Overclock.net banner

41 - 60 of 138 Posts

·
Mouse addict
Joined
·
787 Posts
Quote:
Originally Posted by raad11 View Post

Does the 3360 have smoothing? I have the Nixeus Revel which has a firmware update to remove any smoothing, so assuming it has none... how many milliseconds quicker is 12,000 dpi than let's say 1600 dpi in reacting to acceleration from zero? Should we just use the maximum DPI setting or is there a point of diminishing returns?
That firmware update only removed the MCU smoothing which is not native to the 3360. The 3360 has 32 frames of smoothing at or above 2100 CPI.
 

·
Registered
Joined
·
1,679 Posts
Quote:
Originally Posted by cdcd View Post

That firmware update only removed the MCU smoothing which is not native to the 3360. The 3360 has 32 frames of smoothing at or above 2100 CPI.
How do these smoothing values go with the 3310?
 

·
Mouse addict
Joined
·
787 Posts
Quote:
Originally Posted by Leopardi View Post

How do these smoothing values go with the 3310?
20 frames across all CPI ranges, therefore what applies to the 3366 would apply to the 3310 to a lesser degree as well.
 

·
Registered
Joined
·
72 Posts
To add on to my previous post and maybe make a tl;dr, if you impart any amount of force on your mouse which breaks static friction with the mousepad, the difference in the time it takes for the sensor to detect that movement at 400 dpi and at 3200 dpi is, in the absolute worst case, 1ms (1 polling cycle). I say this because we still have people talking about 2-3 ms delay.
 

·
Registered
Joined
·
359 Posts
Quote:
Originally Posted by wareya View Post

The 3366 one frame of smoothing with updated SROMs. This smoothing is not noise cancellation. Yes, it's extremely small, like so small that it's not perceivable, but it's still there, and it is not able to be called anything else. It's literally an averaging of sensor motion deltas. That's smoothing.
According to my conversations with Francois, it's not averaging motion data. It's related to fixing the cursor shooting north. Everyone complained about that, and now it doesn't happen anymore. Regardless, it's not smoothing.
Quote:
Originally Posted by Conditioned View Post

I don't see anyone here having done either, at least in this thread. I think the question is related to your knowledge, and that is quite unrelated to your intelligence or personality.
The statement I reacted to was that I don't understand what the engineers are telling me. Hard to read that as something other than commentary on my intelligence, or a subtle suggestion that I'm just making stuff up (character).
 

·
Registered
Joined
·
1,699 Posts
Quote:
Originally Posted by CPate View Post

According to my conversations with Francois, it's not averaging motion data. It's related to fixing the cursor shooting north. Everyone complained about that, and now it doesn't happen anymore. Regardless, it's not smoothing.
"Shooting north"? You mean malfunctioning.

In any case, it is 100% certain that there is smoothing. I see no reason that blurring the motion data would make it stop malfunctioning, unless there's a bug in the 3366's silicon that makes it malfunction if it doesn't undergo the smoothing step or something stupid like that.

But it is 100% certain that it does that smoothing step. Smoothing is what it is. It is averaging the last two frames of motion data. It really is. 100%. Period. We measured it. The measurement is objective. It's mathematically impossible for the measurement to represent anything other than smoothing. "It's not smoothing" is just a falsehood. If an engineer thinks it's not doing smoothing, they didn't figure out what the new SROMs are actually doing.
 

·
Registered
Joined
·
359 Posts
Seriously there's no advantage in lying for me, and as a representative of Logitech I'm not permitted to. I'm told it's not smoothing by a guy who helped design it. Believe whatever you want.
 

·
Registered
Joined
·
1,699 Posts
Quote:
Originally Posted by qsxcv View Post

This, basically. It literally cannot be anything but smoothing. Smoothing is the only thing it can be.
Quote:
Originally Posted by CPate View Post

Seriously there's no advantage in lying for me, and as a representative of Logitech I'm not permitted to. I'm told it's not smoothing by a guy who helped design it. Believe whatever you want.
That guy did not design the sensor. Pixart did. In fact, early SROMs don't have the smoothing. Only the newest (?) one does.

