things I noticed:

The truform setting does actually do something, and the change with even tesselation level 1 is noticeable (I actually don't see the point in going higher, the calf section of the armor on the models, for example, looks perfectly curved at 1, even zoomed in...maybe I'd need a higher resolution).

With my image quality heavy settings at 1024x768x32 (every setting in game maximum except physics, I believe, and aniso at 4), I get about 76 fps flyby, 28 fps botmatch. In game, during CTF, my average fps is from 27 fps to 34 fps. This to me indicates that the major loss of fps is either due to the added polygon detail from the characters, or the CPU load from physics and AI.

I turn on truform (level 1 tesselation), and my fps barely drops at all (same range, though overall it may have dropped my 4 or 5). From this, it points out that atleast Truform isn't being used inefficiently, and that it is the CPU that is the limitation.

This is on a KT133 MB with 512MB RAM and a 1.3GHz TBird @ 1.377 by FSB overclock (am I the first one who has posted system specs at this forum at the beginning of a thread? I feel so dirty :lol: ), video card a Radeon 8500 w/128MB.

Let the dissection of the demo and benchmark results begin?

Oh, did I already say that I enjoy the demo, or would that be off topic? ;)

[noko]

Link to the Demo please.

[Dave Baumann]

http://www.nvidia.com/view.asp?IO=dl_ut2k3

[noko]

Thanks Dave, downloading :).

[mat]

mirrors (http://www.unrealtournament2003.com/?downloads)

btw: why are all so excited about this game? i've played it a little and there was not really anything new in it (except nice graphics)

[Sabastian]

Epic has really done some nice work with UT2003. On default settings my GF DDR was fairly choppy, but after I overclocked to 130/330 it played fairly good.(granted I had to drop some settings down a bit.) I found it playable at any rate. Nicely developed game really. I can see this game pushing lower end cards.

I am thinking a Radeon 9000 pro would be a good cheap upgrade, but I am unsure if it would make a great difference as I only am running a P3 700 mhz. Can anyone tell me if a graphics card upgrade would benifit me given my CPU is only 700 mhz or should I consider not bothering?(Please reply soon I have to go to work. thanks in advance.)

[Dave Baumann]

I am thinking a Radeon 9000 pro would be a good cheap upgrade, but I am unsure if it would make a great difference as I only am running a P3 700 mhz. Can anyone tell me if a graphics card upgrade would benifit me given my CPU is only 700 mhz or should I consider not bothering?(Please reply soon I have to go to work. thanks in advance.)

GAH! Read the review - I make exactly this type of point in it!

http://www.beyond3d.com/reviews/ati/radeon9000pro/

(i.e. specifically read the T&L vs Non-T&L)

[Sabastian]

[quote=Sabastian] GAH! Read the review - I make exactly this type of point in it!

http://www.beyond3d.com/reviews/ati/radeon9000pro/

(i.e. specifically read the T&L vs Non-T&L)

LOL, sorry Dave I have not had the chance lately to read much I been called into work more then usual lately. At any rate after a quick skim through it appears that particularly where T&L is concerned there is a worth while increase. Thanks for that Dave. I can get a Radeon 9000 pro for $150 Can.. pretty cheap/good upgrade I figure.

BTW Dave that is a great review that you have managed there, nice to see high, mid and low end systems. Later.

[alexsok]

Funny enough:

Textures in the full game are, on average, 2X higher resolution than in the demo. The textures in the demo were reduced because of download size limitations.
The full game includes over 30 maps, compared to the 4 in the demo.
The full game includes over 50 characters, versus 4 in the demo.
The demo only supports "instant action" mode, Internet and LAN play. The full version also contains the single-player ladder, enabling you to play through the game while managing a team roster.
The demo doesn't include the Unreal editor and content-creation tools which ship with the final game, enabling you to build new levels, modify existing ones, and create game mods.
The full version includes 6 mutators offering new variations on gameplay, versus 2 in the demo.

[Dave Glue]

Textures in the full game are, on average, 2X higher resolution than in the demo. The textures in the demo were reduced because of download size limitations.

Glad to hear that, as actually I'm not that impressed with the demo. I expected a better frame rate at 1024*768*32bit with no aniso or AA on my GF4 4200+XP1800 system. The texture detail wasn't overly impressive.

And where's the detail textures? Is there a setting I'm missing, or did Epic not do them this time around?

[alexsok]

And where's the detail textures? Is there a setting I'm missing, or did Epic not do them this time around?

If I'm not mistaken, they were also missing from that leaked build some time ago...

I'm sure they will make their appearance in the full version...

[Fafner]

Anyone got the game running without artifacts on a Voodoo5? (mark rein said it would run ok with a V5 a while back).

[jackkoho]

is this the same engine that unreal 2 is going to use? sorry i've just always been confused about this.

[alexsok]

is this the same engine that unreal 2 is going to use? sorry i've just always been confused about this.

Yes, it's basically the same engine, although Unreal II may contain more nifty and beautiful effects and higher polycounts, since it's SP oriented.

[jackkoho]

thanks.

[demalion]

Textures in the full game are, on average, 2X higher resolution than in the demo. The textures in the demo were reduced because of download size limitations.

Glad to hear that, as actually I'm not that impressed with the demo. I expected a better frame rate at 1024*768*32bit with no aniso or AA on my GF4 4200+XP1800 system. The texture detail wasn't overly impressive.

And where's the detail textures? Is there a setting I'm missing, or did Epic not do them this time around?

Hmm, there is a detail texture setting in the .ini file, and I could have sworn I saw a setting in game as well.

I've seen what I thought were detail textures on some floors and walls on the indoor DM map and the egyptian bomber ball (or whatever the name is...very much fun IMO) map, but I don't remember every noticing it to any degree on the outside DM map or CTF map.

[Dave Glue]

Hmm, there is a detail texture setting in the .ini file, and I could have sworn I saw a setting in game as well.

I've seen what I thought were detail textures on some floors and walls on the indoor DM map and the egyptian bomber ball (or whatever the name is...very much fun IMO) map, but I don't remember every noticing it to any degree on the outside DM map or CTF map.

The setting is there, I just find them very sparse. You're right, I did notice a couple in the indoor map, but at least in the demo they're quite rate compared to UT. I'll wait until the full game release to determine texture quality however.

[jackkoho]

how do you enable timedemo?

[KnightBreed]

From the release notes:
Due to the lack of certain key functionality like texture compression and
cubemap support TNT2, Kyro II, Voodoo 3, G400, ... cards won't be able
to run the game at the full visual quality.
Since when? I can't seem to get over 5fps, regardless of the visual settings. I'm downloading DX8.1b now. I hope this helps. :(

[duffer]

timedemo: look in the C:\UT2003Demo\Benchmark folder, run one of the batch files you see there, the stats end up in two files:

benchmark.log
lowframerate.log

My impressions:

The graphics (especially the player animations) look great.

The gameplay feels very similar to Unreal Tournament. If you liked UT, you're probably going to like UT2003.

[Typedef Enum]

Hey, the Parhelia side of things is shockingly amazing...

Matrox just released some drivers yesterday, so they may have been preparing for this...and maybe Epic had taken a look @ it over the last few months...

I just fired up the "old" demo a few days ago, and it was definitely playable, but also fairly slow. I was very worried about it before, but now I'm feeling much better :)

From what I have seen on the net, the Parhelia will score better when it counts...in the bot-match and/or with higher IQ settings. Enabling FAA doesn't really drag the performance down to an unacceptable level, and you should just see what it looks like :)

Anyhow, I'm very impressed with the work that Epic has done. It's very clear to me that they addressed a lot of the performance issues with this demo, and I'm sure this will only continue to get better.

BTW, looking forward to the real game with the 2x detailed textures!

[Doomtrooper]

After playing with the Demo last night for a bit I am very dissapointed...this game is not Unreal Tournament...it looks and feels like Quake 4.
Physics are all wrong, model movement , powerups...everything about it takes away what made UT better than Quake 3..Shock Rifle is slow, Rocket Launcher is ruined...but it looks pretty like Quake 3..

I certainly hope they fix the Gamplay physics..almost all my Clan Mates agree..Quake 4 feel.

[Dave Glue]

I certainly hope they fix the Gamplay physics..almost all my Clan Mates agree..Quake 4 feel.

Same here. Just seems...dull. I'm not even sure I'll pick it up when it hits the shelves.

[Reverend]

After playing with the Demo last night for a bit I am very dissapointed...this game is not Unreal Tournament...it looks and feels like Quake 4.
Physics are all wrong, model movement , powerups...everything about it takes away what made UT better than Quake 3..Shock Rifle is slow, Rocket Launcher is ruined...but it looks pretty like Quake 3..

I certainly hope they fix the Gamplay physics..almost all my Clan Mates agree..Quake 4 feel.
Careful now... gameplay has no place at Beyond3D.

