Alright here is the best I could come up with:

cl_mouseAccel ".367879441"
cl_mouseSensCap "5 x my sensitivity"
cl_mouseAccelPower "1"

How I imagine it would look like with acceleration .367879441 (but probably is not since I have no idea how mouse acceleration works):

http://i.imgur.com/KXYt4.png

wouldn't setting accelpower to 1 just practicly disable acceleration and add 1 to your base sens?

Don't understand what is accelpower, but value 1 does not disable acceleration.

edit: power 1 does disable accel

no.

you can get something that is roughly the same shape by using a power inbetween 1 and 2.

and yes, using a power of 1 results in a constant sensitivity.

I have cl_mouseAccelPower set to 1 (just double-checked to be sure) and I still get acceleration.

you can get something that is roughly the same shape by using a power inbetween 1 and 2.

Would you mind explaining what it is and how its values between 1<...<2 can get me similar (but not exactly the same?) accel curves.

edit: power 1 does disable accel

here's an example of sens 2, accel 0.5, and power 1.5

click

If the power is P, the accelerating term is raised to the power (P-1). when P=1, it is raised to the power 0, which nullifies the accelerating term (i.e. it becomes a constant (1) which doesn't vary with the mouse speed). With P=2, the sensitivity increases linearly with mouse speed. Powers inbetween give you something that looks like the above plot.

With P=1, you dont have acceleration, you just have a constant sens that is different to your base sens.

Cool, so what cl_mouseAccelPower 1 does is just add +1 to sens. That's why I was fooled because I had higher sens :p

Powers inbetween give you something that looks like the above plot.

Pretty neat graph, this is what most pros use - however, are you sure that there is no possibility to get the exact same graph as that natural logarithm?

no, it's not possible, even if you shift and scale the logarithm.

the more useful question is why do you even want to use a logarithm? and why specifically the natural log?

no, it's not possible, even if you shift and scale the logarithm.

Ok, what is the most similar curve I can get then?

the more useful question is why do you even want to use a logarithm? and why specifically the natural log?

magic

Two more questions if I may:

1) Can you show cl_mouseAccelOffset in the formula sensitivity+(acceleration*x)^(accelPower-1)? I would like to know then exactly mouseAccelOffset disables mouseAccel entirely.

2) In trying to match the accel curve closely to natural logarithm's curve what scale is most important (from 0 to 10 in wolframalpha)? Is it 10, 100, 1000 or even 10000? I would guess probably 100 or even less. I doubt I can move my mouse "1000 times faster than usual".

http://www.esreality.com/index.php?a=post&...pid2087765

Ah so, the correct answer is that mouseAccelOffset never disables acceleration completely [accel * (speed - offset) --> speed can be infinitely fast, no?], I just didn't move mouse fast enough?

offset only stops accel being applied to low mouse speeds.

if it disabled accel completely, it'd be the same as setting accel=0, and would therefore be redundant.

if it disabled accel completely, it'd be the same as setting accel=0, and would therefore be redundant.

But the answer is that it never really doesn't as long as your are able to move mouse physically fast enough you will get accel, yes?

for example: with sens = 0, accel = 10 and offset = 80 I can barely move mouse, with offset = 100 I can't (theoretically because I do not move mouse*¹ fast enough)

¹*1600dpi mouse, raw input

correct.

Alright, now looking at the complete formula:
B + (A(v-c))^(P-1)

One thing is odd, that should happen according to formula, but does not happen in a game itself and that is with accelpower 1:

sens + (positive or negative number)^0 = sens + 1

so with sens 1 I should get sens 2, but that is not so, explain!

works for me.

the power is only applied if accel is non-zero. also check your senscap.

Yes, you are right, I do in fact get + 1 to sens. It was hard to notice with high accel.

edit: I am dumb, with accelpower 1 there is no accel, so why I didn't notice +1 sens increase in my initial test has nothing to do with how high the accel was set.

Ok, before I will become completely lost on this topic (looks like I managed to repeat myself here about accelpower 1!) - consider this:
B + (A(v-c))^(P-1)

A = .367879441
B = 1
P = 1

Why do I get different results with c = 0 and c = 37? According to formula in both cases I should just get +1 to base sensitivity, but in game with c (mouseAccelOffset) 0 it feels like base sensitivity + some sort of accel

edit: it's not accel, just higher sens

when (v-c) is less than or equal to zero, the accel and power is skipped completely, so you are getting the base sens at low speeds.

if you move the mouse quickly enough, you should get the +1 effect.

it should be noted though that you're not supposed to set the power to 1 in the first place. one should consider valid powers to be greater than 1.

so you are getting the base sens at low speeds

Is there a formula for this? The higher mouseacceloffset the lower the sensitivity I get (at least this is the general trend if I compare 0, 37 and 100)

it should be noted though that you're not supposed to set the power to 1 in the first place

When should it be used at all?

the power should only be set to values >1. Typically values around the default of 2 are nice.

you already have the full formula above, with the conditions that the accel term is bypassed when accel is zero or the speed is below the offset.

the power should only be set to values >1. Typically values around the default of 2 are nice.

