forum

Mouse drift is probably due to rotating mouse; any fixes?

posted
Total Posts
54
Topic Starter
Mio Winter
Mouse drift is when your mouse gradually moves further and further in one direction while Osu!ing, such that you have to lift the mouse off the mousepad and reposition it in order to avoid your mouse falling off your desk. This is bad, because you don't always have time to reposition your mouse when playing Osu! and then your hand ends up in a really awkward position, killing your aim.

Assuming you've put mouse acceleration off, this is what I believe is the main cause of the problem:



In other words: unless you keep your mouse at the same angle all the time, you drift. And the more your change the angle of your mouse, the more you drift.

Solution 1: Play with your arm instead of your wrist. It's easier to maintain the same angle if you let your whole arm do the movement. I don't like this solution, however, since I suspect it's easier to learn how to aim by letting your wrist and fingers do the fine work.

Solution 2: A mouse that tracks using the absolute position of the mouse on the mousepad. Or a mouse that tracks the rotation of the mouse in addition to its velocity, so that it's able to track absolute distance travelled from a set point as long as you don't lift your mouse or hit the screen boundaries. Does anyone know of a mouse that does any of this?

Solution 3: No, I don't want to use pen! : P I started playing Osu! in order to get better at clicking orbs in Gop, and I can't play Gop with pen, so I want to continue practicing Osu! with mouse.

Solution 4: Use higher mouse sensitivity so that you don't have to move (or rotate) as much. I'm considering this, but it also seems likely that it's easier to learn better aiming with the relatively low sensitivity I have right now...
Endaris
Don't play osu! to get better at other games. Play those games to get better at them instead.
Aside from that mousedrift is only a problem on really long maps where the mapper looped himself in the same direction all the time. It's no problem to readjust your mouse if you get like half a second sparetime to do so.
After all you will learn to subconsciously deal with it.
Topic Starter
Mio Winter

Endaris wrote:

Don't play osu! to get better at other games. Play those games to get better at them instead.
Aside from that mousedrift is only a problem on really long maps where the mapper looped himself in the same direction all the time. It's no problem to readjust your mouse if you get like half a second sparetime to do so.
After all you will learn to subconsciously deal with it.
Don't worry, I have plenty of fun playing Osu! : D It's not like it's a chore. After playing a game like Gop for a few years, your skill hits a plateau after exploiting all the obvious ways of getting better to their limit. It means you have to get creative in order to find out how to get better once that plateau has been reached.

I suppose one solution would be just to get good at repositioning quickly, and perhaps get a mouse that doesn't track when the mouse is only slightly above the mouse pad. I hope you're right, and that I'll get better at this. But I would prefer a mouse that tracked absolute position! : P
The Gambler

Timo Timo wrote:

Does anyone know of a mouse that does any of this?
Not perfect but...



Before you ask, the mouse had already died from my abuse... About time I gave it a chance to fight at valhalla.

If I may also add, the aim used for absolute positioning is very weird on a mouse grip... Everything you learned using a regular mouse won't apply here.
Xyrus_old_1
Solution 5: You accept that mouse drift is an inevitable problem on longer maps and work on compensating for it. You do this by readjusting your mouse at opportune moments - You need to learn to quickly lift the mouse, move it to the side / forward / back and place it back down very quickly. This takes time and will cause you to break a few combos while you learn the timing, but it is highly effective once you get the hang of it.

The Gambler wrote:



Before you ask, the mouse had already died from my abuse... About time I gave it a chance to fight at valhalla.
*salutes* :cry:
StephOsu

The Gambler wrote:

Timo Timo wrote:

Does anyone know of a mouse that does any of this?
Not perfect but...



Before you ask, the mouse had already died from my abuse... About time I gave it a chance to fight at valhalla.