:)

FWIW, I have never really liked UT nor Q3. Probably since I have no time for "ah, time for a beer and a blast-fest". The two games have their place but not exactly my cuppa.

Oh, and it looks like Quake4? Really??

[Tahir2]

It is a massive improvement on the leaked 927 build. That build had a couple of bugs in it which would slow your whole system down after finishing with a map.

And it is pretty darn fast on my system at the moment.. but then again it should be:
P4 2.4GHz overclocked from 2GHz
GF4 TI4200 (to be replaced or repaired soon).

I would like the detal textures back.. scouring the .ini file right now :)

Personally speaking.. it is not UT. I think UT was better gameplay wise.

[noko]

Looks OK to me, my GF3 when doing anisotropic filtering seems to lock up the computer peridocally, otherwise relatively smooth. I think I am becomeing CPU limited here more so then Graphic card limited.

[Babel-17]

To view your fps hit Tab and then type in

stat fps

In the games system folder there is the very neat little program benchmark.exe, be careful with comparisons as results between cards that do/don't support advanced features are not one to one. V5 users might experience better luck using ogl. To use ogl merely change

RenderDevice=D3DDrv.D3DRenderDevice
to
RenderDevice=OpenGLDrv.OpenGLRenderDevice

in your C:\UT2003Demo\System\UT2003.ini file under the [Engine.Engine] heading. Lol, I keep a shorcut on my desktop. And yes, always keep a backup of that puppy. :)

I hope this is ok to say, there's imo a good thread over at Slackers on this.

Edit: Edited for spelling and clarity.

[Ozymandis]

Looks OK to me, my GF3 when doing anisotropic filtering seems to lock up the computer peridocally, otherwise relatively smooth. I think I am becomeing CPU limited here more so then Graphic card limited.

I get the same problem on my GeForce3 :evil:

[Nick[FM]]

I certainly hope they fix the Gamplay physics..almost all my Clan Mates agree..Quake 4 feel.
I must admit that when I had played the UT2003 demo for an hour, I had to check if I really was playing UT2003, and not Q3A! :( The graphics are ok (sometimes even great) but not as fab as I expected. The character animations (and physics) are ok too, but still they kinda "glide" all over the place. They lack natural motion.. And the tempo of the game is much closer to Q3A than the original UT.

Duh.. I hope DE & Epic fix the playability for the final full version. I'll stick with BF1942 for now! :P

[Doomtrooper]

Hmmm Worm and Me agree on something http://www.gamers-forums.com/smilies/otn/laughing/astrosmiley.gif

Btw Reverend I meant the feel of Quake 3...or what Quake 4 in my mind would have looked like.

[KnightBreed]

Not that it apparently matters to anyone, but the release notes mention downloading DX8.1b to fix a TnL bug. Man, they weren't kidding. I seemed stuck at <5fps with any details settings. I downloaded 8.1b, and there is a profound speed increase.

* ATTENTION ANYBODY WITH A NON-HARDWARE T&L CARD (especially Kyro users) *

Download DirectX8.1b. It worked for me. :)

Windows 98, ME, and 2000 (http://www.microsoft.com/windows/directx/downloads/drx81.asp)

Windows XP (the fix is included in SP1) (http://www.microsoft.com/windowsxp/pro/downloads/servicepacks/sp1/default.asp)

[Matt Burris]

After playing with the Demo last night for a bit I am very dissapointed...this game is not Unreal Tournament...it looks and feels like Quake 4.
Physics are all wrong, model movement , powerups...everything about it takes away what made UT better than Quake 3..Shock Rifle is slow, Rocket Launcher is ruined...but it looks pretty like Quake 3..

I certainly hope they fix the Gamplay physics..almost all my Clan Mates agree..Quake 4 feel.

I'm totally the opposite. I despised the original UT, the weapons were severely imbalanced, the physics was stupid, and the graphics was just plain horrible as compared to Q3. I enjoyed Q3 more, but not overly much. I like this new UT, it feels and plays much better, IMHO. :)

[borzwazie]

I like the benchmark utility they put in. Kinda neat how it runs you through several scenarios.

I got:

Flyby: 95.820404
Botmatch: 42.343575

This is on an XP1600+, SiS 735, a GF3, Win2kSP3, Detonator 40.41's, 512MB ram, all non-o/c'd, max detail, 1024x768x32.

Game ran great for me, no visual impairments or glitches, framerate stayed high enough for me to wax all j00 fewls :)

[KnightBreed]

I'm totally the opposite. I despised the original UT, the weapons were severely imbalanced, the physics was stupid, and the graphics was just plain horrible as compared to Q3. I enjoyed Q3 more, but not overly much. I like this new UT, it feels and plays much better, IMHO. :)
So the general consensus is:

1) If you like Q3, you'll like UT2k3.
2) If you like UT, you won't like UT2k3.
3) If you like them both (like myself), you'll like UT2k3.

[Nick[FM]]

Hummm.. I read that UT2003 "supports" Truform? How do I enable it? I'm too lazy to start looking it.. :wink: Anyone?

[Ozymandis]

]Hummm.. I read that UT2003 "supports" Truform? How do I enable it? I'm too lazy to start looking it.. :wink: Anyone?

I believe you need to go into the advanced options. Pull down the console and type "preferences". There's a setting there that is called "N-patches". Isn't that Tru-form?

[borzwazie]

edit the ini file, you'll see a line referring to TruForm

[cellarboy]

Funny, reading some of the comments on UT2003 really summed up what I was thinking. The game really doesn't 'feel' like UT. I liked Quake3, but preferred UT and this game just doesn't play as well. For some reason too, the graphics were making me feel nauseous while I played. Strange.

Oh well, at least they fixed the S3TC problem on UT. Back to playing that I guess..... :cry:

[Nick[FM]]

edit the ini file, you'll see a line referring to TruForm
I didn't find any mention of the "TruForm", but "UseNPatches=False" was there. I'll turn that to "True" and see what happens. I hope the difference is noticeable in the game!?

Thanks for the info!

[jb]

Sorry I think the game rocks. I loved UT and thats why I am coding mods for it. But UT had lots of short commings. I think this helps to get ride of some and may introduce new ones. People dont like change but change is excatlly what happens in todays worlds. Again the detail on the Anubis map rock. And the rag doll effecs make it oh soo much fun....

They make use of projectors to help with some really cool lighting effects. The particle system them have is amazing!

The weapons are not balanced yet but thats why they released the demo to get feedback....

What was missing is the water effects. You were suppose to be able to shot the water and see ripples. Or walk in water and see ripples...

[Doomtrooper]

Jb...like I said it looks pretty..gameplay doesn't..something this Demo is lacking. The models GLIDE over the terrain...the model physics are not as realistic IMO...Rag Doll effect is over done IMO.
I hope the final product is better...

[Doomtrooper]

Worm...

[D3DDrv.D3DRenderDevice]
DetailTextures=True
HighDetailActors=True
SuperHighDetailActors=False
UsePrecaching=True
UseTrilinear=True
AdapterNumber=-1
ReduceMouseLag=True
UseTripleBuffering=False
UseHardwareTL=True
UseHardwareVS=True
UseCubemaps=True
DesiredRefreshRate=85
UseCompressedLightmaps=True
UseStencil=False
Use16bit=False
Use16bitTextures=False
MaxPixelShaderVersion=255
UseVSync=False
LevelOfAnisotropy=1
DetailTexMipBias=0.800000
DefaultTexMipBias=-0.500000
UseNPatches=False
TesselationFactor=1.000000 <---Sets your level of tesselation
CheckForOverflow=False
DecompressTextures=False
UseXBoxFSAA=False
DescFlags=0
Description=

[Nick[FM]]

Hmmm.. I enabled the npatches, but didn't notice any difference in the game, nor in performance. I wonder if it really works. Has anyone captured screenshots with it on and off? I'd do it, but I don't have any image editing tools here..

Edit: Thanks DT. I didn't touch the tesselation number as I wasn't 100% sure I had to. Now if someone could shoot up some comparison shots, it'd be nice! :)

[Dave Baumann]

I like the benchmark utility they put in. Kinda neat how it runs you through several scenarios.

I got:

Flyby: 95.820404
Botmatch: 42.343575

This is on an XP1600+, SiS 735, a GF3, Win2kSP3, Detonator 40.41's, 512MB ram, all non-o/c'd, max detail, 1024x768x32.

9700 + 2.53GHz P4, 1024x768

Normal
Flyby: 176.19
Botmatch: 60.02

4X FSAA / 16X 'Quality' Aniso (= Mmmmmmm) :
Flyby: 95.55
Botmatch: 56.42

6X FSAA / 16X 'Quality' Aniso (= Double Mmmmmmm) :
Flyby: 88.71
Botmatch: 51.91

:D

[Ozymandis]

UseXBoxFSAA?

