"pixel perfect" is a bad term to use, as it just causes confusion, as your post demonstrates. however people just cant help themselves.

The basic idea is that you want to know how much DPI is *useful* for a given setup based on what is visible on your screen.

Now there is more than one way to do this, as "usefulness" isn't a well-defined concept. However you can get a *sensible* estimate of how much DPI is useful. In fact there are a number of sensible ways to estimate useful DPI, but none of them are 'special' in any way.

One type of "useful DPI" estimation was featured in Sujoy's ESR MouseScore (link).

I have since developed a slightly more sophistocated estimation (link), which is arguably better than the simple one above.

A lot of the misunderstanding about this issue comes from the term "pixel skipping" which is blatantly WRONG when talking about aiming in 3D games (it should only be used to describe 2D mouse cursors).

Finally, do beware of what others say, as most people are idiots. Anyone using the term "pixel skipping" or "pixel perfect" or interpreting these "usefulness estimates" as magic numbers isn't qualified to give advice.

You are like a mouse/dpi jedi.... respect.

the force is strong with this one

taking another view at the esr mousescore in 2010 reveal some interesting details about gaming hardware imo. the wmo which receives endless praise on these forums scored this:

Perfect Control: 1.00 m/s (40"/s) [1.55 m/s (60"/s) @ 1000Hz]

now i kind of wonder why so many people bitch and whine about recent laser mouse with similar stats. perfect control up to just 1 m/s? according to some "experts" this should already disqualify the wmo as a serious gaming piece of hardware.

do the esl iem lans even allow "overclocking" of the usb hz?

most people with wmos overclock the usb (you only need to go to 250Hz to get the 1.5m/s).

now a top speed of 1.5m/s at a relatively low sens of 30cm/360 means you can still perform a quick 180 flick in 100ms and still get perfect control.

so 1.5m/s is not as low as it sounds for most players.

as for lasers, afaik the reason people bitch is due to tracking reasons (inconsistency, skipping, or only working correctly on certain surfaces), it's not due to stats.

i cant comment on laser tracking as i dont have a laser mouse.

yea, 1,5 m/s is quite some speed but with about 35cm/360 I easily hit more than 2 m/s

haven't used a WMO/IME/IMO for a long time, so I don't know how well I could handle the neg. accel.

PS: some people will always bitch about laser even if we would have the "perfect" sensor

Allthough 400 CPI (for example) are allways as precise regardless of your display resolution it can feel unprecise if you see your crosshair jumps some pixel when doing a small movement.

All those calculations aren't magic numbers (just like injx wrote) but they can give you a routh estimation on how much CPI you really need.

Oh, and you are wrong on the technical side, the wrong term DPI leads to this assumption.

The CPI haven't anything to do with the size of the images the sensor takes (not every sensor takes even images), at least not in the way you think.
The image size is 30*30 pixels on most gaming mice (MX518, DA, DB, CH, G5, G9x/G500, Xai, etc.), the IMO/WMO/IME sensor has iirc a 22*22 image array.

Can you elaborate on how I'm wrong about what DPI refers to? I got my information from this site. http://home.comcast.net/~richardlowens/OpticalMouse/#resolution

I know that site and at no point it says that the CPI refers to the image size of the sensor as you wrote.

Higher CPI don't mean that the sensor takes images with a higher resolution.

OK maybe I phrased it wrong, but the idea is still the same. On the site it says this "Just keep in mind that this is how many optical sensor pixels per inch the camera sees". The point I was trying to make is that the 400 DPI doesn't mean that the cursor moves 400 pixels for every inch of movement on the mouse.

So with higher DPI/CPI the camera sees more sensor pixels. It also says, the DPI is determined by the optical properties of the focusing lens and the physical size of the image sensor. So since the sensor size can't change, I'm assuming changing the DPI of a mouse changes the focusing lens somehow. Ie the image size will be the same, but the lens will focus on it in a different way, so that it becomes more segmented and has more "optical pixels"(example it zooms in for lower DPI, and out for higher DPI).

