Kyro2 image quality in two popular games [Archive] - slackercentral.com messageboard

electragician

07-23-01, 06:00 PM

Alright folks... I made a post over on another forum today, and though I'm not going to repost the full contents over here as I was intending, I would appreciate it if you guys would go and read said post.

Feel free to reply back here in this thread, as I'm not trying to drum up business for the attached site's forums, but rather I'm looking for answers and didn't feel like retyping this time and time again on several sites. I'm steadily getting more upset with this card guys, and I tend to be noisy when upset.

Anyone else, please post in this thread if you've ever witnessed ANY game show this particular problem, and post a link to a screenshot if you can.
http://www.3dgpu.com/cgi-bin/YaBB/YaBB.cgi?board=101&action=display&num=995927795

Ailuros

07-23-01, 06:43 PM

I was able to reproduce some of the artifacts kincaid mentioned in conjunction with vertical or 2x2 FSAA (probably because of some "gaps" in the code) in two racing sims, but then I asked around and it turns up in quite a few cards using OGSS, while the v5 doesn't, but that's completely unrelated to what you mention here. I've seen one or two other fellow users in this forum mention something about "weird" texture behaviour on walls when you move real close up to them (like you showed in one screenshot) but I wasn't able to see any of the kind in the games I play no matter the settings.

In fact I don't have any of the games you mention and I'm not implying that it's game specific it might very well be some hardware/driver issue. Did you forward those to Kristof? Did he give you any anwer yet?

Well in any case if you should notice similar problems in other games, please keep us updated, I might then be able to be at least of some help with screenshots or trying to investigate why/where/how the best I can with my limited knowledge to come closer to maybe understand where it comes from....

Keep us updated :)

electragician

07-24-01, 12:54 AM

I can add this much Ailuros:

1) I'm running the latest versions of all drivers for my rig (no betas anywhere). All Herc and PowerVR driver versions exhibit this problem, across multiple reformats of my system disk.

2)None of the videocards D3D settings affect this problem... it's always there. I don't run FSAA in any games, as for me personally the framerate hit in the games I play doesn't justify the improved IQ.

3)Colordepth doesn't affect the problem at all.

4) The only game that I've noticed the screwy textures in is EQ, and the screenshots others have posted of CS doing it in some maps. I have, however, seen some texture blocks in Giants shake, bu they don't look distorted like the ones in EQ do.

5) I haven't personally forwarded these Pics to Kristof, and to be honest I had assumed that if the vendors play tested EQ then they must be aware of the problem. I'm pretty sure they play tested it, because they are aware of the large UV texture coordinate problem (shaking textures) with the game.

I have no idea if the corrupted textures are tied in with the coordinate problem, but if they are, then more than the EQ engine is evidently afflicted with said problem, as it looks to be the same problem found in some CS maps.

I should also state that although many zones in EQ look like ass due to this problem, the game is imminently stable with the card. I played for 6 hours straight tonight.

System -

Abit KT7A with latest non-beta bios
1.2gig T-bird (non-oc'd)
256megs PC-133 ram @ 133
SB Live
WinME
DX 8.0a
Latest Herc drivers

[ 07-24-2001: Message edited by: electragician ]

EgonOlsen

07-24-01, 01:30 AM

At least somebody who mentions the same problems as i did. I know what you are talking about and i really like your description 'large UV problem'. This makes perfect sense to me. Is this problem documented somewhere? There is a thread here on slacker called KyroII accuracy (somewhere below...don't have the link at the moment), where i was talking about the same problem. It's really annoying and i really hope that they are going to fix it in Kyro3. But as most reviewers don't have a clue about such things, it would be hard to find out without buying the card yourself when it's out... :(

electragician

07-24-01, 02:12 AM

As far as documentation of the shaking textures problem with EQ, Krisrof replied to a poster over on Hercules boards that it was a known problem and not much could be done about it. (Not dogging on you Kristof, as I realise that you are simply the bearer of bad news, in this case)
http://forum.hercules.com/index.php3?pg=readmsg&msgid=21559&forumname=Hercules+Video

I'm just wondering if the blocky corruption is also tied in with this.. I hope not, because I damned sure don't want to hear that there's nothing to be done about that either =(

I'm glad others have noticed the same thing as well, but sorry that they had to notice then in the first place.