I wonder what that means

[nggalai]

9700 + 2.53GHz P4 :D

4X FSAA / 16X Aniso (= Mmmmmmm) :
Flyby: 95.55
Botmatch: 56.42oh I hate you, Wavey. :D

ta,
-Sascha.rb

P.S. runs fine here (Ti4400, PIII-S, Win98SE), but when compared to Q3A, it feels less responsive to me. Not laggy, but more . . . hmm . . . wobbly. Sorry, difficult to explain.rb

[Dave Baumann]

UseXBoxFSAA?

I wonder what that means

Probably as it sounds - they have ported this engine across for the XBox (and other consoles IIRC).

[Doomtrooper]

Flyby: 176.19
Botmatch: 60.02

9700 + 2.53GHz P4 :D

4X FSAA / 16X Aniso (= Mmmmmmm) :
Flyby: 95.55
Botmatch: 56.42


http://www3.sympatico.ca/dimension/buttrock.gif

[jb]

Dave what res?

When I did mine at

1024x768 same settings all sliders maxed

I got 79.8 for flyby
43.4 for bot

thus is on a 2100xp+


Someone else askme what is a projector...here is an example:

http://www.pc-gamers.net/jb/reviews/R9700/Shot00005.jpg

Here you have a light soruce setting above the fan and the opening to this metal mesh. If you look you can how the light souce shinning through that fan and mesh make the pattern on the floor. Basicly the projector is taking the light and passing in through a mask that is simular to the metal patten. Not earth shaking but you can get some dam nice effects when you shine a light through stained glass for example ;)


Also the have a more efficent way to handle active lighting. I could kill your fps in UT with the use of few dynamic lights faster than with more polys. Here they have made some great strided to reduce this and will allow for my dynamic/active lighting...like when light bounces off water:

http://www.pc-gamers.net/jb/reviews/R9700/Shot00006.jpg

[Dave Baumann]

Dave what res?

Same as borzwazie - 1024x768

[Dave Baumann]

oh I hate you, Wavey. :D

:lol:

[Bambers]

4X FSAA / 16X 'Quality' Aniso (= Mmmmmmm) :
Flyby: 95.55
Botmatch: 56.42

6X FSAA / 16X 'Quality' Aniso (= Double Mmmmmmm) :
Flyby: 88.71
Botmatch: 51.91

:D


pfft :( Anyone got 300quid to give me? :D :-?

I couldn't see any difference with truform on or off. The graphics look nice but they almost seem too complex. Probably not used to this kind of detail in a multiplayer fps.

I wouldn't say the feel is like Q3, but its not like UT either. The pace seems a bit slower than both those games.

[Reverend]

Knowing Dave, the first thing he did was run the benchies. Hell, all he ever does nowadays is run benchies, make his house a better home, please the woman and his real day-time job.

Oh, the "I love it even if it hurts" or "I hate it even if it is good" conundrum of wanting to simply game while at the same time getting all those (many) great new toys!

[Humus]

So the general consensus is:

1) If you like Q3, you'll like UT2k3.
2) If you like UT, you won't like UT2k3.
3) If you like them both (like myself), you'll like UT2k3.

I think I'll fit in 3). While I seldom played Q3, I still liked the game anyway when I did. Still UT had a much more interesting gameplay.
I think the major reason some UT fans don't like UT2003 is that it's just different. The first time you play it you suck, regardless how good you are in UT. It just takes some getting used to. Have been playing some yesterday and some today and I'm getting more and more into it. The only thing that's disappointing worth mentioning IMO is the credit-based translocator, I think it was already limited enough from the 432 patch in UT. If only it reloaded a little faster it would be acceptable. Being able to quickly TL through the map to catch up with the flagcarrier is essential for CTF gameplay to be interesting.

[Tahir2]

The more I play it, the more I like it.

And I hated Quake3 Arena. But I'm weird.

I am at the moment wieghing the possibilities of returning my GF4 TI4200 and snagging an ATI Radeon 9700 Pro. Yea I am insane.

[Nick[FM]]

I couldn't see any difference with truform on or off. The graphics look nice but they almost seem too complex. Probably not used to this kind of detail in a multiplayer fps.
I didn't see any difference either.. :-? Oh well.. I just hope someone has the time & energy to take some comparison shots with it on and off (and with different tess values). I agree on the detail level too. The maps have such a high amount of detail that it is almost(!) distracting. I mean, in a single player game, such detail levels would rock, but in a multiplayer DM the detail actually is .. well, more or less disturbing/distracting!

I wouldn't say the feel is like Q3, but its not like UT either. The pace seems a bit slower than both those games.
Slower? Hmm.. I got the feeling that UT2003 is almost at the same speed as Q3A, while the original UT seems a bit more slow tempo. Just IMHO!

[borzwazie]

damnit dave! i'm getting married and can't afford a 9700 :)

I bet the Truform doesn't make so much of a difference because there's so many polys already.

Worm - look for the Adrenaline pills. They're fairly low-poly. See if you can spot the difference there.

edit - I should mention - the game looks so nice I keep getting distracted by the graphics and getting killed. almost too much detail maybe :) but it looks real good. I'm impressed at the performance difference over (ahem) earlier versions. This runs real nice.

[jb]

I think the gameplay reminds me more of Q2 than Q3...come on you all know you loved to double jump to get that super health pack on Q2DM1 The Edge :) (Of course RJ worked too...providing you had a RL). I liked Q2 style more than Q3 but thats just me.

[mat]

This is on an XP1600+, SiS 735, a GF3, Win2kSP3, Detonator 40.41's, 512MB ram, all non-o/c'd, max detail, 1024x768x32.


from the readme
3.6 NVIDIA 40.41 drivers
------------------------

The NVIDIA 40.41 drivers are known to have visual flaws and performance problems (hitching) with the Unreal Tournament 2003 Demo which are not present in the 30.82 drivers. The issue has been addressed by NVIDIA and future drivers released by NVIDIA will contain the appropriate fixes.

If you are currently running the 40.41 drivers please either downgrade to 30.82 or install newer drivers when available.

[Dave Baumann]

Knowing Dave, the first thing he did was run the benchies. Hell, all he ever does nowadays is run benchies, make his house a better home, please the woman and his real day-time job.

Oh, the "I love it even if it hurts" or "I hate it even if it is good" conundrum of wanting to simply game while at the same time getting all those (many) great new toys!

Oh, I've managed to squeeze in a few games.

[borzwazie]

that's odd - the game really does look just fine to me, 40.41's and all. I read that too. I can't spot anything like that.

I was noticing that there were problems with older versions of EAX, too, but it seems to work just fine on my C-Media 6-channel.

[Johnny Rotten]

If you're running a creative soundblaster live under windows XP make sure you are using Software Sound under the sound options tab. Setting it to hardware causes SIGNIFICANT peformance degradation. Creative Labs Live card is apparently a decelerator. :roll:

As to the actual demo, I like it quite a bit. Great visuals and very good performance. Its the first game in a long long time though to make me play in 800x600 to get smooth and steady framerates though. :) The 9700 is looking awfully tempting but I swore to myself no big overhauls until Doom. :wink:

[Bambers]

Hmm I've actually played on some full servers on asbestos now and it does seem as fast as q3 but Q3 had a smoothness that this doesn't have. Maybe its because UT2k3 is only running at 30-50fps or less. :-?

[crypto1300]

I liked both Q3 & UT very much and possibly like UT2003 demo even more. There are some weapon balance & gameplay speed issues, but I think they'll be ironed out for the final release.

Graphically, I can't really complain here. A step above UT & Q3 in my opinion, (awesome particle effects) but the demo is lacking some detail textures which again I think will probably be ironed out in the final.

[Bambers]

Demo textures only go up to normal size instead of the highest that will be used in the game to keep the download size down.

[Fuz]

Whats with the giant Nvidia logo at start up?
Says something about "the way it's meant to be played".

I am sure ATI are happy about that.

I was expecting more eye candy. Graphically, it was a let-down for me.
Still having a blast playing it though.

[Hellbinder]

It tells everyone who dounloads the game that for the *best* gameplay experience in Ut 2003 you had better own an Nvidia card.

Which is of course a DERN LIE. Unless there Is an Nvidia card out there right now, or when the game launches in a coupple weeks capable of this. These are at 1024x768x32 with maxx detail settings Trueform enabled in the system.ini and 4x FSAA + 4x Aniso

http://binderslair.freeservers.com/images/UT2003-3.JPG

http://binderslair.freeservers.com/images/UT2003-4.JPG

[alexsok]

From the readme:

3.6 NVIDIA 40.41 drivers
------------------------

The NVIDIA 40.41 drivers are known to have visual flaws and performance
problems (hitching) with the Unreal Tournament 2003 Demo which are not
present in the 30.82 drivers. The issue has been addressed by NVIDIA
and future drivers released by NVIDIA will contain the appropriate fixes.

If you are currently running the 40.41 drivers please either downgrade
to 30.82 or install newer drivers when available.