If I may also add, the aim used for absolute positioning is very weird on a mouse grip... Everything you learned using a regular mouse won't apply here.
Are you a god
The Gambler

StephOsu wrote:

Are you a god
Naw man, I gotta Steph up my game instead.
StephOsu

The Gambler wrote:

StephOsu wrote:

Are you a god
Naw man, I gotta Steph up my game instead.
But you are a god
You fused two playstyle and create a new playstyle
The Gambler

StephOsu wrote:

But you are a god
You fused two playstyle and create a new playstyle
Shot down by the king of mice...

StephOsu
Because he don't understand the meme new meta
Believe in yourself and make it work
Fxjlk
Solution 6: Send feedback to gaming mouse manufacturers about this problem?

If we tell them they can use it as a marketing tool to sell more mice they might create a solution for it. Already sent a message to Razer and I will probably tell any other companies I can think of.

Mice will probably have to have a second sensor to detect rotation.
Fxjlk
Heres my template in case you guys want one:

Hi

I have a problem playing the game Osu, I get something that's called mouse drift. Its a big problem in game and currently there are no solutions to this problem so most people use tablets. If #company# created a gaming mouse that solved this problem it would put #company# way ahead of the competition and I would definitely purchase it.

Thanks,

#your name#

If you guys have a better template or can improve this one it would be helpful cause my English is pretty bad XD
chainpullz

M3ATL0V3R wrote:

Heres my template in case you guys want one:

Hi

I have a problem playing the game Osu, I get something that's called mouse drift. Its a big problem in game and currently there are no solutions to this problem so most people use a tablet. If #company# created a gaming mouse that solved this problem it would put #company# way ahead of the competition and I would definitely purchase it.

Thanks,

#your name#

If you guys have a better template or can improve this one it would be helpful cause my English is pretty bad XD
Uh, there's not a fix. I was talking with Narrill the other day about the one driver for tablet that people claim has no input lag and I have suspicions that the "input lag" on other drivers is actually masking a forward correction algorithm thus why the cursor on this "lagless" driver seems so damn jittery. It's basically just an inherent fault of sensors. They have error. Unless you literally constrict the movement to within an array of sensors there really isn't a way you eliminate drift. Even then you're just reducing it to negligible levels instead of actually removing it. That is, a driftless mouse would just be a tablet.
Fxjlk

chainpullz wrote:

Uh, there's not a fix. I was talking with Narrill the other day about the one driver for tablet that people claim has no input lag and I have suspicions that the "input lag" on other drivers is actually masking a forward correction algorithm thus why the cursor on this "lagless" driver seems so damn jittery. It's basically just an inherent fault of sensors. They have error. Unless you literally constrict the movement to within an array of sensors there really isn't a way you eliminate drift. Even then you're just reducing it to negligible levels instead of actually removing it. That is, a driftless mouse would just be a tablet.
When I say there needs to be a second sensor i don't mean an array of sensors I mean there needs to be one that detects the current angle of the mouse which transforms x,y movement based on the mouse rotation.

Mouse rotation could be set to zero at the resting position where the x,y movement is not transformed.

Also I don't see how implementing this solution wouldn't remove the drift. If mouse rotation is taken into account every time the mouse reports its position drift should be pretty close to zero
chainpullz
The 2nd sensor would have angular drift. In fact, there's a decent chance higher end mice have multiple sensors. It only reduces but doesn't eliminate drift.
Fxjlk

chainpullz wrote:

The 2nd sensor would have angular drift. In fact, there's a decent chance higher end mice have multiple sensors. It only reduces but doesn't eliminate drift.
x,y Drift is due the loss of absolute positioning due to the sensors inability to detect rotation. What is the angular drift caused by?

