How to make a graph of sensitivity/accel settings?

[SSRI]

Hi!
I have been experimenting with my mouse settings using wolfram to plot the formula I found on this forums (B + (A(v-C))^(P-1)).
I understand what all the variables do and how the line/curve describes the kind of acceleration used.
I guess that y axis is simply the in game sensitivity.
If the x axis is velocity, in what units is it expressed?
It would be useful to know what mouse movement speeds correspond to the exponential acceleration or how quickly I reach the insane speeds.
Is it possible to make individual graphs based on mouse cpi settings?

I would be grateful if you helped me understand the math behind it.

[Lorfa]

I have been experimenting with my mouse settings using wolfram to plot the formula I found on this forums (B + (A(v-C))^(P-1)).

I guess that y axis is simply the in game sensitivity.

If the x axis is velocity, in what units is it expressed?

Instead of WolframAlpha I use Winplot:

http://math.exeter.edu/rparris/winplot2005.html

Just go to window --> 2-dim , then when the window opens up, Equa --> Explicit

From there you have an f(x) = to put in. I set x to be the velocity, which is in counts/millisecond.

For example:

f(x) = 4 + (0.5(x-5))^(2-1)

Using B + (A(v-C))^(P-1)) as a guide, you can see that:

cl_MouseAccelPower "2"
cl_MouseAccel "0.5"
cl_MouseAccelOffset "5"
sensitivity "4"

This makes the x axis the velocity, and the y axis sensitivity.

At any time you can go to Equa --> inventory to see all the equations you put in.
You can also go to Equa --> Segment to put in a vertical line if you want to see where two curves meet up, or you can just set an explicit where y is equal to a constant for cl_MouseSensCap.

For example:

y = 4

Draws a straight horizontal line at y = 4, so you can just imagine your curve plateauing at that point.

To give you an idea of what this looks like here's a screenshot:

http://img542.imageshack.us/img542/2394/offsetgraph.png

Is it possible to make individual graphs based on mouse cpi settings?

I don't see why not!

m_cpi changes the units the equation uses though. When m_cpi > 0, the velocity is in cm/second, sensitivity is in degrees/cm, offset is in cm/second, I guess power is still dimensionless, and mouseaccel would be something like "degrees/cm added per unit of velocity".

The thing is that it might not be clear on the graph that doubling cpi doubles the number of counts at a given velocity.

Remember to do the conversions when using m_cpi, found at:

http://www.quakelive.com/forum/showt...l=1#post137695

Note: the mouse accel one isn't so intuitive because of the ^2 in there. Also I always put the cpi/2.54 part in parenthesis. I don't know if you were around for the whole order of operations debacle that spread through the net about the answer to 6÷2(1+2) , if not see here: http://knowyourmeme.com/memes/48293 :-) Doesn't apply here but I like to be on the safe side :-P

[jigglywiggly]

while all of this ms algebra is gr8
why would you want to use accel

[SSRI]

So if I want to convert the x axis to meters/second and my mouse is 450cpi, do I just multiply x by 1000 to make seconds, divide x by 450 to make inches and multiply x by 0.0254 to make meters?

http://www.wolframalpha.com/input/?i...x+from+0+to+10

Does this work?

One more question.. how does sensitivity work? Is the number of counts sent by mouse multiplied by sensitivity? How does that translate to the movement on the screen (degrees, pixels)? .... well ok that was three more questions hopefully I'm not annoying.

[Lorfa]

So if I want to convert the x axis to meters/second and my mouse is 450cpi, do I just multiply x by 1000 to make seconds, divide x by 450 to make inches and multiply x by 0.0254 to make meters?

http://www.wolframalpha.com/input/?i...x+from+0+to+10

Does this work?

I don't think I quite understand what you did here :-P

CPI/2.54 should be in brackets, to convert counts per inch to counts per centimeter.

When graphing m_cpi > 0, you don't need to change the equation, just the way you look at the same graph.

The x axis will be in cm/second, and the y in degrees/cm, as opposed to cm/second and dimensionless multiplier.

Since counts/ms tends to be smaller than cm/second, and larger than meters/second, you may wish to adjust the way the graph draws the x-axis by adjusting the range that x is shown. I didn't find an option in Winplot to just change the units used outright, so maybe a fancier program is needed.

Perhaps wolframAlpha has this feature. You'd first have to add units when you know them to the equation, and then its output may allow you to show in meters vs. centimeters for example.