I checked the game under WinME with 30.82 detonators and under WinXP with 40.41 detonators on a gf4ti4600 and found no problems or whatsoever...

[Nite_Hawk]

Hi Guys,

I just downloaded the demo for linux and am using nvidia's newest linux drivers. I'm on a 1700+ athlon with a ti200. (though only sdram atm, as my board supports both and I haven't upgraded to ddr yet)

Here are some rough scores for on the outdoor map for those intrested:

1024x768 No AA No AF:

39fps Current
35fps Average

1024x768 2x AA No AF:

27fps Current
25fps Average

1024x768 Quincunx AA No AF:

22fps Current
23fps Average

1024x768 4x AA No AF:

17fps current
15fps average

1024x768 4x Guassian AA No AF:
18fps current
17fps average


And just for fun:

1024x768 4x AA "high" AF:
12fps current
10fps average

Nite_Hawk

[Gollum]

Got to spend a little more time with the demo, but here's my opinion so far:

I dunno about the graphics being "distracting", IMHO its really the first game besides Morrowind that has this kind of visual detail, especially when it comes to level geometry. I for my part am willing to bet that once we get used to the more detailed environments and the game itself, it will get a lot less distracting and at some point it'll be the same way it is with yesterday's and today's games. We've been playing in blocky low-poly worlds for so long (Half-Life anyone?), its almost shocking to suddenly get a much smoother and more detailed game world and at decent framerates to boot! Models and animations look good, but Doomtrooper is right about the models sliding over the surface, it doesn't look very convincing, they gotta tweak that...

About gameplay, I always liked UT over Q3. Although both could be fun in multiplayer, UT was the one with far more creative gameplay in my book and could even be fun in single play for a while. Deathmatch-purists may complain about lack of response in multiplay or weapon balancing, but I for my part am not one of those, I'm in it for fun, not competition and UT had lots of fun. How this translates to Ut2003? I dunno yet, right now it plays a bit worse then UT did IMHO, but it definitely has its strong points and might even turn out better, who knows, its only a demo and lacks a lot of content! Levels, gamemodes and MODs make or break these games, shame assault is gone. The Unreal engine is already a prooven development platform for high-quality MODs though, I can't wait to see what the first good MODs can make out of this toolset!

Oh Hellbinder, good bit about the Ad being "a DERN LIE", I really laughed hard at that one. But please do keep those shots coming, they look great... ;)

[Donald]

Might as well throw in my .02 here too.

I'm loving it so far! So much better than the original UT, which is didn't like all that much so it wouldn't be hard to improve on that one for me. Like the physics, like the animations and models. The level design seems very well done from what I've seen and I'm definitely looking to the full game.

I also think that we will be seeing some really awesome mods for this thing as well. A game like this is worth buying for the community support, especially if something like a DOD is pushed out using it :)

Just a quick note on performance too:

I ran the benchmark.exe program on the following system and received the following numbers:

P4 2.0a GHz
512MB PC2700 DDR
Asus P4S533 Mobo (645DX)
Audigy X-Gamer
9700 Pro

1024x768 - 126.44 flyby
1024x768 - 48.98 botmatch
1280x960 - 95.03 flyby
1280x960 - 48.96 botmatch
1600x1200 - 66.29 flyby
1600x1200 - 44.63 botmatch :o

Oh and those number come with all game options maxed out. Physics set to high and all that fun stuff.

As for video driver settings in control panel:

4x FSAA
16x Anisotropic Filtering
Highest quality textures and highest quality mip-map selected

9700 OWNS!!

[crypto1300]

Anyone having trouble using the "speech key", or is this feature not supported in the demo?

[Snap]

9700 OWNS!!

Heh. I'm wondering if this isn't a bit of an 'own goal' for nvidia. A demo out, with nvidia plastered all over it, that screams for a high end video card upgrade and whats the best on the market?

[Tahir2]

I wonder about the botmatch scores.

Almost everyone with a highend (AtlhonXP 2000+ or P4 2000) CPU and memory configuration seems to be getting about the same results.. about 45 - 55 fps regardless of GF4 TI 4xxx or Radeon 9700 Pro. Does this mean that the CPU is holding the game back.

Again? Like in Unreal Tournament?

Interesting.

[RaolinDarksbane]

You can change the pace of the game you know.... just slide down the slider from 100% to may be 80% if you want the UT physic. The weapons are still nerfed though.

[pascal]

Some layman´s preliminar subjectives impressions:
- the gameplay is good
- the level of detail is good too
- graphically it is a step better than Q3 or UT
- no stutering for me

My system P3-S 1.13GHz/256MB/GF3Ti200 playing smooth at 800x600x32 high detail.

[vogel]

Anyone got the game running without artifacts on a Voodoo5? (mark rein said it would run ok with a V5 a while back).
Sorry, I accidentaly broke that 10 hours before release by changing some high level code code related to handling lack of cubemap support which in turn triggered a Voodoo3 driver bug. D'oh! It currently is fixed and will be in a future update. Sorry about that.

BTW, I heard someone got the game up and running on a Voodoo Banshee.

-- Daniel, Epic Games Inc.

[vogel]

It is a massive improvement on the leaked 927 build. That build had a couple of bugs in it which would slow your whole system down after finishing with a map.

And it is pretty darn fast on my system at the moment.. but then again it should be:
P4 2.4GHz overclocked from 2GHz
GF4 TI4200 (to be replaced or repaired soon).

I would like the detal textures back.. scouring the .ini file right now :)

Personally speaking.. it is not UT. I think UT was better gameplay wise.
Detail texures are in the demo though without anisotropic filtering you won't notice them much. The full game has higher res textures (there is a reason we require 3 GByte of HD space) so it will be far less blurry.

-- Daniel, Epic Games Inc.

[vogel]

UseXBoxFSAA?

I wonder what that means

Heh, that probably should've been removed. It doesn't do anything on PC.

-- Daniel, Epic Games Inc.

[Jazz]

man i cant get this game to run on my Matrox G400 max :cry: Everytime i start up the a dark blue screen comes up and my comp crashes :(

[vogel]

man i cant get this game to run on my Matrox G400 max :cry: Everytime i start up the a dark blue screen comes up and my comp crashes :(
I haven't tried it on a G400 in a while. You could try the OpenGL renderer and see whether it works there.

-- Daniel, Epic Games Inc.

[Matt Burris]

Jazz:

This is in the readme file:

3.9 TNT2 / Kyro II / Voodoo 3 issues
-------------------------------------

Due to the lack of certain key functionality like texture compression and
cubemap support TNT2, Kyro II, Voodoo 3, G400, ... cards won't be able
to run the game at the full visual quality. As uncompressed textures are
much larger than the compressed ones we have to drastically scale down
both size and color depth of the textures in order to fit them into video
memory. This means you will notice banding artifacts - especially in the
menus.

I did hear that someone got the G400 to work by downloading DirectX 8.1b though .... may want to try that.

[Matt Burris]

By the way, just thought I'd say that I've had the most fun in a long time playing on Anubis/Bombing Run with 16 bots on Expert skill. Very very fun! :D

[Jazz]

ive got the service pack for Win xp and to my knowledge it has directx 8.1b already on it and it still doesnt work :(

vogel, ive tried the opengl renderer and i get this error message:

OpenGL renderer relies on DXTC/S3TC support.

History: UOpenGLRenderDevice::SetRes <- UWindowsViewport::TryRenderDevice <- UWindowsViewport::OpenWindow <- UGameEngine::Init <- InitEngine

[vogel]

ive got the service pack for Win xp and to my knowledge it has directx 8.1b already on it and it still doesnt work :(

vogel, ive tried the opengl renderer and i get this error message:

OpenGL renderer relies on DXTC/S3TC support.

History: UOpenGLRenderDevice::SetRes <- UWindowsViewport::TryRenderDevice <- UWindowsViewport::OpenWindow <- UGameEngine::Init <- InitEngine
Ups, forgot about that - yes, the OpenGL renderer requires the S3TC extension.

DirectX 8.1b is required as we trigger a couple of crashbugs in the DirectX 8.1 runtime... that was a nightmare but Microsoft has been *extremely* helpful.

-- Daniel, Epic Games Inc.

[Jazz]

i guess ill to have to wait till monday and post my problem in the matrox support forums as no one in the murc forums has posted a solution to it either :(

[Randell]

Gameplay (naturally) feels much smoother than build 927 and of course the bots AI is up and running properly.

Did you know you can run each part of the benchmark proggie separately. Go into the ut2003demo\benchmark\results folder and run the batch file for each map (flyby or bot). You then get the results saved in that folder with min/avg and high fps.

My results;

XP@1533mhz, 640meg DDRam, 64meg 8500LE @ 275/275

normal demo settings, 10x7, no AA no AF
Flyby - 110.3
botmatch - 42.6

no AA, 8xAF
Flyby - 110 again
botmatch - 42.5

2xAA, 16xAF
Flyby - 59
botmatch - 35

gameplay citadel on server - 30-60 fps with occasionally (1 or 2 times in a ten minute session) drops to single figures.