Higher end mice don't have rotational sensors only additional x,y sensors which are only a partial fix because if you lift your mouse at any point the absolute rotation is lost and you will get drift.
Sup A Noob
Last I used a multi-sensor mouse it fucked so bad I went back to my old one.
Endie-
Mouse players have different ways of dealing with drift, and as Endaris mentioned it's only a problem on long maps without breaks. You'll just have to find a way to deal with it. Personally, I concentrate on keeping my mouse in the same angle, and when there is a little break I simply readjust to where I feel comfortable.
Oh and don't give in to the dark side (tablets....... ) #mousemasterrace
Good luck
chainpullz

M3ATL0V3R wrote:

chainpullz wrote:

The 2nd sensor would have angular drift. In fact, there's a decent chance higher end mice have multiple sensors. It only reduces but doesn't eliminate drift.
x,y Drift is due the loss of absolute positioning due to the sensors inability to detect rotation. What is the angular drift caused by?

Higher end mice don't have rotational sensors only additional x,y sensors which are only a partial fix because if you lift your mouse at any point the absolute rotation is lost and you will get drift.
Either way I think you are still missing the main point. The drift isn't actually coming from rotational issues. It's inherent to sensors in general. The only reason absolute positioning is possible with a tablet is the sheer number of sensors that surround all parts of the play area. Though each sensor itself may have similar error to a mouse sensor, the system of sensors as a whole has very little error. The actual space in which you can spread multiple sensors in a mouse is so small that you cannot achieve much in the way of error correction from adding additional sensors. You aren't onto something revolutionary. You are just misunderstanding the underlying issue.
KanoSet
if you notice that your mouse is drifting in a direction make it drift to the opposed direction yourself
Fxjlk

chainpullz wrote:

Either way I think you are still missing the main point. The drift isn't actually coming from rotational issues. It's inherent to sensors in general. The only reason absolute positioning is possible with a tablet is the sheer number of sensors that surround all parts of the play area. Though each sensor itself may have similar error to a mouse sensor, the system of sensors as a whole has very little error. The actual space in which you can spread multiple sensors in a mouse is so small that you cannot achieve much in the way of error correction from adding additional sensors. You aren't onto something revolutionary. You are just misunderstanding the underlying issue.
No I don't think I'm misunderstanding the underlying issue, you're saying that drift is due to sensor error but it is not. As I explained before and the diagram at the beginning clearly shows, drift is due to rotation which the mouse cannot account for. This drift is not due to random error

chainpullz wrote:

Uh, there's not a fix. I was talking with Narrill the other day about the one driver for tablet that people claim has no input lag and I have suspicions that the "input lag" on other drivers is actually masking a forward correction algorithm thus why the cursor on this "lagless" driver seems so damn jittery. It's basically just an inherent fault of sensors. They have error. Unless you literally constrict the movement to within an array of sensors there really isn't a way you eliminate drift. Even then you're just reducing it to negligible levels instead of actually removing it. That is, a driftless mouse would just be a tablet.
An input lag of 0 is physically impossible, also I don't understand how forward correction has anything to do sensor error. Input lag is to do with the travel time from the sensor not sensor error
N0thingSpecial
tl;dr play more and deal with it
chainpullz

M3ATL0V3R wrote:

chainpullz wrote:

Either way I think you are still missing the main point. The drift isn't actually coming from rotational issues. It's inherent to sensors in general. The only reason absolute positioning is possible with a tablet is the sheer number of sensors that surround all parts of the play area. Though each sensor itself may have similar error to a mouse sensor, the system of sensors as a whole has very little error. The actual space in which you can spread multiple sensors in a mouse is so small that you cannot achieve much in the way of error correction from adding additional sensors. You aren't onto something revolutionary. You are just misunderstanding the underlying issue.
No I don't think I'm misunderstanding the underlying issue, you're saying that drift is due to sensor error but it is not. As I explained before and the diagram at the beginning clearly shows, drift is due to rotation which the mouse cannot account for. This drift is not due to random error

chainpullz wrote:

Uh, there's not a fix. I was talking with Narrill the other day about the one driver for tablet that people claim has no input lag and I have suspicions that the "input lag" on other drivers is actually masking a forward correction algorithm thus why the cursor on this "lagless" driver seems so damn jittery. It's basically just an inherent fault of sensors. They have error. Unless you literally constrict the movement to within an array of sensors there really isn't a way you eliminate drift. Even then you're just reducing it to negligible levels instead of actually removing it. That is, a driftless mouse would just be a tablet.
An input lag of 0 is physically impossible, also I don't understand how forward correction has anything to do sensor error. Input lag is to do with the travel time from the sensor not sensor error
I'm not going to waste any more time trying to spell it out to you but the 3d coordinate systems with 0-2 rotational axis are all isomorphic to each other. Likewise for 2d since 2d is merely a projection of 3d. Thus adding a rotational sensor is no different from adding another translational sensor.
Fxjlk

chainpullz wrote:

I'm not going to waste any more time trying to spell it out to you but the 3d coordinate systems with 0-2 rotational axis are all isomorphic to each other. Likewise for 2d since 2d is merely a projection of 3d. Thus adding a rotational sensor is no different from adding another translational sensor.
It is different because a translational sensor requires a surface, a rotational one does not. People could lift the mouse off the surface and change the rotation which would create drift which is why current multi sensor mice are so awkward.
dung eater
Either do not tilt your mouse while you play, or tilt it in a consistent way to negate the drift when doing opposite motions. The latter might cause problems when you do some unusual patterns of movement for a while. You will develop a mix of both if you just keep playing more.

A mousepad/play area where you usually use just the middle 1/9 or less is pretty good for managing drift.
Cowbrowncow
Interesting thread I've come across here...as a rank 18k mouse player I've always had a bit of an issue not even with drift, but mainly with grip. I fixed the drift problem early on by basically hypertraining my circles at every break.

To explain...when I started playing, I do remember having a slight drift problem at first. I ended up increasing my sensitivity rapidly as my skill grew, and I capped at my current 1.05 which seems to be the most comfortable for both full screen jumps and controlled streams/rhythms. So my first solution really would be to get used to a decently high sensitivity (therefore decreasing your play area, which really really helps with drift).

However, training your spin is -key- to creating good mouse aim (or at least good mouse -stability-) in my opinion. The more perfect you can consistently spin, the more your muscles train toward keeping your mouse at the perfect angle. The ONLY problem I have now, as stated earlier, is sometimes on longer songs my grip gravitates/slips very slowly toward the back of the mouse over time. Eventually, it interferes with my aim. I've begun to remedy this simply by always using a tighter grip than I usually would on other games. It's strengthened my hand, and in turn also increased my speed.

The only other thing I can think of is I don't place my entire palm on the mouse. As a matter of fact, only my fingertips touch it, with the side of my thumb and side of my pinkie on their respective sides for stability. I also place the bottom of my palm -firmly- on the mousepad, using the area in front of it for play. I'd say I play with my wrist, but...honestly because of my palm being lifted I use a lot of those palm muscles, as well as a few arm muscles that connect to my fingers.

To conclude, I think the best "fix" to this problem is to get used to keeping the angle of your mouse consistent. Very weird explanation maybe, idk if it's unique or not, but that's the best way I can explain my style. I guess it works? For me at least xD
Topic Starter
Mio Winter

chainpullz wrote:

M3ATL0V3R wrote:

x,y Drift is due the loss of absolute positioning due to the sensors inability to detect rotation. What is the angular drift caused by?