One thing that I'm constantly amazed by is that some people are basically blind to graphics problems. They either don't notice them, or don't realise that what they are seeing is not what they SHOULD be seeing.

EgonOlsen

07-24-01, 02:42 AM

I see. Thanx for the link. At least they recognized it as a problem. You are right about the people are not being able to spot visual flaws very well. I don't understand that either but it's good for PowerVR ;)
After some thinking i would rather assume that the problem doesn't lie in large UV coords, but in very close ones and/or large triangles. Imagine coords like (0/0) and (250/250). Try to interpolate between them over a 1000 pixel wide span using an accuracy of 16bit (8bit.8bit fixed point format). You'll get (250-0)/1000=0.25. That will fit into your 8bit.8bit format without problems (being 0x00.0x40). But imagine a ground texture stretched over large parts of the terrain. For example, our poly is covered by a part from (0/0) to (5/5).
This will result in (5-0)/1000=0.005. The 16Bit accuracy is not enough to interpolate this texture correctly. You may either take 0x00.0x01 or 0x00.0x02, where the first delta would be too small and the second one too big. Altered use of them (because the polys are moving slightly and other accuracy issues come into play) may cause the so calculated texture to wobble.
I know that PVR doesn't do the texture calculations this way (because this is the way traditional renderers do it), but the idea/problem should be the same.

[ 07-24-2001: Message edited by: EgonOlsen ]

Ailuros

07-24-01, 05:27 AM

Egon,

I did remember the similarity in issues, I just didn't remember where it came from. Keep it coming guys and let's see if those issues can be worked out. Kristof is checking in here quite often ;)

electragician,

I know this is hardly of any help nor really relevant, but Sharkfood (yes the one :) ) showed some texturing problems with EQ on his GF3 some time ago over at B3D and was almost slaughtered for NVIDIA bashing. I would like to believe that on any board you're posting this very interesting piece of information, that no hardcore fanboy ninny will show up and accuse you of possible k2 bashing.

Issues are to be worked and most hardware has a few up to quite many, that depends. If users sit down and try to find where that comes from and why and ring an alarming bell to those who are responsible to look at it, then only improvements can come out of this.

*sigh* If you would only know how many people live in actual denial..... :(

Anything related I should come across I'll let you know :)

Kristof-PowerVR

07-24-01, 05:50 AM

OK, I have seen these posts in a number of places. Unfortunatly I am quite busy with Meltdown UK at the moment. We know about the shaking EQ textures and I have seen the "banded" textures but currently do not have a final answer on those.

I did get one email about these problems with HL but comparisions of the screenshots did not show much difference between multiple boards, the main difference in those turned out to be gamma where KYRO cam out as brighter than the others which made the bands more visible then on other darker setup cards.

So be patient I'll have a hunt about these issues during the next weeks.

K~

Ailuros

07-24-01, 06:08 AM

(How can you not like Kristof....he's trying hard ;) )

Ok I did a small quick experiment and I promise to bring more up as soon as I find more time. I quickly used UT (32bits-bilinear) and q3a 1.17 (32bits-trilinear) and captured 3 screenshots close to walls. If there's something I don't see or should have used other settings feel free to smack me on the head, I'm neither a hardware nor a software expert hehehehe.

The UT shots: one was captured with print screen and pasted into paint while the other with F9 and I adjusted the gamma quite a bit to make things more visible.

http://ailuros31.homestead.com/files/q3close.jpg

http://ailuros31.homestead.com/files/UTclose.jpg

http://ailuros31.homestead.com/files/UTclose2.jpg

Egon,

upon your "alarm" I've been paying attention to walls or textures even more in quite a bunch of games. Let me know if there's something I don't see in the above shots, because only then I will be able to recognize any abnormalities. (again I'm just an average gamer....).

Please note that those shots have a heavy amount of jpeg compression and I'm using win98se, 7.114 and dx8a.