[Matt Burris]

I posted my results on my site, but in case you haven't seen them:

- Intel P4 2.1GHz
- 1GB PC800 RDRAM
- GeForce4 Ti4600 (325/730)
- Detonator 30.82 Drivers

Antalus Botmatch: 11.648977 / 35.513119 / 69.017975 fps -- Score = 35.530331
Anubis Botmatch: 19.529333 / 58.345604 / 120.366623 fps -- Score = 58.348648
Asbestos Botmatch: 21.981621 / 53.147606 / 103.991570 fps -- Score = 53.176178
Citadel Botmatch: 16.390404 / 42.461380 / 88.730461 fps -- Score = 42.526073
Antalus Flyover: 25.512714 / 129.651871 / 522.527954 fps -- Score = 129.720779
Asbestos Flyover: 25.837849 / 162.031189 / 464.948944 fps -- Score = 162.186371
Citadel Flyover: 18.936213 / 101.405579 / 293.091766 fps -- Score = 101.634811

[demalion]

As I said in my first post, the difference is noticeable. To really notice it, you need to play in a way that places you in close proximity with other player models, or you need to stop and look at items that benefit. For me, it is most noticable in the bomb ball (I really should verify that name) match. On the models, the twin feeder tubes in the back of the armor are the easiest place to spot it...they turn sharp corners with it off.

What I'm so pleased about is that it turns the model detail from "high poly" to "nearly perfectly curved" and it barely hits my fps (but that's probably due to the high physics and AI load)... IOW it appears to be about a perfect NPatch implementation (IMO).

I do have screenshots of some end game zooms to show the "perfectly" curved edges I mean. All graphical settings are high (the only thing not a high as possible is physics).

Hmm...perhaps in this case it would be better to use jpgs than pngs since I'm just showcasing edges...I'll link the images inline and offer links for the pngs.

Also note that these are at round end, so the CPU load should be off (D Vogel?), and these FPS being still lower than my flyby score would then seem to indicate that the player models are all rather highly detailed at their maximum level of detail (i.e., viewed close up)...well, the fps and actually looking at the models ;) .

Here it is with it off...some sharp edges, mostly not noticeable except at odd moments or in the model with the twin tubes I mention (sorry, didn't capture a shot of it yet):

http://home.earthlink.net/~demalion/Ut2k3tfoff1.jpg
PNG file (http://home.earthlink.net/~demalion/Ut2k3tfoff1.png)

Here is an atleast similar shot with it on:

http://home.earthlink.net/~demalion/Ut2k3tfon1.jpg
PNG file (http://home.earthlink.net/~demalion/Ut2k3tfon1.png)

Here is a different viewpoint with the calf armor receiving some lighting, and highlighting what I mean by my impression of "perfectly" curved.

http://home.earthlink.net/~demalion/Ut2k3tfon2.jpg
PNG file (http://home.earthlink.net/~demalion/Ut2k3tfon2.png)

My benchmark results were the same for both Truform on or off, both the flyby and botmatch...so either I made a mistake in setting those settings for the benchmark, or it is ignoring it. Or, of course, it might mean it doesn't matter...which means the fluctuation I saw in game between having it on and off was just variance in game play, because the botmatch scores were EXACTLY the same, and that Truform is only used on the models.

[Nick[FM]]

Worm - look for the Adrenaline pills. They're fairly low-poly. See if you can spot the difference there.
No effect on them either.. Am I the only one who can't see no difference? :cry: Do I need to go to the doc and have my eyes checked?

Vogel,

Will there be a TruForm/NPatch option in the game/video settings, or will it be in the Preferences menu only? Also, what's the "UseStencil" option in the preferences? :wink: Enabling it makes the screen go nuts for a second, but then everything looks normal again. Any hint on what it really does?

Now back to test some more..

[Dave Glue]

I wonder about the botmatch scores.

Almost everyone with a highend (AtlhonXP 2000+ or P4 2000) CPU and memory configuration seems to be getting about the same results.. about 45 - 55 fps regardless of GF4 TI 4xxx or Radeon 9700 Pro. Does this mean that the CPU is holding the game back.

Again? Like in Unreal Tournament?

Interesting.

Probably. With my XP1800, as long as I don't go above 1024*768, dropping below does very little for the framerate. This game does indeed need a hefty CPU, although the way prices are these days an XP2800 or PIV 3 ghz will be rather cheap in 6 months time. It took quite a bit longer for most UT rigs to get 60fps consistently.

[demalion]

]Worm - look for the Adrenaline pills. They're fairly low-poly. See if you can spot the difference there.
No effect on them either.. Am I the only one who can't see no difference? :cry: Do I need to go to the doc and have my eyes checked?

Vogel,

Will there be a TruForm/NPatch option in the game/video settings, or will it be in the Preferences menu only? Also, what's the "UseStencil" option in the preferences? :wink: Enabling it makes the screen go nuts for a second, but then everything looks normal again. Any hint on what it really does?

Now back to test some more..

Look at my shots above your post.

Hmm...9016 drivers in your 3dmark link. I had problems with a driver beta somewhere in that vicinity, but that was for stability and I don't remember any comments about Truform not working.

I assume that if you have the 9062 drivers you are aware of the the new Truform setting at the bottom? I think it defaults to "Always off"...just double check it is set to "Application Preference"?

EDIT: An extremely easy way to check the difference with Truform on/off is the 3D View button when picking your character model.

[Nagorak]

After playing with the Demo last night for a bit I am very dissapointed...this game is not Unreal Tournament...it looks and feels like Quake 4.
Physics are all wrong, model movement , powerups...everything about it takes away what made UT better than Quake 3..Shock Rifle is slow, Rocket Launcher is ruined...but it looks pretty like Quake 3..

I certainly hope they fix the Gamplay physics..almost all my Clan Mates agree..Quake 4 feel.

Agree 100%...very disappointed. :cry:

Shock Rifle was rather overpowered in the first one, but now it's so slow that it seems almost useless... I'd rather have it overpowered even though I wasn't a shock whore.

The minigun no longer seems to do any damage???? That used to be my best weapon, but now I fire away and it's like I'm shooting blanks. Edit: Ok my problem was they reversed the fire modes, so I was going around using the slow fire accuracy mode, LOL.

Flak secondary fire no longer kills on a direct hit...WTF! I never hit someone with the flak ball in UT without them dying, unless they had the shield belt or mega health!!!

Rocket Launcher...once again I felt it was somewhat overpowered in UT, but in UT2003 it just seems totally ineffective. The rockets are just too slow, and the weird spinning of multiple shots just makes it more or less useless.

The sniper rifle is now the stupid rail...I mean lightning gun, which has a projectile that BLOCKS YOUR VIEW!!!!.

The physics really need improvement...if you sit there and watch your corpse after you die, you never stop moving completely... There just isn't enough resistance in the model's joints or something. I'm sorry but in real life when someone dies their arm doesn't keep twitching indefintely, it hangs limp.

[Oompa Loompa]

I need to play some more (and try Bombing run, which sounds interesting), but it seems quite nice. The water is particularly gorgeous.

I expected to see more evidence of skeletal animation, but I really don't see much difference in the way characters move. Then again, I suppose a frantic run&gun game isn't the best place to appreciate subtleties of physics. I'm really looking forward to Unreal 2, DX2, and Thief 3, which should have many more opportunities to dazzle the player.

The bummer for me is that Assault mode is my favourite thing about UT, but UT2003 doesn't have it. Hopefully someone will do an Assault mod.

[demalion]

Hmm...perhaps a "UT2k3 things I don't like because they changed" thread should be started over in the Game Talk forum?

That way you could list not only things that changed, but why they actually do suck beyond your just having to adapt, and it won't be completely off topic (since this is a UT2k3 thread in the 3D Technology and Hardware forum)... ;)

*Ducks and covers in anticipation of incoming special smiley fire*

[Nagorak]

One other thing...I realize this isn't RL or anything...but why are the guys running around with 3 inch thick armor and a completely exposed head?

[Babel-17]

Hehe, perhaps a force field covers the head. They don't call it "Unreal" for nothing. Lol, I once posted about how strange it was to have both projectile weapons and a translocator as part of the weaponry. The energy used by a translocator could surely power something on the order of a phaser bank. :)

Hehe, I postulated that perhaps the translocator tapped into an energy grid that was left behind by an alien race and subsequently tapped into by Liandri Industries, I suppose the same could be said for all the energy weapons and with some really tortuous logic the whole system of respawning. I toyed with attempting a short story wherein Xan Kriegor in his youth was shipwrecked in space and near fatally injured when he stumbled on an asteroid that contained a prototypical Tournament setup. It was inhabited by several aliens, remnants of some quasi military organization and they had sublimed partially into a state that made them appear ghostlike. After repairing Kriegor they studied/tortured/toyed with Kriegor in a near endless series of death matches till finally one day they just weren't there leaving Xan, by default, ownership of the tech.

