Overclock.net › Forums › Components › Mice › s3988 vs 3310 vs 3360 Technical Differences?
New Posts  All Forums:Forum Nav:

s3988 vs 3310 vs 3360 Technical Differences?

post #1 of 38
Thread Starter 
I understand that there have been a lot of posts of people debating, but all of the information is scattered through hundreds of different posts, and a lot of important information is kind of lost in heaps of comments that flood virtually all forum posts on nearly every website I've seen on this topic (maybe I just haven't looked hard enough). I don't have a lot of technical knowledge on this type of thing, and this is also a controversial(?) topic. I'd like to appeal to the people that actually know what they're talking about, rather than the ones that go by "feel."

Here's what I at least THINK I know so far (do correct me if I'm wrong and feel free to discuss relevant information that's missing)

3310
-Frame rate is ~6k
-malfunction speed ~5m/s(?)
-3-4m/s small amounts acceleration(?)
-Every CPI step of 50 is native
-LOD of 1-2mm(?)

3310 questions:
-I've found some information regarding "perfect control speeds," and I'm curious to know if the acceleration is present before reaching ~4m/s in minuscule amounts. Is there a perfect steady movement that the ~6k framerate can keep up with? Is this dependent on the CPI? I suppose this is also a question for the higher frame rate sensors.
-Is every 50 step "truly" native? I seem to recall seeing people talking about different "native" CPI performing better or worse than they would on a different sensor. What is the "true" native for the 3310?
-Because of its lower frame rate, does it have I guess what I would have to refer to as "input skipping?" By "input skipping" I guess I mean the 6k frame rate not matching up with the 500/1000hz that the mouse is set to, so it would wait until the next update to the computer before registering the movement from the sensor? Would the movement found by the sensor be lost instead of waiting for the next update (I'm assuming no)? If this IS the case, would it not be better to have the frame rate of the sensor to be lowered to the polling rate of the mouse to the computer? Again, I'm ASSUMING it just waits for the next update, similar to how Vsync works on monitors. But hey, I don't know anything. I suppose this question also applies to the others.

s3988/s3989
-Frame rate is ~12.5k
-Potentially lower LOD than the 3310(???)
-Native CPI is 1800/6400(?)

s3988/s3989 questions:
-I'm EXTREMELY curious about the whole native 1800 cpi thing. Is this dependent on the lens that is used? Are there any implementations of this sensor that have different native CPI setting(s) due to a different lens? What about less smoothing at lower CPI? If I recall, there was a thread which discussed the difference in delay between certain CPI, and the lowest amount of delay for this sensor was sub 1600. If this is true, why is 1800 considered to be the native? What is the interpolation method used to achieve different settings on this sensor? What are the downsides of not using 1800 cpi? This is also a question for the other sensors, but more so for this one.
-What are the benefits of having 12.5k framerate over ~6k? If this has consistently higher frame rate than even the 3360, does it have less acceleration? What are the factors that go into acceleration? If it's at least mostly determined by the frame rate, why do people generally prefer the 3360/3366 to this one? As far as snappiness/responsiveness goes, would this not be almost entirely determined by the polling rate of the mouse rather than the frame rate of the sensor? Are all of the people saying it's snappier unfortunately suffering from a placebo?

3360/3366
-Frame rate is inconsistently anywhere from 6k-12k depending on speed of the mouse
-Similar/same LOD as 3310(?)
-Every CPI step of 50 is native

3360/3366 questions:
-How long does the sensor take to switch from higher to lower frame rates? Is there a delay/skip that happens in tracking as it does so?
-Is the main/only benefit of the inconsistent frame rate less power consumption for wireless mice?
-Why do people generally say this sensor is the best of the 3? Why is it more accurate than the others? I've seen the SRAV tests for this sensor, but haven't delved into those tests for something like the s3988. Is the 3360 better at higher speeds? I guess I'm asking this because I'm a low sens user that has to do lightning movements just to turn around. Would the s3988 be better or worse than the 3360 specifically for low sens users like me?
-Does the mouse require software to gain access to the changes in frame rate?


I know plenty of people have disagreements on which sensor is the best, but I'm curious to know what the more tech savvy lot has to say about this kind of thing, rather than browsing through an infinite pool of bias and subjectivity. I also understand that the implementation from mouse to mouse is very important, so if specific examples or references to certain mice are required to make comparisons, by all means do so.
post #2 of 38
This is stuff I want to know too, at this point I'm mainly interested in the 3360 for the sensor possibly being the most responsive, without any latency or loss of tracking.

