Ex_interp "0"
falloutx2
Join Date: 2003-05-01 Member: 15979Members, Constellation
<div class="IPBDescription">Is the 1/updaterate formula correct?</div> I was messing with various net settings today in an effort to get them to be optimum. I decided to let the game set my ex_interp by using "0" and it came up with .016 ms because my cl_updaterate was "60"
I ended up with the following settings:
cl_cmdbackup "81"
cl_cmdrate "101"
cl_updaterate "60"
ex_interp "0.05"
ex_extrapmax "6"
My concern is that if I allow the game to set ex_interp on it own, the .016 value is very far from correct. I had significant teleporting of players as caused by incorrect ex_interp when using this setting, so I reset to .05 and had much smoother results.
Its not a bug perse, but it seems the default ex_interp setting for ex_interp "0" should get closer to an accurate setting than it currently is.
I don't think this is anything new as I've seen it discussed before. I am considering opening a bug report on this once I get some data as to whether it performs better for users with higher cl_updaterates by default and its just a matter of my personal settings.
It would be could if ex_interp "0" caused a more useable value than it currently does.
It would aid me greatly if people would post the relevant portion of their config (i.e. ex_interp setting and cl_updaterate setting) so I can get a feel for the depth of discrepancy between what ex_interp value is generated and what ex_interp value is actually function.
I ended up with the following settings:
cl_cmdbackup "81"
cl_cmdrate "101"
cl_updaterate "60"
ex_interp "0.05"
ex_extrapmax "6"
My concern is that if I allow the game to set ex_interp on it own, the .016 value is very far from correct. I had significant teleporting of players as caused by incorrect ex_interp when using this setting, so I reset to .05 and had much smoother results.
Its not a bug perse, but it seems the default ex_interp setting for ex_interp "0" should get closer to an accurate setting than it currently is.
I don't think this is anything new as I've seen it discussed before. I am considering opening a bug report on this once I get some data as to whether it performs better for users with higher cl_updaterates by default and its just a matter of my personal settings.
It would be could if ex_interp "0" caused a more useable value than it currently does.
It would aid me greatly if people would post the relevant portion of their config (i.e. ex_interp setting and cl_updaterate setting) so I can get a feel for the depth of discrepancy between what ex_interp value is generated and what ex_interp value is actually function.
Comments
Prior to using ex_extrapmax "6" I had still had problems with "jerky models" using ex_interp "0"
I thought ex_interp was too :/
<!--c1--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>CODE</b> </td></tr><tr><td id='CODE'><!--ec1-->(<current ping> + 10) / 1000 = ex_interp<!--c2--></td></tr></table><div class='postcolor'><!--ec2-->
Is this completely wrong?
I mean ex_interp is afaik for lag compensation. And how should it work without considering your ping?
if it is, would be very nice if next versions of natural selection used this formula to auto-adjust the ex_interp
Ex_interp is a variable the determines how long the prediction engine is in control of the model placement and all of the other fun stuff. The server is not able to give you the position updates 100 times every second, so the Half-life engine on your client has to predict where the models are. So, the optimum ex_interp at 1/cl_updaterate (that's a fraction with cl_updaterate as the denominator), because that is how long it takes between server updates.
While nothing beats an accurate update packet from the server, you don't want models changing position at 30 FPS with your video card rendering the animations on them as well as everything else at 100 FPS, do you? The prediction engine does a very good job of filling in the gaps.
By the way, ex_interp was LOCKED in B4A, but UNLOCKED in B5. If you haven't done so already, I would highly reccommend setting your ex_interp to 0, because Half-life will then automatically set it to 1/cl_updaterate.
I'm not sure about ex_extrapmax; I've never heard of it before.
As for the original post, I think Yumosis hit it dead on. If you change your cl_updaterate to 30 and make sure that ex_interp resets itself correctly, you should be back in business.
Just like how the minimum required updaterate/rate works, it checks the server sv_minupdaterate/rate, I have no idea why Valve just doesn't do this for sv_maxupdaterate and sv_maxrate <!--emo&:(--><img src='http://www.unknownworlds.com/forums/html//emoticons/sad-fix.gif' border='0' style='vertical-align:middle' alt='sad-fix.gif' /><!--endemo-->
Good point, anyone care to enlighten? Just a basic description if what it does and how it affects your gameplay should suffice, aswell as optimal values? <!--emo&:)--><img src='http://www.unknownworlds.com/forums/html//emoticons/smile-fix.gif' border='0' style='vertical-align:middle' alt='smile-fix.gif' /><!--endemo-->
taken from <a href='http://www.foom.net/fortressfiles/netcode.htm' target='_blank'>http://www.foom.net/fortressfiles/netcode.htm</a>