Qrack 2.01 beta public release

[Baker]

Ill try that tripplebuffer trick...

Naus, I see your celshading isnt working weird since you have a newer gfx card...

R00k, you might check out cl_independent_physics in JoeQuake Build 1829. From what I recall of it (but I could be wrong), it appeared to unlock the relation between frames per second and mouse rate.

Not 100% sure of this at the moment.

[Lightning_Hunter]

R00k, you might check out cl_independent_physics in JoeQuake Build 1829. From what I recall of it (but I could be wrong), it appeared to unlock the relation between frames per second and mouse rate.

Not 100% sure of this at the moment.

Interesting, if that cvar actually does work. Even modern games have issues with mouse lag while using V-sync.

Just don't go messing with the mouse code too much. It is finally back to the way I liked it in 1.85 after so many versions.

[R00k]

i KNOW, i just ripped out the m_smooth code, as it was useless internal acceleration based on m_rate and buffer history, phew! I found a bug in the connect "last" function if you issued it twice in a row to the same server the server name got trashed. also writealias was buggered, but that's fixed now.

Actually, since i cleaned up some code in in_win.c i cant make it have mouselag with m_directinput 1 or 0 or vid_vsync 1 or 0.. hmmmm wtf??

edit: good time to back up my source code

[Lightning_Hunter]

R00k, be prepared for some Lightning Hunter complaints if the mouse handles differently in the next version.

Seriously though, I thought the handling in the last uploaded beta was flawless as long as -dinput -m_smooth and -noforcemparms were used in the command lines. In fact, I felt the handling was better than any other game out there.

[ungod]

I've set the mouse back further to mimic the 1.85 build. Seems there was an issue where in 1.90 mouse acceleration was taking priority. Default now is zero acceleration. You may wish to try m_smooth 1, which uses software acceleration with m_directinput 1. That' really the only difference. Also i forgot to mention i added a new cvar "m_noforcemparms" so that you dont need to add it to the commandline.

also try using vid_vsync 0, as setting it to 1 causes mouse lag.

i hate to be the one to ruin the parade but i have tried to do what you suggested but it seems off completely, otherwise i wouldnt mention it.

testing
qrack1.85 build 2194- m_directinput "0", m_smooth "0". vid_vsync "0" = mouse feels good, with minimal acceleration
qrack1.90 build 2422- m_directinput "0", m_smooth "0". vid_vsync "0" = mouse feels good, with minimal acceleration
qrack2.010 build 2878 - m_directinput "0", m_smooth "0". vid_vsync "0" = mouse feels good, with minimal acceleration

qrack2.010 build 3095 - m_directinput "0", m_smooth "0". vid_vsync "0" = hyper sensitivity, no mouse acceleration.
qrack2.010 build 3095 - m_directinput "1", m_smooth "1". vid_vsync "0", m_noforcemparms "1" = no mouse acceleration.

-ps, admittedly for me anyway, mouse accel in quake has never really been "perfect" could be better and mouse accel can really fuck things up if you dont have the mouse drivers to tweak stuff like that, but, just want enough to make fast tracking, 180 degree and 360 degree turns easier without loosing precision aiming needed when shooting rockets across the map, like in most modern games nowadays (halo,cod,fear,etc.).
Any way to get that back in qrack without breaking or overriding anything? because i cant play without it and using dinput and mouse smooth doesnt seem to be adding any mouse accel of any kind.

also alt tabbing out latest build seems to mess up the mouse even more, i can get around that by simply disabling m_directinput then re-enabling it.

cant you just rollback the mouse code back to 1.85?

[R00k]

Okay, give this a test...

http://www.quakeone.com/qrack/dist/qrack201beta.zip

(same filename i just keep updating the zip file so this thread always has current beta version...)

Personally, mouse acceleration is the worst idea for FPS games since the trackball, and I'm trying to remove all existence of it in Qrack. Toggling m_directinput 0 to 1 and moving the mouse feels very close to the same, no half sensitivity for me on this end.
Also test with -noforcemparms too...

[Lightning_Hunter]

Mouse feels perfect to me! I noticed that M_showrate doesn't seem to work now, however.

Personally, mouse acceleration is the worst idea for FPS games since the trackball, and I'm trying to remove all existence of it in Qrack.

I agree. I HATE mouse acceleration. There is an awesome port for Quake2 called Quake2XP, but it has one problem... YOU CAN'T GET RID OF MOUSE ACCEL! I can't stand it! I refuse to play with mouse accel.

[R00k]

Mouse feels perfect to me!

Ya m_showrate was tied into m_smooth, and i always got 125hz no matter what value m_rate was set to :S
On another note, I tested cl_independent_physics from JQ, which was a port from ZQuake. Though, one thing to explain about how that command works for QuakeWorld; qw takes a series of frames and creates a buffer of input from the client. It then uses that input history buffer to create a prediction table and interpolate it with the history from the server to create a virtual lag free environment. The timing between input frames is normally based on the ticrate of the server, but with independent_physics this timing is overridden by the maxfps of the client. Well, none of that applies to netQuake.
netquake already has the raw benefits of this when the cl_maxfps variable was implemented instead of the locked 72fps of old glQuake. So, in a nutshell, this command is useless for netQuake

[Lightning_Hunter]

My mouse is at 1000Hz, so I wanted to make sure Qrack was using the correct polling rate. I guess since I can't use m_showrate, I can't check to see!

[R00k]

Okay, I'll work the m_showrate back in just to ease your worried mind

[Lightning_Hunter]

Okay, I'll work the m_showrate back in just to ease your worried mind

When you put it that way, I feel paranoid now.

[R00k]

Well actually, Quake only calls the CL_SendCmd function, which reads the keyboard, joystick and mouse, based on (1 / cl_maxfps) of a sec. And only 1/72 if playing locally. So, unless you are getting 1000fps, i dont see how reading the mouse that much would make a diff.

[ungod]

Personally, mouse acceleration is the worst idea for FPS games since the trackball, and I'm trying to remove all existence of it in Qrack.

hehe while i agree with you about the trackball, please dont remove mouse accel from qrack one reason you guys hate mouse accel as did i at 1st is because you dont spend the time tweaking it to get the right feel. to do this i have always played around with mouse driver sesnsitivity, ingame sensitivity, Dpi, and polling rate untill i was happy with the results.

Okay, give this a test...

http://www.quakeone.com/qrack/dist/qrack201beta.zip

(same filename i just keep updating the zip file so this thread always has current beta version...)

okay thanks let me check it out

[lohphat]

2.01 has a problem with shambler.mdl. It says it's a wrong version number. I did install the new shambler model (it's now a pk3), which didn't work in any engine. DP will use the .mdl, and it's the old shambler model. If I load the map E2M2 with this mdl, Qrack will crash.

[MH]

That's not an mdl, it's an md3 (or something else) with the extension changed. I really really wish that DP hadn't supported this kind of tomfoolery in the first place...