But really, it depends on the company, and how well they do with the implementation. Which you would think wouldn't be hard, just get rid of any weird calculations and keep it as raw as possible.
post #3 of 38
Thread Starter 
Totally guessing because I have no idea what I'm talking about here, but isn't having fewer calculations one of the reasons behind the MLT04 not functioning very well at higher speeds? At lower speeds, I assume it doesn't take very long to process and seems very "responsive" because of this, but wouldn't the difference be virtually negligible at that point and still be mostly determined by the polling rate rather than the sensor itself? The delay could be more than 1 refresh from the polling rate which is when it would matter, but aren't all of the modern sensors still on an extremely low delay? I doubt anyone is going to feel the difference between something like 4ms and 7ms (disposable subjective garbage, but still relevant).
post #4 of 38
3310
variable framerate depending on speed (3modes, very roughly 2000, 4000, 7000)
same smoothing on every step (~4ms, lower at higher framerate modes)
native steps in increments of 50
pcs aroudn 6-7m/s

3988
variable framerate depending on speed (3modes, very roughly 3000, 7000, 12500)
variable smoothing depending on dpi setting. at <= 1550dpi it's the same as 3310, above it's more. (someone please find the post where i wrote all of the estimated values. i'm stuck in this guesthouse with 0.2mbps internet)
native steps of 50. for the razer abyssus14, if you use synapse, the sensor is set to the highest dpi and every step is scaled down from firmware and/or drivers. of course this is an issue with the mouse/driver not the 3988 sensor.
pcs around 7-8m/s

3360/3366
(possible) sensor rattlelelelelelelelel
variable framerate depending on speed (4modes, roughly 4100, 4900, 5900, 12000)
native steps are 50 to 12800 in increments of 50 for 3366. logitech limits their mice to 200-12000 for whatever reason
native steps are 100 to 12000 in increments of 100 for 3360.
infinite pcs. no seriously though i managed to get my g900 to 10m/s... so go figure

all of them have variable framerate. not sure about 3310/3988 but for 3360/3366 there's no way to control the framerate; it's completely automatic. i suspect it's the same for 3310/3988 as every mousetester plot i've seen shows variable framerate for those.

there is no delay for framerate changes. but for sensors with smoothing, there will be a skip/jump on framerate transitions. i posted about this in the thread "visualizing smoothing in mousetester"; should be pretty easy to find (sorry no link, again i don't want to click around with this 0.2mbps internet tongue.gif). for 3988 i can literally see the skipping on the windows desktop when i use >=3200dpi and scale down the windows sensitivity.

no 100% reputable source for this, but SRAV is better on 336x than on 3310 or 3988. although for srav estimates done by hand (the swipe-fast-right-swipe-slow-left test), i suspect having feet near the sensor makes a bigger impact than the differences between the sensors.

also srav for 336x should be more consistent from mouse to mouse because the 336x led is integrated, whereas for 3310/3988 it's up to the manufacturer to decide which led to use and how to ensure it's aligned with the lens.

given the way math works, higher framerate generally allows for smoother tracking. e.g. with 12500, you get either 12 or 13 frames per usb poll (assume 1ms), whereas with 2500, you get either 2 or 3 frames per usb poll. of course this rule is broken if framerate is exactly a multiple of the usb polling rate. imo sensor differences affect things more though. e.g. i think a 3366's low speed tracking is smoother than a 3090's tracking even though 3090 gets 6500 and 3366 is <5000 for low speeds.
yes it doesn't directly affect responsiveness much. some people may find smoother tracking to be more responsive though.
Edited by qsxcv - 6/22/16 at 8:35am
main
(15 items)
 
old
(14 items)
 