Again, why? And why is there value 1 at all?

full formula above, with the conditions that the accel term is bypassed when accel is zero or the speed is below the offset

I must be a blind, but it doesn't explain why I get lower sens if P=1 and c=100 compared to if P=1 and c=0

Also you said:

when (v-c) is less than or equal to zero, the accel and power is skipped completely

Is not evident in the formula either because 0^0=1

setting P=1 is a silly thing to do. It's like setting your screen resolution to 0x0 pixels and then asking me why you can't see anything, it's just not supposed to be set to that value.

the reason is as I explained above: with P=1, sens is constant, and if that's what you want, you shouldn't be using accel in the first place.

I must be a blind, but it doesn't explain why I get lower sens if P=1 and c=100 compared to if P=1 and c=0

I've already explained that above. with c=0, you are getting a +1 effect from the accel term. with c=100 you are NOT getting the +1 effect, as the accel term is skipped completely.

...Is not evident in the formula...

no, it's not evident in the formula, which is why i wrote it separately.


None of these odd effects are relevant when you are using the power correctly (P>1), they are purely due to the nature of raising numbers to a power of 0.

with c=100 you are NOT getting the +1 effect, as the accel term is skipped completely.

What effect is it then? And why is there a difference between c=37 and c=100. And how to calculate the way it affects the base sens?

edit: difference between 37 and 100 is not as large as difference between 0, 2 , 5, 10 and 37. So the effect on base sens is probably like this:

for c=0 we get +1 sens and for c>0 we get gradual decrease near base sens. The decrease is probably not linear (because of bigger difference between c=1 and c=10 than c=37 and c=100) and never really reaches base sens [lim(base sens) or however you call it in math]

no.

you will get +1 when v>c and +0 when v<c. as you raise c, you're less likely to get this +1 effect with a typical mouse speed.

the formula is all that appears in the source code, there is nothing extra. (positive) accel will never cause the base sens to be reduced, even if you are using a retarded setting like P=1.

1) there is no visible accel with P=1
2) and again you are saying it's either +0 or +1, but there is a measurable difference with for example c=5 and c=35

there's a measureable difference because v isn't constant in practice.

Ok I give up. I still don't understand anything about c and why sens is different even though formula would imply c value doesn't matter if P=1. But thanks for trying to explain.

B + (A(v-c))^(P-1)

If P = 1 you get (A(v-c))^0

When offset is a large value, and A is >0 , (A(v-c)) will usually be negative.

-x^0 = -1

So, cl_MouseAccelPower 1 != cl_MouseAccel 0, as B will be subtracted by 1 until it reaches your very high test value for the offset at which point you will instantly get B+1 instead of B-1.

You can't get B + 0 without cl_MouseAccel being 0.

With power 1 it is silly though, because it will most of the time have the effect of +1 sens, or -1 sens. Would be much better to just change \sensitivity in that case.

-x^0 = -1

Anything to the power of zero equals one

x^0=1

edit:
http://en.wikipedia.org/wiki/Exponentiation
http://planetmath.org/?op=getobj&from=objects&id=3948

2nd edit:
the reason why calculator gives you an answer "-1" without "()" in "(-1)^0" is that it assumes -x^0 is actually -1*x^0 = -1*1 = -1

Yes but perhaps this is what's happening. I can think of no other possible explanation.

(A(v-c))^(P-1) must evaluate to a negative for B to be lowered.

It would be weird for the code to get the order of operations wrong, but without seeing the code itself I cannot rule this out.

I think what should be done is to precisely measure the difference in sens when c is a high value vs. when not, if it is the same as -1 sens then the order of operations thing is probably what's happening.

If not it would indicate some other cause, but what?

you didnt read any of my posts did you.

when (v-c) is less than or equal to zero, the accel and power is skipped completely, so you are getting the base sens at low speeds.

the accel term is bypassed when accel is zero or the speed is below the offset

(positive) accel will never cause the base sens to be reduced

what he's experiencing is purely down to this:

you will get +1 when v>c and +0 when v<c. as you raise c, you're less likely to get this +1 effect with a typical mouse speed.

because v isn't constant

Gotcha. Thanks for the info.

What units does v use with m_cpi 0 and m_cpi >0?

counts per millisecond, and cm per second.

The master has spoken!

coollers sens equals all natural logarthims

wow, so much shit over fucking pc game

It's not shit, curves are important!

just type sens 4 and get used to it D:

To be completely honest with you this is almost what I have (assuming sens 4 applies to dpi 400):
sensitivity "1.0753" for all weapons @ 1600 dpi + no accel

But I am very tempted to start using acceleration, because in theory it has a few benefits over no acceleration. The problem is hard to understand commands and therefore impossible to visualize (before this forum thread and injx replies) what's going on.

sens 4 accel 0.1 :D

magic quake numbers!

Actually I just started using acceleration after 20 years of gaming and it's made a huge difference. So no it's only useless if you don't care about aiming.

thank you for your reply!

btw i have sens 6 accel 1 now :<

that is a hell of a nerdy question

i still wait for one of you computer scinetists who seem to play quake and throw alot of numbers come out with a demo or movie showing what kickass aim and movement all that messin gets you.

anybody know how to make mouse accel curve like natural logarithm using povohat mouse accell driver?