The Blur Busters Mouse Guide


A Blur Busters Feature
Written by Ben Hansen (sharknice) of EliteOwnage.com

razer-deathadder-gallery-3[1]Mouse settings become even more important when you are using high frame rate and low persistence, blur free displays such as LightBoost. With increased motion clarity, the fluidity and precision of mouse movements become much more apparent. This guide goes over the most important mouse behaviors and settings for blur-free gaming.

Polling Rate

What is Polling Rate?

The communication between a mouse and computer is not instantaneous or constant. USB mice communicate with the computer in set intervals of time. Polling rate is typically measured in polls per second also known as hertz. The default polling rate for most USB mice is 125 Hz (hertz).

Faster is Better

17697[1]If you use a higher polling rate you will have less delay between when you move your mouse and it moves on screen. When using a 125 Hz polling rate the mouse is being updated in intervals of 8 milliseconds. This means you have a potential of 0 to 8 ms of mouse lag added to every frame depending how long after the last mouse poll the frame begins rendering.

Most new gaming mice now have a 500 Hz and 1000 Hz setting. If you are using 1000 Hz polling the added delay is only 1 millisecond. (It is important to note that this is only the mouse latency, not total input lag).

More Noticeable at Higher Frame Rates

Polling rate becomes increasingly important with higher frame rate displays. If you’re using a standard 60 Hz display and 125h Hz mouse polling the delay between frames would be 16.7 ms and the delay between mouse updates would be 8 ms. The delay between mouse updates is less apparent because the delay between frames is so high. Using a 120 Hz refresh rate and 125 Hz mouse polling brings the frame rate interval down to 8.3 ms which is almost as frequent as the mouse updates.

Here is a simple example of input lag during 120 frames per second (1/120sec = 8.3ms) being compared between a 125Hz mouse and a 1000Hz mouse:

addedlag

Note: This simplified diagram excludes other parts of the whole input lag chain, such game engine latency and display latency.

As you increase your frame rate, the frequency of mouse updates per frame goes down and the delay between when you move your mouse and the effect of it on screen becomes more apparent. If you play at a higher frame rate than your mouse polling rate some frames will be updating with no new mouse input. At this point, you’re basically just viewing some frames instead of actively participating in them, and the mouse movement will be very choppy.

pollinginterval

Some games actually update the game state at a rate faster than the video frame rate to reduce input lag. These games take advantage of higher polling rates even with a lower frame rate. For example, a game could update the physical interactions between objects based on player input at 500 Hz despite only running at a 60 Hz video frame rate. In games like these a faster polling rate reduces mouse input lag even more.

Which Polling Rate Should I Use?

To decrease input lag you can increase the USB polling rate of the mouse. Many modern gaming mice support polling rates up to 1000 Hz which drastically reduces mouse lag in comparison with 125 Hz. Most gaming mice have adjustable polling through software or switches built into the hardware itself. For mice that don’t have a polling rate setting (e.g. Logitech MX518) you can often manually overclock them to a higher poll rate.

In addition, if you are using Windows 8.1, there is a system-wide 1000Hz mouse fix which serious gamers should install to bring maximum mouse performance to all games and Windows applications.

VSYNC

What Does VSYNC Do?

VSYNC synchronizes the games frame rate with the displays frame rate to reduce tearing and stuttering. As long as the game updates as fast as the monitor’s refresh rate there will be no tearing or stuttering. Without VSYNC a game will update as fast as it possible can send updates to the monitor. This can lead to tearing (see TestUFO animation) if the frame rate is faster than the monitor’s refresh rate. It also can cause stuttering because the the time the frame was created to the time it is displayed is constantly changing.

Input Lag

VSYNC can increase input lag because while using VSYNC the game will take a mouse measurement, render the frame, then wait to display that frame until the next monitor refresh. When VSYNC is off the game will constantly update with new frames without waiting for the display so the input will be fresher with less delay.