Perhaps driven half insane by those torturous matches, at any rate motivated, the once ne'r do well playboy Xan uses the tech to leverage his father out of the family business and subsequently establishes the Tournament on earth. Has he become like the Chistopher Walkin character in Deer Hunter seeking out the thrill of near death experiences or does he have other, perhaps alien, motives?

:) Lol, I think this post helps define "Off Topic". Feel free to delete. :)

[Gollum]

Below at least "skilled" level, the bots seem to be more in a chillout than killing mood - must be smokin' pot or something because they sure ain't trying to kill you very hard... ;)

Maybe this thread belongs in the Game Talk forum after all, eh?

[Humus]

LOL @ Babel-17's theories :lol:

Just play the game ffs :)

[Grall]

I played some insta-kill CTF just now (about half an hour's worth), and what strikes me is: this game is FAST. Bloody fast! Everything feels speeded up like 50% compared to old UT. Jumping, running, everything. Fast!

I like the fact you can't crouch-walk over a ledge and kill yourself, but disappointed here's yet another FPS where you're basically a disembodied head floating around. Look down, and there's nothing there! No torso, no legs, no feet. I really REALLY hope they'll give us a body in the future because not having one is major stupid.

Can't give much of an impression of the weapons, I've only played a little against bots and on low diff setting too so they were pretty much meat waiting to be harvested. Minigun does seem a bit ineffective, range of flakcannon is better, but still not good. Well, it's not really that type of weapon anyway, so no complaints from me.

Looks damn purty, but it's not next-gen as far as graphic's concerned. It's just more of the same, with exception of cubemaps, which I haven't seen used in a game before.

How come reflective surfaces don't reflect the players? Surely they could be rendered into those cubemaps as well. Looks kind of stupid when you're standing right in front of a shiny wall and you're INVISIBLE...


*G*

[fresh]

Runs sweet on my Gf4600 and 9700. I prefer playing on the 9700, cause I can crank up the aniso & AA.

[MuFu]

Has anybody else noticed that this benchmark totally destroys your vid card overclock?

I have my Ti4400 set to 310/700MHz upon bootup and the highest I normally run 100% stable for benchmarks, games etc is 315/720MHz. Well the highest I could complete the benchmark with no FSAA, no aniso and vsync off with was 300/650MHz...

1024x768
Flyby: 146.46
Botmatch: 55.22

Any higher on the memory and I got a general protection fault (due to a texturing error) or a lockup in the Antalus Botmatch. Any higher on the core (this is bizarre...) and it caused a few pauses of about 4-5 seconds in the Antalus Flyby! :eek:

The highest I could complete the benchmark with FSAA or at 1600x1200 was only 300/550MHz (default memory clockspeed)...

1024x768 (4xFSAA)
Flyby: 68.61
Botmatch: 43.36

1600x1200 (no FSAA)
Flyby: 71.57
Botmatch: 43.29

That Antalus level makes really heavy use of quad-texturing. I bet that is totally raping the memory subsystem like no other game/benchmark has ever done before. As for the bizarre pauses I have no idea.

MuFu.

[Brent]

Ooops, a little late to the thread

I'll just toss this in here for what it's worth, heh, prolly already been said though as I haven't read the whole thread yet, forgive me:

UT2K3 Benchmarking Results

Gigabyte 8IHXP
P4 2.53B
512MB PC1066 Rambus
Radeon 9700 Pro Cat 2.3 driver

All default settings, No Overclocking, No AA No AF, all defaults

1024x768
Flyby: 186.7
Botmatch: 64.9


1280x960
Flyby: 160.2
Botmatch: 64.8


1600x1200
Flyby: 113.5
Botmatch: 62.7


seems the Flyby is fillrate limited, and the Botmatch is CPU limited

I'll do some overclocked results later

[Brent]

Just overclockd the CPU to 2.81Ghz (148x18)

1024x768
Flyby: 203.3
Botmatch: 72

1280x960
Flyby: 164.5
Botmatch: 71.7

1600x1200
Flyby: 113.6
Botmatch: 68

what's goin on there at 1600x1200 in the flyby, it's the same as the result with the cpu at 2.53Ghz, vid card limited? or still cpu at this rez

[Randell]

as you said. flyby fill-rate or bandwidth ( :O ) limited. Bots CPU limited.

[Chalnoth]

The more I play it, the more I like it.

And I hated Quake3 Arena. But I'm weird.

Same here.

UT2k3 definitely has none of the things that really made me dislike Quake3, though it does remind me of it.

Here are some of the things I didn't like about Quake3:

1. Rocket launcher: the splash damage was almost nonexistant in Q3, and when you hit somebody head-on and gibbed 'em, it sounded more like an egg launcher. UT2k3's rocket launcher definitely sounds more solid, and has respectable splash damage.

2. Gameplay flexibility. With only single-fire weapons and few power-ups, Quake3 was just boring to me. The alt-fire modes in Unreal and Unreal Tournament were always things that I really liked (particularly the shock rifle...and the shock rifle combo effect for 2k3 looks awesome!), and I think adrenaline really adds something to the game (Similar to UT's relics, after a fashion). My belief is that for any reasonably-skilled player, options are always good. But, I imagine it makes it a little bit harder for the newbie figuring everything out. Speaking of which, I should probably see if I can figure out a place that describes all the possible adrenaline combos.

3. Game modes! One of the reasons I never got into Quake3 was the fact that it shipped originally with nothing but deathmatch.

[Rodéric]

Slow and laggy.
Commands don't respond fast enough (far slower than Q3 and slower than UT)

Cannot use my 4th and 5th mouse button
Couldn't find how to use Adrenaline.

Athlon 700 256MB SDRAM133 + ATI RadeOn 8500
everything to low or minimum, 800*600*32 and it still don't find it perfectly smooth.

I don't like the new interface that much, couldn't find how to activate the console, change the Renderer (except in the ini, of course), neither how to make it show the players stats or the fps...

Very disappointed by this demo.
It's nowhere near what UT was.

(For Info I like both Q3A and UT very much)

[Chalnoth]

Slow and laggy.
Commands don't respond fast enough (far slower than Q3 and slower than UT)

Your CPU is a bit on the slow side for this game. I would suggest turning down a lot of graphics settings. Turning off mouse smoothing may also be nice.

Cannot use my 4th and 5th mouse button
Couldn't find how to use Adrenaline.

Can't help you with the mouse buttons, but Adrenaline is based on combos. For example, press forward a few times quickly to execute the speed combo (increased speed until adrenaline runs out). I've also managed to get an invisibility and health combo (I'm not entirely certain what invisibility was, but I think health was just pressing back a few times...).

I don't like the new interface that much, couldn't find how to activate the console, change the Renderer (except in the ini, of course), neither how to make it show the players stats or the fps...

Changing the renderer is only in the .ini. Just press ~ to activate the console, or tab to type in quick commands.

[Chalnoth]

Ah! Found a thread that describes the adrenaline combos:

http://www.ina-community.com/forums/showthread.php?s=7961ea866c30640480d4e709b5624bc9&threadid=204322

[Nagorak]

For what it's worth, I played the game more and now I'm loving it. It wasn't what I expected and my first impression was bad, but once you get the hang of it, the game is really fun! I take back most of my complaints (honestly, the reverse fire modes on the minigun really screwed me over, hehe). Bombing Run is also great!

So if you're not liking the game, my advice is just to give it some time. After two or three hours of play, you'll probably start to like it. If not, you can uninstall it. :wink:

But anyway this is a 3D "technology" forum so we can't talk about game play. :roll: (Edit: it's been moved so no problem!) So, in that vein, the game runs damn well on my Radeon 8500...granted not with everything maxed and with no AA or Aniso, but still damn good. Detail textures had to be disabled though, because they caused some weird shimmering errors. I think detail textures are overrated anyway though. :)

[Nagorak]

Changing the renderer is only in the .ini. Just press ~ to activate the console, or tab to type in quick commands.

Is there any reason to change renderers? Does OpenGL run better at all?

[Chalnoth]

Is there any reason to change renderers? Does OpenGL run better at all?

I don't know. OpenGL felt a little bit smoother, but I don't think it uses all of the same graphical options. I think OpenGL is mostly there for people that have trouble running in Direct3D mode (aside from the obvious portability part...). Given that a Linux server is emminent, I would imagine a Linux version of the game will be released, and so the OpenGL renderer should eventually be updated.

[Nick[FM]]

Hmm...9016 drivers in your 3dmark link. I had problems with a driver beta somewhere in that vicinity, but that was for stability and I don't remember any comments about Truform not working.