Hope that helps at least a bit :)

EgonOlsen

07-24-01, 06:30 AM

Ailuros: I can't find anything wrong in the shots you've posted. But regardless of the filtering problems i got in HL and OpFP, the shaking of textures is present in more games, but it can't be captured in a screenshot. I don't know about UT but in DEUS EX, some textures are shaking too (not much, but it's there).
Try to move close to a wall and move your mouse very slowly. You'll notice the shaking if it is present. But as i said: I don't know about UT. Maybe it's all fine with it.
Do you have Operation Flashpoint? It's really worse in this game...it's impossible not to see it.

[ 07-24-2001: Message edited by: EgonOlsen ]

EgonOlsen

07-24-01, 06:41 AM

Originally posted by Kristof-PowerVR:

I did get one email about these problems with HL but comparisions of the screenshots did not show much difference between multiple boards, the main difference in those turned out to be gamma where KYRO cam out as brighter than the others which made the bands more visible then on other darker setup cards.

I'm not talking about the banding, but about the rendering accuracy. In the shots using pick-nearest, you should see accurate borders between the single texels (like the V5 shows). On the Kyro, this is not the case. The borders between the texel are looking ragged (don't know a better word for it). Move the mouse and the textures start to dance somehow.

kincaid

07-24-01, 09:10 AM

If this problem occurs in some games and not others, doesn't that mean it's a game specific problem? I've noticed it in Half-life but only when I went up to a wall. I've never seen it at all in Q3A.

What's a large UV coordinate problem? Is this a game specific problem? Should the programmers be using large UV coordinates in the first place?

EgonOlsen

07-24-01, 09:25 AM

Originally posted by kincaid:
If this problem occurs in some games and not others, doesn't that mean it's a game specific problem?

Not necessarily. If the chip is lacking accuracy, this may effect one game more than another.
It's like a person that can't handle numbers greater than 10. You won't notice it, until you are giving him a task (calculate 9+2 for example) where he reaches his limits. But you can't blame the task for this, because it can be easily be calculated by most other people out there. It's only a problem for this specific person. Now imagine this person to be named 'Kyro'...and there you have it ;)

07-24-01, 01:07 PM

I noticed that you're running dx8.0. Does EQ use D3D? If it does, upgrading to the latest build of dx8.1(currently 648) may fix your problem.

When I used to use dx8.0 with my Kyro2, I had some IQ issues in some games, but dx8.1 fixed all that.

electragician

07-24-01, 03:18 PM

Let me rephrase that... I'm running the latest non-beta DX version from Microsoft's site. I'll wait till the next version goes final before upgrading again.

I doubt that the DX version will have much effect, as this problem occurs in some games in both OpenGL and DX. It really appears to be more or less a hardware/driver issue with the Kyro2.

electragician

07-24-01, 03:28 PM

Originally posted by kincaid:
If this problem occurs in some games and not others, doesn't that mean it's a game specific problem? I've noticed it in Half-life but only when I went up to a wall. I've never seen it at all in Q3A.

What's a large UV coordinate problem? Is this a game specific problem? Should the programmers be using large UV coordinates in the first place?

No other card I've ever used with these games has exhibited these problems. Both CS and EQ work fine with a Voodoo2, Voodoo3, Voodoo5, TnT2 Ultra, Geforce MX and a Geforce2 GTS, in my experience. If the Kyro2 based cards are the only ones that can't do the math to get the images stable and the textures uncorrupted, well then it's pretty evident where the problems lie.

BTW... I hope that I'm not appearing as if I'm completely down on the Kyro2 (because I'm not, it's a decent card in other games). I do however expect the card to perform as well as traditional cards, or I expect a fix to be implemented to bring performance and IQ up to acceptable levels. Anything other than that is passing the buck. Passing the buck is something that the involved companies can't afford to do if they plan on staying in the retail vidoecard business for the longhaul.

The problems we're reporting are Kyro2 problems, as it's the only card showing said problems.