Low Persistence Displays (LightBoost)

120hz-monitorLow persistence displays include CRTs and new strobe-backlight gaming displays with LightBoostULMB and Turbo240. With a low persistence display, there is less blur to hide stuttering and tearing caused by fluctuating frame rates.

During LightBoost, mouse movement can be more fluid with VSYNC ON when running at full frame rate without frame drops. However, it is also creates added input lag.

If you’re playing a competitive game like Counter-Strike, it is typically better to use VSYNC OFF, as less input lag is more important than maximum fluidity. On the other hand, single-player games like BioShock Infinite can be more enjoyable with the fluidity provided by VSYNC, assuming you have enough GPU power to keep frame rates maxed out, with frame rates matching refresh rates.

Some games have built in frame rate limiters that behave similar to VSYNC ON, but instead of rendering the frame right away the frame is rendered at the last possible moment to reduce input lag. It does not work perfectly, but it is a good compromise between the fluidity of full frame rate VSYNC ON (120fps at 120Hz), and tearing of uncapped frame rates during VSYNC OFF. One example is the fps_max setting found in Source Engine games (e.g. Counter-Strike: GO).

G-SYNC and FreeSync

gsync-logoNVIDIA G-SYNC (and eventually, AMD FreeSync) aim to completely eliminate stuttering and tearing without increasing input lag. These technologies work best with a high mouse poll rate (1000Hz).

Using these variable refresh rate technologies, the monitor can be updated at varying intervals so a variable frame rate will not cause stuttering since the monitor will always be in sync with the games frame rate. Blur Busters’ GSYNC Preview (Part 1 and Part 2) covers more information about G-SYNC.

Addendum: Photos of 125Hz vs 500Hz vs 1000Hz

Originally created in a Blur Busters Forums thread, and now a part of this guide, this is a photo comparision of 125Hz versus 500Hz versus 1000Hz mouse poll rates. 500Hz versus 1000Hz is more clearly human-eye visible when enabling blur reduction strobing (e.g. LightBoost) as well as G-SYNC where NVIDIA recommends a 1000Hz mouse.

mouse-125vs500vs1000

You can see this by enabling motion blur reduction on your 120Hz monitor, and then drag a text window. Fewer microstutters makes text easier to read while dragging.

The gapping effect is caused by the harmonic frequency difference (beat frequency) between frame rate and mouse poll rate. It is clearly visible when no other sources of microstutters exist; e.g. fast GPU, fast CPU, low-latency USB. This mouse microstutter is clearly visible in Source Engine games on newer GPUs at synchronized framerates.

During 125Hz mouse poll rate versus 120fps frame rate (125 MOD 120 = 5), there are 5 microstutters per second. This results in 1 gap every 25 mouse arrow positions.

During 500Hz mouse poll rate versus 120fps frame rate (500 MOD 120 = 20), there are 20 microstutters per second. This results in 1 gap every 6 mouse arrow positions.

These mouse microstutters become especially visible on low-persistence displays such as strobed monitors or CRTs, during window-dragging. 500Hz vs 1000Hz difference is amplified during LightBoost, ULMB, Turbo240, and BENQ Blur Reduction.

Microsoft Agrees 1000Hz Makes A Difference

See this Microsoft Research video on 1000Hz touchscreens. This is not for mice, but it demonstrates how surprising for humans being able to detect experiences still improving at 1000Hz.

Also, 1000Hz is NOT Final Frontier!

People are now overclocking computer mice to 2000Hz+.  There are some very minor benefits, but we can confirm we notice the benefits on an ultralow-persistence monitor (e.g. ULMB).

2000hz-mice-needed

Questions?

Discuss further about gaming mice in the Blur Busters Forums!

Ben Hansen (aka sharknice) is an indie game developer, and runs EliteOwnage.com, a website with a very detailed mouse sensitivity guide.