I assume that if you have the 9062 drivers you are aware of the the new Truform setting at the bottom? I think it defaults to "Always off"...just double check it is set to "Application Preference"?
Heh.. The project is OLD. I have installed the latest 6166 driver (and swithced to XP) installed and the TruForm is set to "application preference".

Now after a good nights sleep, I'll do some more testing! :wink:

[Nagorak]

How come reflective surfaces don't reflect the players? Surely they could be rendered into those cubemaps as well. Looks kind of stupid when you're standing right in front of a shiny wall and you're INVISIBLE...


I think this is more of a situation where other games over-do reflective surfaces. If you stand in front of a gold plate, yeah it's shiny, but unless it's really buffed you can't see your reflection in it. It's pretty unlikely people would go around buffing up the walls in a bloodsport arena.

[Rodéric]

yep, after the post I did some search an found the Adrenaline combo too.
:)

I switched the sound to failsafe or something like that (last option), and it works much better, but since I also changed the renderer to OpenGL, I'm not sure which is responsible for the huge speed up.

the game is fun when it runs fast :)

I feel like the character are dying in slow motion, weird, but the physics are rather cool :)

Still slow when on the top of a tower in the capture the flag map.

Something weird with the water, looks nice but no splash sound I remember off, neither particles or ripples when I fall from high in it :(

[Sabastian]

Well from the little that I have played the game I find it is considerably better then any other FPS I have ever played. Graphics are gorgeous really, I think Epic has done a fine job. Most bugs I would imagine will be worked out before the final version is out. My Geforce DDR is lacking though and more then I thought it would. To get decent playability I have to over clock. I notice that there are some corruptions though. I am pretty much forced to play at the lowest res so that I can use AA with acceptable frame rates but still a bit choppy. I notice the water is pretty but no movement effects it.(Not a big deal really.) The game play is considerably faster then UT and it does remind me of Q3 to a degree in terms of how fast the game play is but that is where I would stop the comparisons between the two games. Graphically it is superior to any game I have ever seen.(actually played on this machine.) It kind of reminds me of the multiplayer aspect of Halo but doesn't run nearly as smooth on my PC.... sigh. This game looks like it should outshine UT in a massive way. BTW I wonder how much nvidia paid to have their logo plastered on the opening of the game? LOL "the way it is ment to be played" never mind that it runs the best on the Radeon 9700... argh.(Wishing I could manage this card. Even If I was to buy it I would have to upgrade my whole system..) I tried the OpenGL hack and got about 3 frames per sec.... needless to say I was a bit disappointed. Anyhow I am looking forward to the full version and other maps it really does look like a decent game.

[easyride]

I'm disappointed. The game is just too slow. In-game I'm getting about 30-40 fps average. No matter how many settings I turned down or off, only DM-Asbestos could be considered playable.

I do meet recommended specs
Athlon XP 1.47 GHz
512 MB PC133 RAM
GF2 GTS 32 MB
SB Audigy

I tried both Win98 and 2000, with different drivers. I must be limited by CPU and/or system bandwith, since even 800*600*16 with bilinear and really small textures hardly helped.

Looks like I won't be buying the full game for a long time. UT with high res textures is better in every way that matters.

[demalion]

But anyway this is a 3D "technology" forum so we can't talk about game play. :roll: (Edit: it's been moved so no problem!)

I'm actually more than slightly peeved that after several requests to curtail the gameplay discussion, that my discussion of UT 2003 as pertinent to the "3d Hardware and Technology" forum is instead curtailed by the thread being moved, with no "Moved->" indicator no less. I just now found it again, I had thought it had just not been posted to in a while.

I guess this week is "only rudeness gets results" week?

Oh well, I guess I'll just keep referring people to the screenshots I wasted my time posting here whenever it crops up in the "3D Hardware and Technology" forum, eh? :evil:

[BRiT]

AMD Thunderbird 1.58Ghz [143*11]
ATI 8500 64M [XP-6159 drivers]
16x AF + MAX Quality Texture/Mips
SB Audigy
512M DDR Memory

Running the game with maximum details cranked, with D3D+EAX sound, the gameplay is extremely smooth at 1024x768-32. At 1280x1024-32 it's extremely playable, but it doesn't feel as smooth as 10x7. I've been playing at 12x10-32 though.

The "benchmark.exe" results are:
1024x768-32 : flyby 111.84, bot 44.23
1280x960-32 : flyby 78.42, bot 42.02

The game runs so well at either resolution, that I don't feel the urge to even upgrade my aging system (18 month old cpu, 10 month old vid card). I used to upgrade parts on a 5-6 month schedule. Now don't get me wrong, I'm still lustful over the ATI-9700...

--|BRiT|

[Tahir2]

yea i know how u feel about that lust

like a teenage boy again i feel.... and to temptation i have fallen

my wife.. kill me she will...

d'oh!

[Nagorak]

I'm disappointed. The game is just too slow. In-game I'm getting about 30-40 fps average. No matter how many settings I turned down or off, only DM-Asbestos could be considered playable.

I do meet recommended specs
Athlon XP 1.47 GHz
512 MB PC133 RAM
GF2 GTS 32 MB
SB Audigy

I tried both Win98 and 2000, with different drivers. I must be limited by CPU and/or system bandwith, since even 800*600*16 with bilinear and really small textures hardly helped.

Looks like I won't be buying the full game for a long time. UT with high res textures is better in every way that matters.

I think you're both CPU limited and graphics card limited. That GF2 GTS is over 2 years old, which isn't helping you. But I noticed a major performance increase going from 1800+ to 2100+ (overclock, actually spees 1533 and 1725).

[easyride]

Well, a new graphics card is on the top of my upgrade list, but I don't think it will help much.

It's no big loss. I'll just wait for Unreal 2 and Deus Ex 2. 30-40 fps is adequate in slower paced single player games.

(edit)
Hmm, I guess I'm more graphics limited than I thought. After more tweaking I'm now mostly in the 30-50 fps range. In the heaviest scenes it still drops to the 20s, though.

[Brent]

I have some questions so far about some settings in the ut2003.ini file, i'm not sure what they do.

all under

[D3DDrv.D3DRenderDevice]

DetailTexMipBias=0.800000
DefaultTexMipBias=-0.500000

UseStencil=False

(by enabling stencil what is actually happening, because it reduces my framerate by about 10fps on a gf4 ti 4600)

MaxPixelShaderVersion=255

CheckForOverflow=False


if anyone can provide info on those settings I would greatly appreciate it!

[jb]

Guys,

keep in mind UT2k3 is a joint effort. In fact Digital Extremes did about 3/4 of the game content. Epic did about 1/3 of the maps, tweaking, all AI and other core functions. Also the demo was designed to get some feedback. So please if you have some constructive feedback let them know about it...

[Nagorak]

Well, a new graphics card is on the top of my upgrade list, but I don't think it will help much.

It's no big loss. I'll just wait for Unreal 2 and Deus Ex 2. 30-40 fps is adequate in slower paced single player games.

(edit)
Hmm, I guess I'm more graphics limited than I thought. After more tweaking I'm now mostly in the 30-50 fps range. In the heaviest scenes it still drops to the 20s, though.

On a Radeon 8500 this game kills my performance on Antalus. It's actually fairly ridiculous... I can go through Asbestos and come in first with like 25 kills and 2-5 deaths, then I hit Antalus and with my crap 30 fps in firefights end up somewhere in the middle of the order with kills=deaths. I'm very unhappy with the performance discrepancy between inside and outside levels...they really need to institute separate settings for both inside and outside maps or even per map, because I don't want to run around Asbestos with everything turned down and blurry (when I don't need it), just so Antalus runs at a decent speed.

I'm disappointed. The game is just too slow. In-game I'm getting about 30-40 fps average. No matter how many settings I turned down or off, only DM-Asbestos could be considered playable.

I do meet recommended specs
Athlon XP 1.47 GHz
512 MB PC133 RAM
GF2 GTS 32 MB
SB Audigy

I tried both Win98 and 2000, with different drivers. I must be limited by CPU and/or system bandwith, since even 800*600*16 with bilinear and really small textures hardly helped.

Looks like I won't be buying the full game for a long time. UT with high res textures is better in every way that matters.

I think you're both CPU limited and graphics card limited. That GF2 GTS is over 2 years old, which isn't helping you. But I noticed a major performance increase going from 1800+ to 2100+ (overclock, actually spees 1533 and 1725).

Strange I get the same fps with a much lower system running in 800*600*16
Athlon 800
384 MEG PC100(compatible till 166)
Geforce2 400 mx 64MB

So you should get much higher fps then I get which is avarage 30-40 with a max off 60 now this is way to low for online matches..........

I get 80-100+ avarage fps on almost every map in UT

Most dissapointing is footsteps sounds to weak or not at all (alot not always) taunts to unclear, instagib sound of the enhanced shockrifle is just way to weak and the beam looks like crap(if it where fired with a 640*480*8 res.) Movement improved alot and the soundtracks for the maps is awsome..... Offline I can run it with all specs set to high in 1024*768*32 and play an acceptable botgame..... so even with lower fps the game just looks much smoother then ut and also the code for mp is improved a lot on servers with even 32 players I didn't experience no lag unlike ut these days where it seems lag is now part of the game.....