 
CPUMotherboardGraphicsRAM
4770k maximus vii impact nvidia gtx 970 crucial ballistix tactical 16gb 
Hard DriveCoolingOSMonitor
crucial mx100 noctua nh-c14 windows 7 ultimate sony cpd-g520 
KeyboardPowerCaseMouse
kbp v80 matias quiet silverstone sx500-lg ncase m1 v3 logitech g100s with mcu replaced by teensy2.0 
Mouse PadAudioAudio
allsop raindrop xl chord mojo hifiman re-600 
CPUMotherboardGraphicsRAM
i7 920 evga x58 sli le galaxy gtx 460 crucial something 3x1gb 
Hard DriveCoolingOSMonitor
intel 330 180gb scythe kotetsu windows 8.1 pro sony cpd-g520 
KeyboardPowerCaseMouse
logitech k120 silverstone st75f-gs nxzt h440 evga torq x5 
Mouse PadAudio
allsop raindrop mobo 
  hide details  
Reply
main
(15 items)
 
old
(14 items)
 
 
CPUMotherboardGraphicsRAM
4770k maximus vii impact nvidia gtx 970 crucial ballistix tactical 16gb 
Hard DriveCoolingOSMonitor
crucial mx100 noctua nh-c14 windows 7 ultimate sony cpd-g520 
KeyboardPowerCaseMouse
kbp v80 matias quiet silverstone sx500-lg ncase m1 v3 logitech g100s with mcu replaced by teensy2.0 
Mouse PadAudioAudio
allsop raindrop xl chord mojo hifiman re-600 
CPUMotherboardGraphicsRAM
i7 920 evga x58 sli le galaxy gtx 460 crucial something 3x1gb 
Hard DriveCoolingOSMonitor
intel 330 180gb scythe kotetsu windows 8.1 pro sony cpd-g520 
KeyboardPowerCaseMouse
logitech k120 silverstone st75f-gs nxzt h440 evga torq x5 
Mouse PadAudio
allsop raindrop mobo 
  hide details  
Reply
post #5 of 38
Thread Starter 
This is by far one of the best responses I have seen on any post so far. Thank you. How/why would the tracking be different with the placement of the mouse feet? You just turned my whole world upside down.
post #6 of 38
see pg 13 of https://media.digikey.com/pdf/Data%20Sheets/Avago%20PDFs/ADNS-3090.pdf
since the lens magnification (and hence the actual sensor resolution) depends on tracking height, when you press the mouse, on a soft pad, the resolution will increase as the regular mousefeet will push down on the pad and the tracked surface will be raised relative to the sensor/lens. but if you have feet near the sensor, that feet will also push down the pad, so it keeps the distance between the sensor/lens and the tracked surface more constant

it's a minor thing but it's comparable in magnitude to the srav of optical sensors
Edited by qsxcv - 6/22/16 at 8:50am
main
(15 items)
 
old
(14 items)
 
 
CPUMotherboardGraphicsRAM
4770k maximus vii impact nvidia gtx 970 crucial ballistix tactical 16gb 
Hard DriveCoolingOSMonitor
crucial mx100 noctua nh-c14 windows 7 ultimate sony cpd-g520 
KeyboardPowerCaseMouse
kbp v80 matias quiet silverstone sx500-lg ncase m1 v3 logitech g100s with mcu replaced by teensy2.0 
Mouse PadAudioAudio
allsop raindrop xl chord mojo hifiman re-600 
CPUMotherboardGraphicsRAM
i7 920 evga x58 sli le galaxy gtx 460 crucial something 3x1gb 
Hard DriveCoolingOSMonitor
intel 330 180gb scythe kotetsu windows 8.1 pro sony cpd-g520 
KeyboardPowerCaseMouse
logitech k120 silverstone st75f-gs nxzt h440 evga torq x5 
Mouse PadAudio
allsop raindrop mobo 
  hide details  
Reply
main
(15 items)
 
old
(14 items)
 
 
CPUMotherboardGraphicsRAM
4770k maximus vii impact nvidia gtx 970 crucial ballistix tactical 16gb 
Hard DriveCoolingOSMonitor
crucial mx100 noctua nh-c14 windows 7 ultimate sony cpd-g520 
KeyboardPowerCaseMouse
kbp v80 matias quiet silverstone sx500-lg ncase m1 v3 logitech g100s with mcu replaced by teensy2.0 
Mouse PadAudioAudio
allsop raindrop xl chord mojo hifiman re-600 
CPUMotherboardGraphicsRAM
i7 920 evga x58 sli le galaxy gtx 460 crucial something 3x1gb 
Hard DriveCoolingOSMonitor
intel 330 180gb scythe kotetsu windows 8.1 pro sony cpd-g520 
KeyboardPowerCaseMouse
logitech k120 silverstone st75f-gs nxzt h440 evga torq x5 
Mouse PadAudio
allsop raindrop mobo 
  hide details  
Reply
post #7 of 38
In non scientific stuff from personal use