[ 07-24-2001: Message edited by: electragician ]

[ 07-24-2001: Message edited by: electragician ]

kincaid

07-24-01, 04:36 PM

EgonOlsen and electragician, I'd very appreciate it if you could post your thoughts about the FSAA issue. Now after this texturing problem I'm not convinced anymore that the FSAA problem is a game problem because maybe it's also an accuracy problem. Please reply in the "Kincaid concerning F1 C" thread. And if you are able to, please test out the F1RC demo with other videocards FSAA. Thanks.

Ailuros

07-24-01, 04:52 PM

Wohoooo easy here......we're not even close as to where these problems come from and you want to attempt to even come close to a conclusion?

Egon,

I'm a passionate UT player since it was released and I played my butt off with it on my former v3 and still playing it. If I would have noticed something wrong I have no problem at all to say so. I'm keeping a benefit of doubt because, since you brought it up I am indeed running close to a wall once in a while in several games, but I'm exactly playing the whole game glued to a wall, I do want to win once in a while too hehehehehe. I have a shitload of maps added in UT, those screenshots above were from Gothic and dm6 in q3a.

It would help if someone could point to a certain map, space etc etc where I'll be able for sure to notice it. I'm more than finicky with IQ and if you look closely to the q3 shot you'll see that's with 2x vertical on and the lodbias set to -2.

Let me list some of the weird findings I had so far with this card and had been reproduced by others too:

a) Lens flares shining through walls with FSAA on in Serious Sam.

b) UT for some weird reason allows TC enabled in the drivers tab with 32bits, making actually no difference for IQ or performance (while it does work in other unreal based games.....hmmmmmm....), yet crashes to the desktop when you attempt to run it in 16bits with TC on.

c) All UT maps with high resolution textures (f.e. DM Falkenstein) and games like Undying have some horrible texture flashes when FSAA is enabled, but can be controlled when setting 32bit textures and a lower mipmaplevel in the ini.

d) Several racing sims (f.e. NFS-PU) with vertical or 2x2 FSAA enabled show in the long distance on mostly road textures some weird lines or artifacts (appearing/reappearing for seconds or fractions of seconds). It doesn't appear with Horizontal FSAA on.

e) Heavy stuttering in games that use a lock in the code, where the game goes constantly lock-render-lock-render (example Grand Prix3)

Damn I'm sure I have collected more of those, but they really skip my memory at the moment. Oh yes I read the other day about an issue with ST Dominion Wars where the game has severe slowdowns and it runs only fine when the disable the text (wtf?) or something like that. Then there's always the linear "non-performance" in Giants in all resolutions, the known Tribes2 issues and of course Half-life/Counterstrike.

As I said I will spend some more time trying to find a spot where something similar occurs, but if you can point me to a spot in UT where it appears I'd appreciate it, it might save me a lot of time and effort. I have no doubt at all that there's something there, once the findings do not appear only on one machine, I just want to see it myself. I can't be that blind hehehehe or such a novice with those things.... :)

Ailuros

07-24-01, 05:19 PM

No other card I've ever used with these games has exhibited these problems. Both CS and EQ work fine with a Voodoo2, Voodoo3, Voodoo5, TnT2 Ultra, Geforce MX and a Geforce2 GTS, in my experience. If the Kyro2 based cards are the only ones that can't do the math to get the images stable and the textures uncorrupted, well then it's pretty evident where the problems lie.

Everquest worked horrible on the GF3 upon release, but I don't know if has been fixed yet. No problems there indeed with the cards you listed and you might add the Radeons too :)

Let's see the GF3 has some serious issues in some older D3D games.....would it be exageratted if I go out on a limb and would throw in as a possible option that older DX games have some "objections" with any form of HSR?

Take MBTR. Remember Anand's preview? Some minor mistakes in rendering (you can still find the article for reference), the newly released GF3 used to skip hole fractions of scenes in that demo, which I believe has been fixed now.....

Can any of you guys give the following registry setting a shot for D3D?