Higher end mice don't have rotational sensors only additional x,y sensors which are only a partial fix because if you lift your mouse at any point the absolute rotation is lost and you will get drift.
Either way I think you are still missing the main point. The drift isn't actually coming from rotational issues. It's inherent to sensors in general. The only reason absolute positioning is possible with a tablet is the sheer number of sensors that surround all parts of the play area. Though each sensor itself may have similar error to a mouse sensor, the system of sensors as a whole has very little error. The actual space in which you can spread multiple sensors in a mouse is so small that you cannot achieve much in the way of error correction from adding additional sensors. You aren't onto something revolutionary. You are just misunderstanding the underlying issue.
I get how drift can come from a mouse that fails to track distance travelled consistently. So if I move my mouse 5 cm to the right, and the sensor accidentally only detects 4 cm, and then I move it 5 cm back, and my mouse detects 5 cm, then I'll have drifted 1 cm to the right. These sensor misses won't be biased in any direction, however, so if they're small enough, they'll cancel each other out over time. If the sensor misses are frequent and large, then we'd expect to notice that the mouse has drifted into an awkward position in any direction. What I do see, however, is that my mouse drifts consistently to my right and upwards. Since sensor drift is random, and my observed drift is not random, my observed drift is probably due to something other than sensor drift.

Yeah?

And would higher polling on the mouse reduce random sensor drift? It seems likely to me, but I'm far from an expert.
Topic Starter
Mio Winter

Cowbrowncow wrote:

Interesting thread I've come across here...as a rank 18k mouse player I've always had a bit of an issue not even with drift, but mainly with grip. I fixed the drift problem early on by basically hypertraining my circles at every break.

To explain...when I started playing, I do remember having a slight drift problem at first. I ended up increasing my sensitivity rapidly as my skill grew, and I capped at my current 1.05 which seems to be the most comfortable for both full screen jumps and controlled streams/rhythms. So my first solution really would be to get used to a decently high sensitivity (therefore decreasing your play area, which really really helps with drift).

However, training your spin is -key- to creating good mouse aim (or at least good mouse -stability-) in my opinion. The more perfect you can consistently spin, the more your muscles train toward keeping your mouse at the perfect angle. The ONLY problem I have now, as stated earlier, is sometimes on longer songs my grip gravitates/slips very slowly toward the back of the mouse over time. Eventually, it interferes with my aim. I've begun to remedy this simply by always using a tighter grip than I usually would on other games. It's strengthened my hand, and in turn also increased my speed.

The only other thing I can think of is I don't place my entire palm on the mouse. As a matter of fact, only my fingertips touch it, with the side of my thumb and side of my pinkie on their respective sides for stability. I also place the bottom of my palm -firmly- on the mousepad, using the area in front of it for play. I'd say I play with my wrist, but...honestly because of my palm being lifted I use a lot of those palm muscles, as well as a few arm muscles that connect to my fingers.

To conclude, I think the best "fix" to this problem is to get used to keeping the angle of your mouse consistent. Very weird explanation maybe, idk if it's unique or not, but that's the best way I can explain my style. I guess it works? For me at least xD
Interesting. Training spinners to get better at not rotating your mouse. I'll have to test it.

I haven't thought about it sufficiently, but my current leading hypothesis is that by letting your fingers and wrist take care of the movement of the mouse, I'm better able to learn fine and accurate movement. My arm is much heavier and, it seems to me, not built for as fine and precise control as my fingers are. If I try to move my mouse with my arm rapidly, it's going to be hard to stop the mouse at precisely the desired spot, for the same reason a large and heavy car can't reduce its speed from 100 km/h to 0 immediately: inertia.

But it also seems to me that if you let your fingers and wrist take care of mouse movement, you tend to rotate the mouse more, thus creating more drift than what you would if you let your arm do the moving. Given the benefits of finger/wrist movements I explained above, switching to moving with my arm is not an option for me.
Topic Starter
Mio Winter

The Gambler wrote:

Timo Timo wrote:

Does anyone know of a mouse that does any of this?
Not perfect but...



Before you ask, the mouse had already died from my abuse... About time I gave it a chance to fight at valhalla.

If I may also add, the aim used for absolute positioning is very weird on a mouse grip... Everything you learned using a regular mouse won't apply here.

This is so brilliant. D: +10 Timopoints to you.

