Monday, January 18, 2010

1.8.0 Beta is Out

The Slipgate Complex
Castle of the Damned
The Necropolis
The Grisly Grotto
Ziggurat Vertigo
Gloom Keep
The Door to Chthon
The House of Chthon

Unless your Quake engine can run those the way they are meant to be run, no matter how fast or pretty it might be, it ain't worth a damn.

1.8.0 Beta made it through this evening and so it is now unleashed.

I now consider the 1.8.0 codebase to be feature-complete, so start yelling about what you'd like to see in 1.8.1; there will however be one full and final release of 1.8.0 before that, which will incorporate any bugfixes required between now and then.

24 comments:

=peg= said...

Wow, out of the blue.. there it is! Gonna test it straight away :D

rovaman said...

Great to see another release :)

So far, only thing I've seen that's unusual is some missing weapon and model textures (!)

Took some caps for you:
http://i47.tinypic.com/24q8x82.jpg
http://i49.tinypic.com/2ufq1cz.jpg

Clean install, vanilla quake, no old cfg files, no additional pak or pk3 files present in id1 dir.

Any ideas?

Keep it up, I'm lovin this performance boost you gave the game!

mhquake said...

Odd that, the missing textures seem to be the ones that have fullbright colours. What's your 3D card?

rovaman said...

Here's a cap of my console at start...

http://i47.tinypic.com/312ee4n.jpg

So far,
I've see this prob on these things:
grunts
player weapons (except axe)
flames
lava chunks
grenades
pentagram
lightning bolts

Didn't have this behaviour on previous release.

Anything else I can do to help , lemme know.

mhquake said...

Definitely fullbrights, I need to check out that code and see what I'm missing. Can you check and see if it also happens on world textures, such as around slipgates? It might be useful to me if I can compare the two codepaths.

rovaman said...

So far, I haven't seen this on any world textures at all, just on models.

Slipgate textures seem to be fine here.

If I see anything else, I'll report back.

=peg= said...

I noticed a few things with r_novis 1:
Disappearing sky

Disappearing world surfaces:
screenshot1, screenshot2

Setting r_novis 0 again does not fix it however, not until a client-restart that is..

Btw, nice captcha :D

mhquake said...

I wouldn't rely on r_novis 1 to be stable; firstly I hadn't tested it, secondly it isn't really intended for regular gameplay use. I guess what's happening is that you're probably overflowing a VBO here, so in this case it's something that could happen in regular gameplay - so I'll see what's needed to fix it.

Isn't Marcher already vised for translucent water anyway?

I don't mean to sound ungrateful for the bug report by the way, it's more that I'm kinda amused (and bemused) at the obscurity of bugs you can pull out - who woulda ever thought of running r_novis 1 on an map that already has translucent water? Do you work as a software tester, cos you seem pretty good at it!

On the fullbrights thing I think I might have a solution. What I have in mind will require breaking my alias model batching a little, so I need to cross-check it with performance in some maps to ensure it's viable.

=peg= said...

Hehe, no I do not work as a software tester, but I use a button to cycle through various translucent water settings..

Opaque water:
r_novis 0
r_wateralpha 1

Translucent water (high fps):
r_novis 0
r_wateralpha .3

Translucent water (low fps):
r_novis 1
r_wateralpha .3


This comes in handy when playing online, where some maps are vis'd for translucent water and some are not..

For anyone interested: here is the setup I use for online fragging (proquake 4.00)

mhquake said...

@rovaman - try this one: http://www.sendspace.com/file/a2yg3s

rovaman said...

That appears to have cleared up the prob, mh. :)

=peg= said...

Something that is not 100% perfect (or at least something that glpro does better) is maps compiled with quarks ETP format..
I'm referring to light-maps in particular, compare this map in DirectQ vs glpro to see what I mean..

As for feature requests:

I'd love to see an option to make DirectQ render the viewmodel (gun in hand) translucent.
(r_gunalpha maybe?)

Normally I play with r_drawviewmodel 0 but since DirectQ renders the guns properly with high FOV settings, I prefer to use r_drawviewmodel 1

However, the viewmodel still blocks a small part of the screen, especially when aiming down.. Translucent viewmodel would help there ;)

mhquake said...

@rovaman: "That appears to have cleared up the prob, mh. :)"

Ahhh good. Marked for inclusion in the full release then.

@=peg=: I'll need to check out the ETP thing. I remember we briefly touched on it maybe this time last year. I'm not aware of any specific differences between DirectQ and ProQuake texcoord generation (aside from general code structure) though. One possible cause is that DirectQ uses low-precision floating point (it's a D3D thing); I can possibly work around it by keeping the calcs as integers as far as possible, multiplying before dividing to preserve precision, and so on.

How does using values of r_drawviewmodel between 0 and 1 sound for enabling translucency? (0 = fully off, 1 = fully on, intermediate values translate to an alpha value.) This one should be real easy to get working.

=peg= said...

"One possible cause is that DirectQ uses low-precision floating point.."

Yeah, my bet would be on that too ;)

"How does using values of r_drawviewmodel between 0 and 1 sound for enabling translucency?"

Thats sounds like music to my ears :D

mhquake said...

Hmmm - another possibility would be my lightmaps tend to be tall but narrow (32x4096 on most cards). If your lightmaps look fine in the horizontal but off in the vertical that's also a likely suspect.

Although I wouldn't be so willing to change that as it's quite performance-critical and ETP is a non-standard hack that just happened to work most of the time.

rovaman said...

Is there a setting that can be adjusted to increase the degree to which one can "look up"? I find myself sometimes limited because the crosshairs don't go to a full 90 deg straight up and down.

If there isn't a setting that controls this, could one be made easily?

mhquake said...

That's quite easy to implement, there are just a few places where viewangles are bounded and all that needs to be done is conditionally remove (or extend) the limits on them. I know a few other engines do this so I can just look at what they have called the cvar (consistency).

=peg= said...

cl_fullpitch is most common I guess (set to 1 to enable 90 degrees up and down pitch, and 0 for the defaults)

I think most multi-player server-side mods stuffcmd this setting to the clients to force the default pitch angles.. so If you adopt cl_fullpitch that should work fine ;)

mhquake said...

cl_fullpitch is in.

r_drawviewmodel intermediate values are in.

r_novis 1 thing is fixed.

No idea on the ETP thing; tried a few changes but no joy. May need to go back to last year and see what happened then.

=peg= said...

"cl_fullpitch is in."
nice!

"r_drawviewmodel intermediate values are in."
Sweet! :D

"r_novis 1 thing is fixed."
great!

"No idea on the ETP thing; tried a few changes but no joy. May need to go back to last year and see what happened then."

To be fair, I think all versions of DirectQ had this issue.
Back then I figured it was close enough for something non-standard, and I figured, I should not bother you about it any further, for my guess was (and still is) that you don't really care to fix it even if you could.
I don't blame you (who am I to do so anyways...), it's not a big deal, but then again.. it would be nice if DirectQ rendered ETP without glitches all the same ;)

mhquake said...

Quark ETP thing is now fixed.

=peg= said...

!! That was.. like.. quick! :D

Floating point precision then?

mhquake said...

Floating point precision.

I've re-enabled the preservation of double-precision FP on D3D startup and all is resolved; keeping an eye on the old performance and so far it doesn't seem to have impacted much; it drops a coupla frames in ID1 timedemos but the important stuff (gameplay) is still solid.

=peg= said...

Well thats just great then :D