My belief is that DPI/CPI has nothing to do with pixels/accuracy/etc, and it is only a hardware sensitivity setting, analogous to changing the volume on a pair of headphones either by the volume in windows, or on the headphones themselves. At higher DPIs, you don't need to move the mouse as far, for the sensor to register a movement, hence the increased sensitivity. How does that relate to anything on the screen other than the fact the cursor will just move faster? Higher DPI/CPI might be useful for people playing at higher resolutions or sensitivities, but only because it allows them to keep everything else at default, ie their window sensitivities, or their game sensitivities, to prevent interpolation/pixel skipping(which isn't even necessary, based on injx's earlier post).

If I'm wrong about this, please let me know.

the point is that the sensor's CPI determines the smallest movement you can make that will be registered (i.e. a single count). this is very much related to accuracy, and so it's more than just a sensitivity setting.

the technical details about how the sensor is engineered and the size/resolution of the imaging array isnt really important.

So is just raising your ingame sensitivity somehow inferior?

Is 800CPI @ a sensitivity of .5 better than 400CPI @ a sensitivity of 1?

technically yes (as long as the mouse tracks and responds correctly), as small precision movements will be more accurately reproduced.

so the big question is "how much is enough?" or "how much is actually useful?", and that is exactly what i was talking about above.

Are people even able to take advantage of such small movements though? It's very rare to hear of someone who uses high DPI, but doesn't reduce ingame sensitivity or windows sensitivity to compensate.

Read Memento Mori's post.

I don't really know what a sensor exactly does if it is set to different CPI values (it's on my list :D) but it can't adjust the lens, ther are no moving parts in an optical sensor.
As injx wrote, all that technical stuff isn't really important.

Your are right that the cursor will not move 400 pixels if you move a mouse with 400 CPI 1 inch (allthough it can with the right settings in an 3D enviroment)) but the sensor/mouse will send 400 counts (most likely not exactly 400 counts depending on some factors) and ofc higher CPI also means smaller minimal movement to send a count.

The question for everyone is how much CPI do I need and can I really benefit from more CPI?
With my sensitiviy @400 CPI at a distance of 40m 1 count send by the mouse represents a difference off a liitle bit less than 5cm (between 2 shots).
So for me doubling the CPI doesn't really sound like a practical increase of accuracy (in real world gaming).

Oh, and sensor have a tendency to be less precise at higher CPI, allthough this gets compensated to soem degree by the lower ingame/software sensitivity.

I made a long post about this some time ago but I cannot find it. I will cut things down here, but I hope it's as clear.

What people should understand is that if you have a mouse with decent DPI (which is any existing mouse, from 400 up), rising the resolution will not influence anything. Instead of skipping 1 or 2 pixels, you might skip 2 or 4, but in the end that's just a visualization of the in-game rotation, which is exactly the same . The ingame rotation precision is only influenced by the DPI and your sensitivity, and not by the resolution.

The stage when the resolution matter is when you at the screen to see what is going on. The problem (if any) is not a high resolution, but rather a low resolution, which in extreme cases (very low res, very low sensitivity and very high DPI) might put you in the situation where the ingame rotations you are able to produce are not noticeable on the screen. But really... this won't ever happen in a real case.

So, stop worrying about DPI and resolution, it's just one of those viral topics which have no grounding.

/e: I just found my post:
http://esreality.hexus.net/?a=post&id=1843131#pid1843131

I read your linked post, and I agree. It fits with what I've personally experienced with mice, and makes the most sense.

Plenty of in-depth explanations here...but wot about the issue that only old game engines doesn't care of high DPI/CPI like Quake? State of the art games like Crysis 2 prefer HD resolutions while input preferences are weak...

Are you saying you read MM's posts, found them interesting, yet didn't understand anything or are you trying to convince yourself buying a 800dpi mouse?

edit : I mean, coding a fps in such a way that it would need more than 1 tick to change the inside representation would be absolutely stupid considering the amount of people playing with insanely high sens on stock 400dpi dell or logitech mice.
And the default ingame sens value is an arbitrary one.
What might have thrown you off is that some game apply a flat accel curve to your mouse movements ( I think I read that about q2 or q3 ).

"...DPI... actually refers to the resolution of the images taken by the optical sensor in the mouse...."

Negative!.

The sensor resolution of a typical mouse sensor is (30x30) ~900pixels...is not like the typical camera ( ~~MPixels = 1000000 pixels ;) ) XDDDD

4 more infos:
http://home.comcast.net/~richardlowens/OpticalMouse/

& you can check several mouse sensor datasheets using google.

You haven't read the whole thread, right?