"DepthBufferLessHSR" "1"

I know I'm merely speculating...... :(

EgonOlsen

07-24-01, 05:39 PM

Ailuros: I installed UT again (just for you... :D ) and tried to produce some visual flaws. I couldn't. But i never claimed that i could do this. I'm always using OpFP, HL and DEUS EX as examples. In all three, the error occurs on large polygons and/or polygons covering just a tiny little bit of the whole texture. This is an accuracy problem to me as long as anybody gives me a damn good explanation why it isn't.

EgonOlsen

07-24-01, 05:48 PM

I just tried a very simple OpenGL-demo i found on my HD. It just renders a terrain with some textures stretched over it (not very detailed, but that's what i wanted...). No lighting, no dynamic LOD, no other fancy tricks. Just polygons, the texture and you...and it's the same here. Blocky filtering on some parts and shaking of the textures if you get really (and i mean reeeaalllyyy) close. :(

electragician

07-24-01, 07:37 PM

I put the DepthBufferLessHSR command in the global D3D properties and ran EQ for a while. It didn't seem to make any difference at all, for better or worse =)

Thanks for the effort, Ailuros.

Ailuros

07-24-01, 10:23 PM

Originally posted by EgonOlsen:
Ailuros: I installed UT again (just for you... :D ) and tried to produce some visual flaws. I couldn't. But i never claimed that i could do this. I'm always using OpFP, HL and DEUS EX as examples. In all three, the error occurs on large polygons and/or polygons covering just a tiny little bit of the whole texture. This is an accuracy problem to me as long as anybody gives me a damn good explanation why it isn't.

Hmmmmm maybe we could be getting close to a common baseline....Would it be too much to ask if I told you to give DM Falkenstein a shot? I haven't yet but it's a demanding map in textures and comes close to games like Deus Ex or Undying. I'm running out of time right now. I'll check every possibility I have later on today.....

PS: you shouldn't be able to find any of the kind in standard UT maps, if my suspicions are right it should be related with higher resolution textures......let's see ....

EgonOlsen

07-25-01, 07:29 AM

I may try that if time permits. But why are you so keen to find this bug in UT?

Ailuros

07-25-01, 08:46 AM

Because I would have to go out and actually buy all the other games you guys mention to be able to follow you. If you should run into any other games giving the same behaviour let me know. I just glued myself to walls in Serious Sam in 4 different maps. Nothing strange either.... I'm merely looking for a game you guys and I should have to see that behaviour first hand. It's not like I can get behind it in no time when I merely spend about an hour daily actually playing a game here. It took me several days to find where kincaid was pointing at and that trying several racing sims with variations of FSAA settings out.

I proposed Falkenstein cause it's an awesome map anyway and worth a shot, definitely fun to play ;)

EgonOlsen

07-25-01, 01:01 PM

Ok, i downloaded that map and checked it out. The problem is there, but not as evident as in OpFP or HL. Just stand still somewhere, look to the groud (i checked it out in a kind of cave and on the castle) and crouch/standup several times. You'll notice that the textures are shaking a little when doing so. This is most evident on the edges between polygons. You may have to crank up the brightness to notice it better (i had to). It's nothing spectacular in this game, but it is there...

<TdB>

07-25-01, 01:14 PM

the banding texturebug is also in quake 3 arena, in q3tourney2 (hunters map)it is most visible. if you stand by the lightning gun, and then zoom in on the very tall texture(looks like a tall skeleton), you will see it

Ailuros

07-25-01, 01:45 PM

Originally posted by <TdB>:
the banding texturebug is also in quake 3 arena, in q3tourney2 (hunters map)it is most visible. if you stand by the lightning gun, and then zoom in on the very tall texture(looks like a tall skeleton), you will see it

With a little help, some things get easier hehehe....

I believe you mean that spot here in the screenshot that follows. No texture compression on, 32bits/32bit textures, trilinear etc etc. I zoomed on that spot with the mouse, although it is already visible without zooming in. For comparisons sake can anyone find a screenshot from the same spot on alternative hardware? I can clearly see it now:

http://ailuros31.homestead.com/files/q3T2.jpg

Egon,

I tried it for over half an hour in Falkenstein with the brightness maxed out, crouching constantly up and down and I still couldn't see it there. I however asked for a friend to bring the Operation Flashpoint demo along to pop it and see if it's evident there......