----

Edit, full explanation:

Those jumps down are framerate transitions. When the framerate changes, the amount of physical motion encoded in each frame is reduced. For the first three framerates, the plot is one dot per frame, instead of one dot per 1/8000th of a second.

When the framerate increases, motion data from the previous frame is a lot higher than the motion data for the new frame, because the new frame took up less time than the previous one. This makes the plot drop.

This drop should be 100% clean, and with sensors with zero smoothing, it is. If this drop is not 100% clean, that means that old frames are being mixed into the current frame somehow.

In the mathematics of filters, in order to mix frames together without adding latency (group delay), you need to be able to see the future. Mice can't see the future. So when mice mix frames together, they're weighted slightly towards older motions, which increases latency.

In the case of the 3366, this latency is significantly less than a single millisecond, and it's less latency than is added by holding the sensor's motion data until the next USB frame, but it's still added latency, and it's still smoothing.

That you can't possibly feel it, or even be affected by it ingame because the difference it makes is like a single pixel when moving at a high speed, doesn't make it stop being smoothing.
 

·
Registered
Joined
·
779 Posts
To the people actively debating CPate, please make an effort to be less antagonizing.

It's nice to have a direct link for communicating with Logitech's engineering and design teams. I don't think we have that privilege with any other major manufacturers, and other smaller mouse companies have done it with far less competence (*cough* FM *cough*).

But if he feels constantly disrespected in our discussions, he may very well start contributing less or stop contributing altogether.
 

·
Registered
Joined
·
153 Posts
Quote:
Originally Posted by cdcd View Post

That firmware update only removed the MCU smoothing which is not native to the 3360. The 3360 has 32 frames of smoothing at or above 2100 CPI.
Quote:
Originally Posted by cdcd View Post

20 frames across all CPI ranges, therefore what applies to the 3366 would apply to the 3310 to a lesser degree as well.
Is there a list or thread somewhere keeping track of which sensors/mice have how much smoothing?
 

·
Mouse addict
Joined
·
787 Posts
Quote:
Originally Posted by raad11 View Post

Is there a list or thread somewhere keeping track of which sensors/mice have how much smoothing?
Not that I'm aware of.
 

·
Registered
Joined
·
21 Posts
Discussion Starter · #53 ·
Quote:
Originally Posted by VolsAndJezuz View Post

To the people actively debating CPate, please make an effort to be less antagonizing.

It's nice to have a direct link for communicating with Logitech's engineering and design teams. I don't think we have that privilege with any other major manufacturers, and other smaller mouse companies have done it with far less competence (*cough* FM *cough*).

But if he feels constantly disrespected in our discussions, he may very well start contributing less or stop contributing altogether.
This.

It's fine to challenge and question the info that is given but it could be done in a more positive way that will be more beneficial.
 

·
Registered
Joined
·
1,699 Posts
When he keeps repeating something even though people keep telling him that we know that it's wrong, like it's going to change anything, what would you do? Politely tell him, again, that we know that it's wrong? No, I'd much rather post why we know it so people know he's posting marketing BS. If he didn't work for Logitech you wouldn't defend the fact that he's incessant about this at all.
 

·
Registered
Joined
·
1,262 Posts
Slightly off topic but in the 3361 did rocat just move the switch over to 32 frames smoothing from 2000 to 3000CPI?
 

·
Registered
Joined
·
404 Posts
Quote:
Originally Posted by wareya View Post

When he keeps repeating something even though people keep telling him that we know that it's wrong, like it's going to change anything, what would you do? Politely tell him, again, that we know that it's wrong? No, I'd much rather post why we know it so people know he's posting marketing BS. If he didn't work for Logitech you wouldn't defend the fact that he's incessant about this at all.
If he's wrong he's wrong and I'm all for BS being called out, but I wonder if reality is not a bit more subtle: before arguing whether or not the 3366 has smoothing, it may help if you both reach the same definition of smoothing first. Can't help but feel you two are not in agreement over that.
 

·
Ex-resident Kinzu shill
Joined
·
2,074 Posts
There is not a subjective definition of smoothing in the community though, that's the thing, smoothing is the same to pretty much every single person (except MaxKohler, who tried to argue that the definition is the same as we define...), smoothing is averaging of the sensor's motion data over a period of X, regardless of what X is, be it 2 frames or 2ms.