One more question.. how does sensitivity work? Is the number of counts sent by mouse multiplied by sensitivity? How does that translate to the movement on the screen (degrees, pixels)? .... well ok that was three more questions hopefully I'm not annoying.

Not annoying at all.

Just using m_cpi = 0 for now, sensitivity is a multiplier. It multiplies by m_yaw and m_pitch to produce the final mouse sensitivity.

m_yaw (and m_pitch) is in degrees per mouse count. So if you set sensitivity to 1, and m_yaw to 1, your mouse would move one degree per count. The quake guy has a maximum of 65536 iterations in a 360° turn, so to turn one of these per mouse count you'd set sens 1 and m_yaw 0.0054931640625 (360/65536).

How a change in degrees translates to pixels on screen is more complicated, and I don't feel comfortable enough in my understanding of it to say much.

The 3d world is viewed with a frustum (looks like this: http://www.lighthouse3d.com/tutorial...ustum-culling/ ), and the value of cg_fov is the horizontal fov. The vertical determined by the aspect ratio. The pixels change more on the sides than in the center as a result.

[SSRI]

OK, I think I now know how to express what I wanted to do.
When I plot my settings, with m_cpi = 0, where the y axis is the sensitivity and x is counts/millisecond, can I also read off/calculate what is the speed that I move my mouse at at any given point on the line?

I also did an experiment with m_cpi = 450. With these settings:
http://www.wolframalpha.com/input/?i...x+from+0+to+10

The graph shows that at 10 cm/s there will be 9*10 degrees movement. I used a metronome and a ruler to move 10 cm/s on dm6 looking at the doors separated by 90 degrees and it works out nicely.

So now, with this other method where m_cpi = 450, y is degrees/cm and x is cm/s can I also read off/calculate what is the sensitivity at any given point on the line?

[Lorfa]

OK, I think I now know how to express what I wanted to do.
When I plot my settings, with m_cpi = 0, where the y axis is the sensitivity and x is counts/millisecond, can I also read off/calculate what is the speed that I move my mouse at at any given point on the line?

If you know your mouse's CPI, you can convert any x-axis value to m/s:

[(v * 1000) / (CPI / 2.54)] / 100 = m/s

Where v is in counts/msec.

I also did an experiment with m_cpi = 450. With these settings:
http://www.wolframalpha.com/input/?i...x+from+0+to+10

The graph shows that at 10 cm/s there will be 9*10 degrees movement. I used a metronome and a ruler to move 10 cm/s on dm6 looking at the doors separated by 90 degrees and it works out nicely.

I assume you set your base sensitivity to 0 here as a test :-)

So now, with this other method where m_cpi = 450, y is degrees/cm and x is cm/s can I also read off/calculate what is the sensitivity at any given point on the line?

Yes.

Here's a more reasonable graph using more likely real world values:

http://www.wolframalpha.com/input/?i=plot+y%3D%2818%2B+%280.2%28x%29%29^%282-1%29%29%3B+y+from+18+to+40+x+from+0+to+100

Mouse accel 0.2, the equivalent with m_cpi 0 would be ~0.37, a bit high but good for an example.

Sensitivity is 18 degrees/cm, which comes out to a very normal 20 cm/360.

Now you can pick out an x value, say 40, for 40 cm/s or 0.4 m/s.

At 0.4 m/s y is 26 degrees/cm. To get cm/360 just divide 360 by this number. For example 26 degrees/cm is (360/26 = 13 cm/360).

You can see that this accel is quite powerful, at a "slow and steady" 0.4 m/s the sensitivity has already almost doubled.

[SSRI]

Thank you! This will guide me in my experimentation and give me something to visualize.
And yes, I did play some FFA for the last couple of days with sens 0, accel 1 ,accelpower 2, just to see what it is like (my accuracy was near 30%).

[Lorfa]

Thank you! This will guide me in my experimentation and give me something to visualize.
And yes, I did play some FFA for the last couple of days with sens 0, accel 1 ,accelpower 2, just to see what it is like (my accuracy was near 30%).

Wild :-)

I tried base sens 0 a few times and it is like, really difficult to move :-P

Massive delta between motion and stop there :-))

[megaman3]

Thank you! This will guide me in my experimentation and give me something to visualize.
And yes, I did play some FFA for the last couple of days with sens 0, accel 1 ,accelpower 2, just to see what it is like (my accuracy was near 30%).

Please get some cfgs from dreamhack summer (2 months ago, because I don't have hopes on quakecon staff) and post those graphs