More to come....

EgonOlsen

07-25-01, 02:00 PM

I'm sorry that you couldn't see it in UT, but some sports don't hurt... :p
As i said, it's not very obvious in UT. Anyway, i checked out the OpFP demo (in comparision to the retail version) and while it is present in the demo, it's not as evident as in the retail version. Something seems to be different in the retail version.
For HL, i made an animated GIF for you that shows the problem. It can be found at kyro-issue (http://www.jpct.de/kyro/). It's around 1MB in size, so be patient...

Ailuros

07-25-01, 02:03 PM

Hahahaha OT that was quick Egon....you're not sitting there all day refreshing that page are you? j/k

I need real time exercise badly not in a game but that's another story.

I'm hitting your gif right after adding this reply and I'll let you know.... :)

EgonOlsen

07-25-01, 02:16 PM

There you are:

chan530-a044.otenet.gr - -[25/Jul/2001:21:03:24+0200] "GET /kyro/kyrotexturing2.gif HTTP/1.1" 200 1116431

I'm not just refreshing slacker every millisecond, i'm also watching my logfiles the whole day :D

Ailuros

07-25-01, 02:22 PM

ROFL yeah :D

I popped it in Adobe Image Ready to have be able to study it better and I can clearly see what you mean. Gif's are far more helpful than still images :)

Hmmmmmmm PowerVR your turn....

<TdB>

07-26-01, 06:41 PM

the texturebanding is not present on a matrox g400 max, i can´t post any screenshots though, it IS a bug, i dunno if its a hardwarebug or a driver bug

Ailuros

07-26-01, 06:59 PM

Originally posted by <TdB>:
the texturebanding is not present on a matrox g400 max, i can´t post any screenshots though, it IS a bug, i dunno if its a hardwarebug or a driver bug

Makes me wonder, what keeps you posting a screenshot? The G400 is not an unreleased product and definitely not under NDA hehehehe ;)

Just hit F11 and convert it to a jpeg. Homestead is still offering free space to upload stuff, even for anonymous posters :)

Ailuros

07-27-01, 04:31 AM

Well since no one else bothered, I asked a friend to shoot me a Radeon shot in the same q3a location, with the exact same settings:

http://sirpauly.homestead.com/files/Ailuros1.jpg

TdB,

that doesn't really support your point does it, hm? If necessary I can ask for a GF3 shot too, but I hardly think there's a need to. I won't even start why and where the G400 lost the race against it's competitors in the first place but it's drivers where one of the weakest links back when it launched... Still an excellent 2D sollution ;)

[ 07-27-2001: Message edited by: Ailuros ]

<TdB>

07-27-01, 09:25 AM

well apparently i misunderstood the subject, i thought you were talking abuot this problem: http://212.227.228.93/pky/discus/messages/37/281.html?993811506

which isn´t visible on the g400 max, and yes, matrox hasn´t made a deacent 3d performer in 2 years, so i agree with you there, but it was the only other card i have, that can run quake3, i just mentioned it, because you were interested in other hardware.
if texture banding isn´t what i thought it was, what is it then??? :confused:

EgonOlsen

07-27-01, 03:56 PM

Actually, we are talking about both ;)

Ailuros

07-27-01, 04:46 PM

For what it's worth I think I did see something there when you pointed at the q3a spot. I'm not so sure about that particular case anymore. I'll keep my eyes open though from now on. Although it's quite stupid to constantly glue yourself like a Nun to walls or start "bunny-jumping" in games hehehehe I am actually doing it, whenever I find time for some gaming until I can see something...

What Matrox concerns, uhhhhm wait until they decide to make their comeback in the 3dgaming market. Despite their constant business-only orientation lately they have something impressive in the works as they seems... As for the G400 I was rather hinting on the lack of driver support when it launched (and quite a few motherboard incompatibilities) that sadly enough made it lose grounds against the S3/NVIDIA (GeForce256) offerings at the time.

I have the highest possible respect for Matrox because they have always been quite innovative on useful features......