17 comments on “The Blur Busters Mouse Guide

  1. Pingback: New Mouse Guide for the LightBoost/G-SYNC era. | Blur Busters

  2. Berserker says:

    Mice actually have more delay than just polling rate. High end sensors still have 30×30 image array and doing insane dpi’s without jitter or other tracking errors requires some anti-ripple(jitter) correction/processing or “smoothing” as people call it, its not done real time. I can easily notice latency difference between mouse with old low dpi sensor and new one.

    • ahigh says:

      Interesting. I am very curious to see your evidence. My belief was that latency for a 1000hz mouse is 2ms or less. I would be surprised if I was wrong, actually. However, I love learning new things, so tell us more?

      • Chief Blur Buster says:

        There may be mouse filtering going on in some 1000Hz mouse, but not all of them.

        If filtering is done, then filtering is mathematically simple — doing jitter filtering on one lookahead coordinate should add only 1ms extra latency. The average movement latency in a _good_ 1000Hz mouse (e.g. Razor Deathadder) is still far less than the average movement latency in 125Hz mice.

        • ahigh says:

          Yeah, I have always believed that zero 1000Hz mice on the market use any filtering that would create more than one sample of latency. But I am open to changing my beliefs if the evidence is there. I just don’t really see that it is. If you look at the Wikipedia entry for PID filters, I have implemented these filters many MANY times over the years, and they can absolutely introduce latency — in general with a larger derivative component. It’s just that I don’t see this as being something you would want to put in a mouse .. at all, and I’ve always assumed that any filtering would be more akin to error elimination that trying to smooth out the response curve to the inputs, which is the characteristic of PD or PID filters that generate latency (smoothing the output rather than providing the actual data arriving as quickly as possible — IE unfiltered).

  3. Pingback: Anonymous

  4. Pingback: System-Wide Fix for Gaming Mouse Fluidity in Windows 8.1 | Blur Busters

  5. Fabulist says:

    Incomplete information on this article:

    1) You do not mention at all the side-effects a mouse turned up to 1000Hz may produce, such as jitter due to processor / motherboard incompatibilities (newer and old ones). This has been confirmed by both professional players and companies like SteelSeries.

    2) You do not mention at all the fact that 1000Hz is not always better when we are having extremely variable or capped framerates of 60 and below (500Hz and 125Hz will be more accurate). This has been confirmed by both professional players and ESReality in the past.

    3) You also do not mention that 500Hz is the universally most stable and compatible rate for most mice, and is the rate that has always been preferred from Counter – Strike and Quake 3 professionals for the past decade, that do care and can feel that 1ms difference, but prefer to sacrifice it. Again, this is due to the reasons mentioned before.

    4) And lastly you forget to mention the different effects polling rate jitter can create based on whether the mouse is laser, optical, is running on native DPI / CPI and to what level it has smoothing (for DPI / CPI interpolation – angle snapping effect). Again this is confirmed by professional players and various community sensor experts (e.g. in Overclock.net).

    • Chief Blur Buster says:

      Correct — Most of these were definitely true, but has increasingly becoming less and less true over the last several years, as 1000Hz mice become better and better. When buying the most modern 1000Hz mice with the best sensors, the compromises/problems of 1000Hz is slowly becoming smaller than the improvement that 1000Hz can now give, as well as and more and more professional/competitive players are now resuming using 1000Hz. Not all of them, and not on all systems/mice.

      LightBoost and GSYNC amplifies the difference between 500Hz and 1000Hz, to the point where the tradeoffs are now smaller than the bigger difference between 500Hz vs 1000Hz (amplified by LightBoost, ULMB, GSYNC, or other game motion fluidity improving technology)

      Many competitive players, do not use LightBoost, as that can add input lag, though faster reaction times can outweigh, depending on the use case (e.g. Scout versus Scout in Team Fortress 2, or high-speed low helicoptor flybys in Battlefield games) that force eye-tracking (LightBoost benefits more) versus keeping eyes stationary on crosshairs (LightBoost doesn’t benefit much, and can hurt competitive FPS gameplay).