Have you actually tried playing like this? : P If it were to actually work, I assume you would need a larger tablet to move your mouse/pen on.
Topic Starter
Mio Winter

Xyrus wrote:

Solution 5: You accept that mouse drift is an inevitable problem on longer maps and work on compensating for it.
Just a general life tip: Only accept the inevitability of something until you've explored all ways of fixing it. : P Perhaps it is inevitable, but, for myself, I'm not done trying to find solutions.
Phobos

chainpullz wrote:

I'm not going to waste any more time trying to spell it out to you but the 3d coordinate systems with 0-2 rotational axis are all isomorphic to each other. Likewise for 2d since 2d is merely a projection of 3d. Thus adding a rotational sensor is no different from adding another translational sensor.
Sorry to break it to you chainpullz, but m3atlov3r is right in this case. I think you're confusing some issues here, just because a sensor is incremental (mouse) rather than absolute (tablet) doesn't mean it cant track accurately in an absolute reference frame. High accuracy machine tools almost always use incremental sensors because they're simpler than absolute sensors. Granted the sensors they use are way more consistent than mouse optical sensors, but they actually work in pretty much the same way. You just need to know where you started, then you know where you are based how far your sensor says you've traveled.

There will be some drift over time due to the inaccuracy of a mouse sensor, but with the good sensors we have in many mice these days, this effect is pretty small and takes a relatively long time to become a problem. This video does a good job of showing this effect for a couple of mice, and I'm not convinced that even these drifts are entirely due to the sensors or could be related to their testing jig:
https://youtu.be/wQxw-pX4dak?t=141

But even if you had a perfect sensor with absolutely no inaccuracy at all, you'd still get mouse drift, as in enough mouse drift that you could run out of mousepad after playing for 30 seconds, depending on the song/sensitivity/how you move your arm. The reason is as m3atlov3r says, the coordinate system in which the mouse is measuring movement rotates with the mouse. Timo Timo's original post does a good job of illustrating this effect. Of course when we're playing we're not rotating the mouse 90 degrees, but we are rotating it smaller increments.

The mouse has 3 degrees of freedom, 2 translational and 1 rotational. Current optical sensors only read translational movement (there is no reason why they couldn't detect rotational movement, its just not implemented), so that leaves one degree of freedom unaccounted for, which results in the drift. If you measure the rotation as well, either by reading rotation from the existing sensor, or adding another translational sensor spaced a few cm away from the original, you could detect rotation and account for all 3 degrees of freedom. You could then use that information to track the mouse in an absolute reference frame instead of one that rotates with the mouse.

This does not eliminate the inherent inaccuracy of the sensor, but would make mouse drift a nonissue unless you're playing 30 minute marathons that have no breaks. Oooooooor you could just use a tablet and not worry about all this bullshit
Fxjlk

Timo Timo wrote:

What I do see, however, is that my mouse drifts consistently to my right and upwards. Since sensor drift is random, and my observed drift is not random, my observed drift is probably due to something other than sensor drift.
Is the drift in the same direction for every map? Because I get drift for certain maps in the same direction too, especially ones that don't let me re position so easily. Some maps its away from my palm yet others drift my mouse towards it
Endaris
Many people like to spam clockwise patterns in their maps and rarely go counterclockwise (lefthanded players the other way around). That's my personal observation at least. Might be faulty.
Fxjlk

Phobos wrote:

There will be some drift over time due to the inaccuracy of a mouse sensor, but with the good sensors we have in many mice these days, this effect is pretty small and takes a relatively long time to become a problem. This video does a good job of showing this effect for a couple of mice, and I'm not convinced that even these drifts are entirely due to the sensors or could be related to their testing jig:
https://youtu.be/wQxw-pX4dak?t=141
That video is about wireless mouse tracking in a high interference environment so I'm not sure if that drift is due to the interference or sensor inaccuracy. He did say later on in the video that the testing jig is very well isolated from the sensor so I don't think the jig causes any drift.
show more
Please sign in to reply.

New reply