Just a small rant to avoid misunderstandings :)

<TdB>

07-28-01, 07:28 AM

i also believe that matrox will make a comeback in the near future. as for driver support, i don´t renember any problems im my case, the g200 is a completely other story!!!(it couldn´t even do opengl)
the g400 still can´t play Glquake thuogh.

BTW what IS texture banding???
is it the way it blends textutres together, while revealing that 32 bit color isn´t that
great, or what?

the problem i was talking about was probably the kyro chipsets inferior dither quality, i really hope this is not the case. :(

<TdB>

07-28-01, 05:20 PM

I found another bug!

it is most visible in unreal 1 version 226, in the map "skytown" (medival city on floating mountain).

the sky animations are very jerky, just look up you can´t miss it, it is also visible(but not much) in UT, in the map "CTF-hydro16" that came with the UTbonuspack
(i am running these games in D3D)

it doesn´t have anything to do with the bug you are talking about, i just thought you would like to know.

does anybody on this board have any relations
to Powervr, appart from using it,I mean are Powervr reading this?

Dabson_Hykydo!~

07-31-01, 10:16 PM

[QUOTE]Originally posted by electragician:
Alright folks... I made a post over on another forum today, and though I'm not going to repost the full contents over here as I was intending, I would appreciate it if you guys would go and read said post.

Feel free to reply back here in this thread, as I'm not trying to drum up business for the attached site's forums, but rather I'm looking for answers and didn't feel like retyping this time and time again on several sites. I'm steadily getting more upset with this card guys, and I tend to be noisy when upset.

Anyone else, please post in this thread if you've ever witnessed ANY game show this particular problem, and post a link to a screenshot if you can.
http://www.3dgpu.com/cgi-bin/YaBB/YaBB.cgi?board=101&action=display&num=995927795[/ QUOTE]

Funny I have no problems running Tribes 2 it run smooth and runs well. Hmmmmm Maybe he missed something I set everything at full! and at 1024x768 32bit... I didn't try higher yet but that was what i was running and it ran great!

Dabson_Hykydo!~

07-31-01, 10:20 PM

Are those pics with your Voodoo5 ??Ailuros:
:p

Ailuros

08-01-01, 12:15 AM

Originally posted by Dabson_Hykydo!~:
Are those pics with your Voodoo5 ??Ailuros:
:p

I never had a v5 Dabby and the second shot is from SirPauly's Radeon.

electragician

08-01-01, 04:50 PM

Kristof,

If you are still keeping up with this, there's a guy thats posted a shot of Hidden and Dangerous over in Paraknowya's Buglist Submit forum... he also describes the same exact problems as EQ has (shaking textures and incorrect rendering of textures at close range).

I told him I'd try and get you to take a look at it, as I don't know if you are actively monitoring that forum.

Just a heads up,
Matt K

Wipeout

08-02-01, 12:11 AM

Electragician;

Just to let you know that I am working on a Kyro 2 review and all of these types of problems will be included. I'm kinda sick and tired of no one ever reporting what's wrong and only what is good.

If you have anything else to contribute please feel free to email me at wipeout@voodoosource.net

Thanks!

Ailuros

08-02-01, 02:04 AM

Hmmmmmm I was just fiddling around with NFS5 and UT a bit, and I finally got to see what Egon mentioned for UT. Funny though it appears only when triple buffering is enabled, which I actually never use for unreal engine based games........weird.... :rolleyes:

EDIT: by the way Wipeout, what's wrong with voodoosource.....Server again?

[ 08-02-2001: Message edited by: Ailuros ]

EvilEngine

08-02-01, 07:36 AM

For comparisons sake can anyone find a screenshot from the same spot on alternative hardware?

Sorry for the late post, here is some GF2 shots. If you will like high quality Voodoo5 shots too , let me know.

http://evilengine.apokalyptik.com/pix/shot0002.jpg
http://evilengine.apokalyptik.com/pix/shot0004.jpg

Kristof-PowerVR

08-02-01, 07:38 AM

Yeah, yeah, seen it... know it... don't know what to say about it at the mo...