Smoothing could also be called interpolation of motion data, or filtering, or many other things, but we generally just call it "smoothing" to give it a standardized term.
 

·
lololol
Joined
·
4,373 Posts
Quote:
Originally Posted by CPate View Post

CPI is not just a measure of what is output to the host - it's also the measure of how far the mouse physically has to move before the sensor detects motion. An optical mouse sensor at 400 CPI can detect movements as small as 63.5 micrometers (microns), while at 3200 CPI it's 7.9375 microns. The sensor only can report integer counts, so the higher the resolution the less physical travel is required before the sensor can see and send its first integer count. When adjusting sensor CPI, there is no change in the sensor pixel counts, nor in lens magnification. Resolution scaling is done by subpixellization, which permits the sensor to detect movements that are only fraction of a physical pixel.
it's misleading to quantify as "2-3ms" or any value of time since it obviously varies with acceleration.

Quote:
Originally Posted by CPate View Post

Also, 3366 still has zero smoothing.
Quote:
Originally Posted by CPate View Post

Seriously there's no advantage in lying for me, and as a representative of Logitech I'm not permitted to. I'm told it's not smoothing by a guy who helped design it. Believe whatever you want.
i don't remember but did you ask said guy about
edit: nvm, saw above post. my response is few posts below
Quote:
Originally Posted by wareya View Post

Quote:
Originally Posted by qsxcv View Post

This, basically. It literally cannot be anything but smoothing. Smoothing is the only thing it can be.
 

·
Registered
Joined
·
1,291 Posts
Quote:
Originally Posted by nillington View Post

This doesn't add up. I'm a film grad but I'm trying my hand at math here, someone correct me if I'm wrong.

1 meter = 1,000,000 microns. 1 millisecond = 1/1000 second (1000hz). 1G = 9.81m/s^2. This means that 1G also = 9,810 microns/ms^2. If I accelerate my mouse from a standstill with a force of 1G for 1 millisecond, it will be moving with velocity 9,810 microns per millisecond after that millisecond.

Now, let's say I'm making a small movement, and I accelerate the mouse from a standstill at a rate of 0.1G, that's 981 microns/ms^2. I begin moving the mouse immediately after it has sent a report to my pc, so the next report will be sent in 1ms. In that timeframe, I move the mouse a certain distance. that distance is defined by this formula, where v = initial velocity (0), t = the time between reports (1ms), a = acceleration (981 microns/ms^2):

x = vt + ½ (a)(t)^2 ----> x = (0)(1ms) + ½(981microns/ms^2)(1ms)^2

> x = ½(981)(1) = 981 / 2 = 490.5 microns.

So even making a fairly slow movement at 0.1G, in the time it takes for one report to be sent (1ms), i move the mouse 490.5 microns. According to your post, "An optical mouse sensor at 400 CPI can detect movements as small as 63.5 micrometers (microns)". How can it take 2-3 ms longer for the mouse to detect and send that movement when the threshhold for detectable movement is almost 8 times less the amount moved in just 1ms @ 0.1G acceleration? Wouldnt the difference in time between when movement starts and when it's reported to the pc be, at worst, 1ms, especially for the 0.5G and 2G examples you referenced? 2ms TOPS if the movement is really slow, at 0.01G for instance.

For everyone's reference, at 400 cpi the sensor can detect movement of 63.5 microns; the width of the average human hair is around 100 microns. My point here is that the '2-3ms faster' being put out there doesn't make mathematical sense, and that you should just use whatever dpi you find comfortable for desktop use in-game as well. There's no need to turn it up, even if you would consider 2-3ms to be absolutely gamebreaking. I don't think I'm disagreeing with CPate's ultimate point at all, just trying to be as accurate as possible here.
You made a very common dimensional analysis mistake and your calculations are off by a factor of 1000. 9.81m/s^2 is actually 9.81microns/ms^2. You most likely forgot to convert both units of s to ms.

Edit: So if you use something like 1g of acceleration you end up with 4.91 microns after 1ms, 19.6 microns after 2ms, and 44.1 microns after 3ms.
 
41 - 60 of 138 Posts
Top