Forum

What interp???

Created 15th November 2010 @ 20:47

Add A Reply Pages: « Previous 1 2 3 ... 6 Next »

fe

You need to have next frame to do interpolation. Next frame comes in cl_interp_ratio/cl_updaterate, which usually is 1/66=0.0152s, but, in real life (in Soviet Russia;), sometimes we lose frames, so we need 1 more frame to have smoothly motions. It’s 0.0304s. When you have interp 0.0304, lerp is white – good indicator of working interpolation.

Spira

Well yellow is bad and orange is good; just less than default.

AnimaL

you wont get yellow unless ur in shit server

Orchid

|=3

Im so confused about all of this right now.
No idea how to get it right!

Koeitje

AUTOBOTS

It will probably determine about 0.5% of your total performance in the game anyways, so who cares.

IPZIE

SUAVE

Just like skinnie said. You take this too sirious :D

octochris

(0v0)

Quoted from fe

You need to have next frame to do interpolation. Next frame comes in cl_interp_ratio/cl_updaterate, which usually is 1/66=0.0152s, but, in real life (in Soviet Russia;), sometimes we lose frames, so we need 1 more frame to have smoothly motions. It’s 0.0304s. When you have interp 0.0304, lerp is white – good indicator of working interpolation.

nope.

besides, in reality it’s about what you’re used to. my values are optimal (as tested through a few methods) on most setups but if you’re used to some other setup I wouldn’t recommend changing.


Last edited by octochris,

Rake

Lutunen
[hePPa]

I’m curious as to what kind of tests – did you use something other than just a bunch of people trying different settings and seeing what felt right, or did you use some black magic testing methods.

E: you wouldn’t know anything about bad hitreg, being dutch and all koeitje. Nothing makes a difference when you ping <10


Last edited by Rake,

octochris

(0v0)

Quoted from Rake

I’m curious as to what kind of tests – did you use something other than just a bunch of people trying different settings and seeing what felt right, or did you use some black magic testing methods.

Scripted hit/register tests on a few servers.

Quoted from Rake

you wouldn’t know anything about bad hitreg, being dutch and all koeitje. Nothing makes a difference when you ping <10

I don’t know why people bother to factor in ping in their interpolation values, it’s nonsensical — that’s not what it compensates for and is not a factor in interp calculation. Latency cannot be alleviated with interpolation.

pykow

using bad connection config by chris. worsk greate for me.
ATM i dont use anyscript this time, cuz im to lazy to do it and to busy with other stuff.

Dr-GimpfeN

9g1c

i am using 0.01337 and i am only doing l33t h34dsh0tZ

fe

Quoted from octochris

[…]
nope.

So, if you have not next frame (update from server), you interpolate between last update and? You need next update anyway, or predictions can go wrong.

octochris

(0v0)

Quoted from fe

[…]
So, if you have not next frame (update from server), you interpolate between last update and? You need next update anyway, or predictions can go wrong.

The source engine already has native compensation for prediction errors. Interpolation is not the way to do that.

Waster

For hitscan it doesnt matter. As soldier playing with interp of 0.1 is like playing with 100 ping.

Koeitje

AUTOBOTS

Quoted from Rake

E: you wouldn’t know anything about bad hitreg, being dutch and all koeitje. Nothing makes a difference when you ping <10

I ping 20 sometimes :(.

Anyways you can just use a higher interp on higher ping connections (If your connection is unstable, there should be no need if its stable and you dont drop packets). Your hitboxes wont get worse, you will just see your opponent later when both of you come around a corner.


Last edited by Koeitje,

Add A Reply Pages: « Previous 1 2 3 ... 6 Next »