K-

<TdB>

08-02-01, 07:47 AM

if you zoom in at that wall in q3a with a kyro 2, you will see the problem im talking about.

EvilEngine

08-02-01, 07:49 AM

Voodoo5 pics:
http://evilengine.apokalyptik.com/pix/shot0003.jpg
http://evilengine.apokalyptik.com/pix/shot0001.jpg

Ailuros

08-02-01, 11:44 AM

Evil,

thanks for the feedback. So we have on this page shots from a KYRO II, Radeon64mb, GF2 PRO and V5 5500. Can someone point to what the heck is there supposed to be wrong in that particular spot?

TdB,

instead of constant useless yadda yadda and flip-flopping I'd appreciate some actual proof. You were the one that pointed at that specific spot in the first place in Q3a. Care to elaborate? A screenshot would be more than helpful here.

At least Egon showed me what he meant and I was able to see it.....

08-02-01, 12:43 PM

i have no idea what is supposed to be wrong with those q3a shots, but it isn't anything to do with what the rest of us are trying to get to the bottom of

kristof,
i posted some more shots and a savegame over at paraknowya

Ailuros

08-02-01, 01:36 PM

deviant,

can you please check wherever it appears if you have the possibility to disable triple buffering. I was trying to get to it in UT and Egon stated that it's a bit apparent in DM Falkenstein.

Although I was crouching constantly up and down, nothing was there, until I reenabled triple buffering.

It's just a lucky guess and I don't know if it's related, but if someone can it might be worth a shot .......

08-02-01, 02:06 PM

i've included a shot of the settings panel here:-
http://212.227.228.93/pky/discus/messages/37/281.html?996771216
ufortunately triplebuffer is already off
there are no additional reg settings / a cfg file
16/24/32bpp all show the same issue

kincaid

08-02-01, 02:17 PM

What happened to the guys legs? Is that a bug?

08-02-01, 02:42 PM

he's crouching down you daft git! ;)

anyway, re/enabling triplebuffer did nothing and v/sync did nothing (didn't expect it to but there's not a lot of options to play with)

08-02-01, 03:05 PM

kincaid,
there's another piccy over @ paraknowya now, fer th' benefit of yerself and others!!!

electragician

08-02-01, 04:35 PM

Yup, it look like the screenies I posted over at 3dgpu from Everquest. Extremely similiar anyway.

In EQ I've tried EVERY option in the drivers and nothing at all helps the problem. Heck I've even installed W2K to see if it was WinME that was causing the problems, but no such luck.

Chiles4

08-02-01, 05:03 PM

:p Does anyone really know what this thread is about? :p

08-02-01, 05:08 PM

yeah,
it's about the stuff on this page which relates to the "incorrect filtering" topic at paraknowya - not the q3a stuff that hijacked the thread earlier

08-02-01, 05:22 PM

chiles4,
i'll explain the last couple of posts at paraknowya:
when kristof asked if he could see fog in my piccy, i thought he was onto something; a relationship between games that use fogging and the errors we are experiencing
the fog wasn't clear in my original pic so i did another one in which it (and it's errors) were more clear, then anotated it
kincaid however, may have inadvertedly brought something to light...
mainly that kristof was merely commenting on the appearance of the affected texture (seeing as it looks like the guy is standing in it) and not on the brink of some amazing revelation after all
my last pic-post at paraknowya was just to avoid any more confusion

kincaid

08-02-01, 07:51 PM

Originally posted by <deviantchild>:
chiles4,
i'll explain the last couple of posts at paraknowya:
when kristof asked if he could see fog in my piccy, i thought he was onto something; a relationship between games that use fogging and the errors we are experiencing
the fog wasn't clear in my original pic so i did another one in which it (and it's errors) were more clear, then anotated it
kincaid however, may have inadvertedly brought something to light...
mainly that kristof was merely commenting on the appearance of the affected texture (seeing as it looks like the guy is standing in it) and not on the brink of some amazing revelation after all
my last pic-post at paraknowya was just to avoid any more confusion

Please explain this again.