3360 > 3310 because
Literally no acceleration, I tried and tried again.. with the 3366 I land in the same spot everytime
higher PCS - I can make a 3310 malfunction in real usage (but you can avoid it by not going too nuts lol)
Umm as for feeling I guess it just feels a little better no big difference there
And 3366 seems to handle 1000hz better although I think I'm going to use all my mice is 500hz mode from now on, 500 seems to be more 'consistent' if that's what you're looking for

Hope this helps... kind of lol
post #8 of 38
qsxcv does the MLT04 remain at 9000 frame rate when active?
post #9 of 38
Quote:
Originally Posted by ncck View Post

higher PCS - I can make a 3310 malfunction in real usage (but you can avoid it by not going too nuts lol)
you sure? you really really really need to try in order to hit >6m/s

it may be more susceptible to malfunctioning if you begin a swipe midair before the mouse touches the pad (i.e. lifting the mouse)

all sensors can handle any hz fine. it's entirely up to the firmware.
Quote:
Originally Posted by realistic01 View Post

qsxcv does the MLT04 remain at 9000 frame rate when active?
ya
Edited by qsxcv - 6/22/16 at 9:08am
main
(15 items)
 
old
(14 items)
 
 
CPUMotherboardGraphicsRAM
4770k maximus vii impact nvidia gtx 970 crucial ballistix tactical 16gb 
Hard DriveCoolingOSMonitor
crucial mx100 noctua nh-c14 windows 7 ultimate sony cpd-g520 
KeyboardPowerCaseMouse
kbp v80 matias quiet silverstone sx500-lg ncase m1 v3 logitech g100s with mcu replaced by teensy2.0 
Mouse PadAudioAudio
allsop raindrop xl chord mojo hifiman re-600 
CPUMotherboardGraphicsRAM
i7 920 evga x58 sli le galaxy gtx 460 crucial something 3x1gb 
Hard DriveCoolingOSMonitor
intel 330 180gb scythe kotetsu windows 8.1 pro sony cpd-g520 
KeyboardPowerCaseMouse
logitech k120 silverstone st75f-gs nxzt h440 evga torq x5 
Mouse PadAudio
allsop raindrop mobo 
  hide details  
Reply
main
(15 items)
 
old
(14 items)
 
 
CPUMotherboardGraphicsRAM
4770k maximus vii impact nvidia gtx 970 crucial ballistix tactical 16gb 
Hard DriveCoolingOSMonitor
crucial mx100 noctua nh-c14 windows 7 ultimate sony cpd-g520 
KeyboardPowerCaseMouse
kbp v80 matias quiet silverstone sx500-lg ncase m1 v3 logitech g100s with mcu replaced by teensy2.0 
Mouse PadAudioAudio
allsop raindrop xl chord mojo hifiman re-600 
CPUMotherboardGraphicsRAM
i7 920 evga x58 sli le galaxy gtx 460 crucial something 3x1gb 
Hard DriveCoolingOSMonitor
intel 330 180gb scythe kotetsu windows 8.1 pro sony cpd-g520 
KeyboardPowerCaseMouse
logitech k120 silverstone st75f-gs nxzt h440 evga torq x5 
Mouse PadAudio
allsop raindrop mobo 
  hide details  
Reply
post #10 of 38
Quote:
Originally Posted by qsxcv View Post

3988
variable framerate depending on speed (3modes, very roughly 3000, 7000, 12500)
variable smoothing depending on dpi setting. at <= 1550dpi it's the same as 3310, above it's more. (someone please find the post where i wrote all of the estimated values. i'm stuck in this guesthouse with 0.2mbps internet)
native steps of 50. for the razer abyssus14, if you use synapse, the sensor is set to the highest dpi and every step is scaled down from firmware and/or drivers. of course this is an issue with the mouse/driver not the 3988 sensor.
pcs around 7-8m/s

I'm pretty sure the Abyssus 2014 is 3888 and not 3988.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Mice
Overclock.net › Forums › Components › Mice › s3988 vs 3310 vs 3360 Technical Differences?