New Mouse Sensitivty and Mouse Accel Features

[Lorfa]

Thanks for the advice, it's a nice idea, but unfortunately I think this is purely a mathematical problem. Setting mouseAccel and mouseAccelPower to some values that I don't use will only give me the wrong offset value, as you can see in the last part of the equation.

Given what I said in my previous post, I was able to guess that A() is not a function but a multiplier.

B + (A(v-c))^(P-1) = 2B
(A(v-c))^(P-1) = B // raise both sides to power ^(1/(P-1))
A(v-c) = B^(1/(P-1))
v-c = (B^(1/(P-1)))/A
c = v - (B^(1/(P-1)))/A

Now I only need to know how v is calculated in order to find the correct offset.

v is calculated in counts per millisecond with m_cpi is 0.

When m_cpi is > 0 it uses centimeters per second.

I recommend WinPlot for analyzing various curves and things:

http://math.exeter.edu/rparris/winplot.html

Just use v as your x so for example: f(x) = 5 + (0.1(x-0))^(3-1)

[Phlogiston]

Thanks for the advice, it's a nice idea, but unfortunately I think this is purely a mathematical problem. Setting mouseAccel and mouseAccelPower to some values that I don't use will only give me the wrong offset value, as you can see in the last part of the equation.

Given what I said in my previous post, I was able to guess that A() is not a function but a multiplier.

B + (A(v-c))^(P-1) = 2B
(A(v-c))^(P-1) = B // raise both sides to power ^(1/(P-1))
A(v-c) = B^(1/(P-1))
v-c = (B^(1/(P-1)))/A
c = v - (B^(1/(P-1)))/A

Now I only need to know how v is calculated in order to find the correct offset.

Just tried it and it works with:

sensitivity "5"
cl_mouseAccel "999999999"
cl_mouseAccelPower "99"
cl_mouseAccelOffset "10"
cl_mouseSensCap "10"

At slow mouse speeds it uses the base sens
At high mousespeeds it uses the senscap sens.
The gradient of the crossover between the two is determined by the accel curve (cl_mouseaccelpower and cl_mouseaccel combined)
The place where the crossover occurs is determined by the offset value.

Input your base sens as sensitivity, double it for senscap, then choose a point inbetween which suits your style - or solve your equations unnecessarily - Try it - it works fine (and is a complete head**** for me to try and play with)

[bkraptor]

Thanks, both of you.

I was under the wrong impression that somehow the required offset c depends on A and P (and it does, but not in the way I thought). I am now using A and P with really high values that cause a really steep (vertical) gradient for the accelerated sensitivity which is then clipped off with cl_mouseSensCap at a sane value.

BTW, my base sens is 1 and the accelerated sens is 2, playing with an 1800 DPI mouse (CM Spawn). I have yet to try the m_cpi setting, but I will once I get used to these settings.

[Phlogiston]

Np...looking back I think I got the bit about choosing the offset wrong... since that's based upon the speed at which the mouse is already moving before it uses accel, not upon the sensitivity at a certain point... which is confusing enough to type, basically upto a certain mouse speed X cm/s it will use your original sens then above that it will use the capped sens... which isn't related to the actual sensitivities at all...

TLDR: It works still, but as I originally said you need to just find a nice number for the offset value by playing around.

[DCR]

3+2 different settings on mouse sensitivity in ql that is a must:
m_cpi 0 or "" - default:0
cg_zoomsensitivity 0 or "" - default:1 - coupled with using always zoom (make zoom fov your fov, keep fov 130)
cl_mousesenscap 0 or "" - default:0
Last i checked the defaults were as mentioned.

[MacsPower]

So players can run backwards as fast as forwards and you are ****ing about with this!!!!! Players can jump 27 times in a row and you are worried about the mouse sensitivity. Players with qaud damage can be killed by machine gun on a "skilled matched" game. Fix the big problems first!

[DCR]

27 times in a row!!!!!! holy ****balls!!!!111

[Phlogiston]

So players can run backwards as fast as forwards and you are ****ing about with this!!!!! Players can jump 27 times in a row and you are worried about the mouse sensitivity. Players with qaud damage can be killed by machine gun on a "skilled matched" game. Fix the big problems first!

lol

Quad doesn't make you invulnerable, it just increases your damage output. Try to kill the other person before they hit you 20 times with the MG...

[vanBasten]

Da FUX im using 1600 DPI in a SteelSeries mice the formula says i have to use 90 in sensitivity is this correct damn 90

SENS 90 = 6.5 (0.022*1600/2.54)

[Lorfa]

Da FUX im using 1600 DPI in a SteelSeries mice the formula says i have to use 90 in sensitivity is this correct damn 90

SENS 90 = 6.5 (0.022*1600/2.54)

I believe the equation should instead go:

new_sens = (360 / (old_sens * old_yaw * (m_cpi / 2.54))

New sens should be 3.9965034965034965034965034

Reasoning:

new_sens = old_sens * (old_yaw * (m_cpi / 2.54))

new_sens = 6.5 (0.022 * (1600/2.54))

new_sens = ~90 cm/360 (sensitivity is in cm/360 when m_cpi is set to the mouse cpi)

This is definitely wrong though, 6.5 sens at 0.022 yaw and 1600 cpi does not result in a brutally low 90cm/360. Also missing from the original equation was an extra parenthesis for (m_cpi / 2.54) to prevent an order of operations error (left to right remember!).

Traditional way to determine cm/360 is this:

(360 / (y * d * w * s)) * 2.54
y = yaw
d = dpi
w = windows sens multiplier (should be essentially forced to 1 assuming in_mouse 2)
s = sensitivity

(360 / (0.022 * 1600 * 1 * 6.5)) * 2.54

= 3.9965 cm/360

Formula per: http://www.funender.com/quake/mouse/index.html
Thanks Injx.

It could be rewritten to do to the CPI to Counts/cm in the center, and also leave out the 1 like so:

(360 / (0.022 * (1600/2.54) * 6.5)) = 3.9965 cm/360

So m_cpi 1600 with sensitivity 3.9965034965034 should be identical to sens 6.5 m_yaw 0.022 and m_cpi 0.

This is because m_yaw (and m_pitch) are how many degrees the quake guy will move per mouse count (Lower than 360/65536 would move less than the smallest iteration per count) , and this is then multiplied by the sensitivity value.

So let's say we had sensitivity 1, m_yaw 0.022. This would mean 0.022 degrees of movement per mouse count. In order to get cm/360, we need to know how many counts there are in a centimeter. In this case there are 1600 counts per inch. To get that into counts/cm we need to divide it by 2.54, and we get about 630 counts/cm.

How many 0.022° turns are there in a 360°? 360/0.022 which is ~16363. That's how many mouse counts it takes for a 360° turn. Since one of these is occurring every mouse count, and we have 630 counts/cm, just need to divide 16363/630 to find the cm/360 which comes out to ~25.9 cm/360 about what we would expect from sens 1 at such a high cpi value; this should illustrate the reason why the equation needs a dividend of 360.