[easyride]

I run at 800x600x32 with settings at normal, all effects but dynamic lighting and decals off. I have also set reduce mouse lag to false, as it gives me a nice fps boost. With bots I get 30-60 fps, very rarely dropping just below 30.

It's fine for bot matches, and it's faster without bots. I would like to stay above 40 at all times, though.

[Thowllly]

For ppl with Creative soundcards; switching to software 3d sound might give a nice fps boost.

[vogel]

For ppl with Creative soundcards; switching to software 3d sound might give a nice fps boost.

There is a reason the game defaults to software 3D sound and everyone at Epic uses it ;)

-- Daniel, Epic Games Inc.

[Brent]

one more question for you Vogel if you dont mind

what is the difference between the lowframerate.log file and the benchmark.log file, they both appear to give the Same information except the lowframerate.log file organizes it better by showing the map name and putting the total score under each test

and also in that same file lowframerate.log is it grouped by Min frame rate/Avg. Framerate/Max Frame rate and then the overall Avg. score?

here's an example:

dm-antalus
28.693314 / 43.461670 / 165.342514 fps
Score = 43.480888

I take it the 28.693314 is the Min frame rate the flyby achieved, and then next is average and then next maximum frame rate? If so why does the Score (43.480888) and the midle Average (43.461670) differ?


And last question, is there anyway to edit something so that when you run a series of test from a batch file with the resolution switch to test at say three different resolutions, is there a way to make it so it adds what resolution the test was run at inside the lowframerate.log file?

[Galilee]

Great demo, lots of improvement over UT which I loved.
Beside the weapons being more accurate I don't see the connection to Q3.

Gameplay is something that grows on me. I thought it was medium fun the first try, but now after playing it some days (10+ hours) I fucking love it. Have set up a 16player ctf-server and it's full all the time :)
All the new stuff like wall-jump, double-jump together with dodging will make this a game a pro game I think. Lots to learn to be good.

Graphic is great and I can't wait for the new textures :)

Athlon 2100+
GF4-4200
1024:

FLyby: 131
Botmatch: 50


Daniel: Send a big thanks to the people at Epic from me.

[alexsok]

Yep, I agree with Galilee, GREAT game!
And the best part of it is that the full version has gone gold and will be available by the end of this month!

[Nagorak]

For ppl with Creative soundcards; switching to software 3d sound might give a nice fps boost.

There is a reason the game defaults to software 3D sound and everyone at Epic uses it ;)

-- Daniel, Epic Games Inc.

Hasn't Nvidia run Creative's SB department out of business yet?

[easyride]

Hasn't Nvidia run Creative's SB department out of business yet?

If they haven't, I seriously think they will. Now that I'm not very graphics limited (various effects off), software audio gave me a boost of about 10 fps in-game!

Needless to say, I'm no longer very happy about my Audigy...

[Tahir2]

I cannot even get EAX to work on my Audigy.

Seriously I cannot recommend this card to anyone.. the amount of hassles it has caused me is way too much and outweigh any benefits.

Back to UT2003, I dont know what I did but suddenly it performs a lot faster (40-80fps) .. so much faster I have 4x fsaa and 16x aa enabled at 1024*768 :)

[John Reynolds]

http://jedi.home.attbi.com/ATILogoReplacement.zip

For those 9700 Pro owners such as myself who don't want to see an Nvidia logo every time they fire up the demo. 8)

Just extract to your UT2003/textures folder, overwriting the existing file, and you're good.

[CosmoKramer]

The slowdown is Epic's fault - the game uses an OpenAL emulation driver as default instead of the hardware accelerated one (OpenAL32.dll in \windows\system\). There are a couple ways to fix that. Hopefully the full version will fix that.

[vogel]

The slowdown is Epic's fault - the game uses an OpenAL emulation driver as default instead of the hardware accelerated one (OpenAL32.dll in \windows\system\). There are a couple ways to fix that. Hopefully the full version will fix that.
Seeing that we use Creative's OpenAL->DS3D wrapper this just means that Creative's DS3D drivers are "suboptimal" ;)

You are right that the native OpenAL drivers Creative ships with perform better but they have a couple of minor bugs... and lock up my machine from time to time if I use them.

Change UseDefaultDriver from true to false to use the native driver if you have an Audigy and the latest drivers installed.

-- Daniel, Epic Games Inc.

[Galilee]

I noticed on Creative's last driver:
*Fixed hardware accelleration of OpenAL (not excact words).

Will this help any?
(the driver from 26/8-02)

[Sabastian]

http://jedi.home.attbi.com/ATILogoReplacement.zip

For those 9700 Pro owners such as myself who don't want to see an Nvidia logo every time they fire up the demo. 8)

Just extract to your UT2003/textures folder, overwriting the existing file, and you're good.

I think that is pretty cool. IMO it even looks a little better.

http://216.185.114.72/images/news/1032991313QTmqfjeyfc_1_1_l.jpg

[renmike]

I find it odd that people are lame enough to give a damn either way about the intro movie....

I mean, get over it....do Pepsi drinkers get pissed when their favorite character on Friends drinks a Coke?

[Gubbi]

The nVidia logo in the beginning cracked me up too when I upgraded my ti4200 to a R9700.

Now running everything maxed with 4 x FSAA and 16X AF, and it is *still* noticeable smoother than the ti4200 was. The R9700 is the most exciting piece of graphics hardware I've owned since V2 SLI :)

Cheers
Gubbi

[alexsok]

On a related note, Anandtech report the following:

UT2K3 Benchmark v1.001 @ [H]
September 26th, 2002 11:34 PM - Kiran Venkatesh
Source: [H]ard|OCP
Category: Other


[H] has already posted a UnrealTournament 2003 v1.001 benchmark after some retooling with some help from NVIDIA. Download it here. Here's a clip:

Working with NVIDIA, we have made a single change to our high quality UT2K3 benchmark settings. It is actually a pretty simple change but seems to have an adverse on High Quality score on cards such at the 4200 and "lower". The changes made should not affect visual quality in any way shape for form. We have verified the information given to us by NVIDIA with EPIC Games.
Brent will be posting a follow up to this as well. Thanks to NVIDIA and the guys at EPIC for helping us make it a better utility and setting us straight. Hell, I am sort of surprised we only got one correction. Brent deserves props as well for good work.

The updated benchmark can be downloaded here:
http://hardocp.com/files/ut2k3_demo_benchmark_v1001.zip

[Brent]

On a related note, Anandtech report the following:

UT2K3 Benchmark v1.001 @ [H]
September 26th, 2002 11:34 PM - Kiran Venkatesh
Source: [H]ard|OCP
Category: Other


[H] has already posted a UnrealTournament 2003 v1.001 benchmark after some retooling with some help from NVIDIA. Download it here. Here's a clip:

Working with NVIDIA, we have made a single change to our high quality UT2K3 benchmark settings. It is actually a pretty simple change but seems to have an adverse on High Quality score on cards such at the 4200 and "lower". The changes made should not affect visual quality in any way shape for form. We have verified the information given to us by NVIDIA with EPIC Games.
Brent will be posting a follow up to this as well. Thanks to NVIDIA and the guys at EPIC for helping us make it a better utility and setting us straight. Hell, I am sort of surprised we only got one correction. Brent deserves props as well for good work.

The updated benchmark can be downloaded here:
http://hardocp.com/files/ut2k3_demo_benchmark_v1001.zip

As has been explained to me now from Epic and NVIDIA the UseStencil has nothing to do with Stencil Shadows like I originally thought :p

I'm suprised myself this has been the only major issue hehe

I'm working on really learning the benchmark setup and how it works for the full version game...

Thanks again Vogel :)

[demalion]

You did this Brent?

If so:

1) Why did you hardcode the paths instead of using relative directory specification? UT2003 doesn't support it? EDIT: Well, a quick look at the files there shows it does support this.

2) Why didn't you have your settings files as newly named files, and then refer to them by name in the .bat file command lines?

[Brent]

You did this Brent?

If so:

1) Why did you hardcode the paths instead of using relative directory specification? UT2003 doesn't support it? EDIT: Well, a quick look at the files there shows it does support this.

2) Why didn't you have your settings files as newly named files, and then refer to them by name in the .bat file command lines?

i'm not a batch file expert, i'm a hardware guy, heh

this will not be the final version... i am going to get some nice batch files in there that will let you use it no matter where ut2003demo folder is located and it will make a backup of the .ini file's it replaces and then when the benchmarking is done copy the backup back in there so nothing changes... it will be all automatic...

there are a lot of changes to be done, i know that... and i am learning and working them out :)

I do appreciate everyone's feedback cause i couldn't improve it without them!