      However, players that want to enable LightBoost and GSYNC, should be aware that the difference between 500Hz and 1000Hz become amplified by the improved motion clarity/fluidity, to the point where the mouse microstutter differences are more apparent than they used to be a few years ago. Our article is targeted towards “Better Than 60Hz Users”. 500Hz used to be best for 60Hz gameplay, but this isn’t necessarily true anymore with the best/newest displays (when using the best 1000Hz mice) for more current games.

      That said, a lot of your concerns still ring true today for many configurations. It merits a mention, especially with lower quality 1000Hz mice or those with bugs/unwanted sensor effects/negative acceleration/etc. I’ll bring your comments to sharknice’s attention, he can follow up with more comments to your good comments!

  6. Fabulist says:

    Please, allow me to reply:

    “professional/competitive players are now resuming using 1000Hz”

    I know but 2 professional FPS players (Counter – Strike, Quake) that use 1000Hz, and their accuracy is not their pride. If you are talking about SC / DoTA / LoL players, they both do not have the necessity for such a level of accuracy nor have the necessary skills (in terms of sensitivity) to realize the difference, as accuracy in both games is not as significant as it is on a FPS game. Even if they do realize the difference, maybe for their kind of games they prefer to sacrifice accuracy for the 1ms deduction – this could make some sense.

    “Not all of them, and not on all systems/mice.”

    However, it is still not specified which mice or CPUs are the cause of this. Having inconclusive data and no official statements, the fact that they are improving is only speculation.

    Our only hope would be for someone from Steelseries / Razer to clarify it (being gaming-equipment oriented), since nobody will be getting a response about this in this life from MS / Logitech. However, although many people discussed and asked this, no response was ever published, they probably have not determined it themselves yet either.

    “Many competitive players, do not use LightBoost, as that can add input lag, though faster reaction times can outweigh, depending on the use case (e.g. Scout versus Scout in Team Fortress 2, or high-speed low helicoptor flybys in Battlefield games)”

    I am a competitive player and you can be / are a competitive player. Anyone can be and at any game in a sense. The sample should always be professional, awarded and recognized players when discussing matters such as mouse accuracy. There are no designated professionals in BF or TFT as there are in CS, Quake, SC etc. Henceforth, it is asinine to sample TFT ‘competitive players’ to CS and Quake professionals that have been playing for over 15 years and gaming-equipment companies employ them, sponsor them, and base their products on them, for them, and also have gotten their primary feedback from them.

    And I can assure you that you will not find a single serious, respected and awarded professional player that will not take anything instead of input lag or any kind of delay for that matter.

    “500Hz used to be best for 60Hz gameplay, but this isn’t necessarily true anymore with the best/newest displays (when using the best 1000Hz mice) for more current games.”

    This statement is untrue. We used to play with CRT monitors that went up to 120Hz and run CS & Quake on 100 and 120 / 125 FPS respectively and 100 / 120 and 120 Hz respectively years before 500Hz was invented. Unfortunately, due to the hardware limitations and issues discussed earlier, 500Hz is still the king on all possible refresh rates, and all games if accuracy is to be a moderate, and above, concern. Unless, there is valid proof that 1000Hz works as intended under specific hardware or under specific circumstances (there is no such proof until now that is publicly available).

    “That said, a lot of your concerns still ring true today for many configurations. It merits a mention, especially with lower quality 1000Hz mice or those with bugs/unwanted sensor effects/negative acceleration/etc.”

    Sensor quality may or may not be relevant, I have no proof of that nor anyone really knows but companies like Avago, since something that you may consider to be a sensor quality problem it could actually be a firmware issue, cable issue, compatibility issue and so on. However, I can assure you that if you have accuracy issues with one mouse running at 1000Hz on a specific system, you will have a problem with all of them – I have tried most mice and you can confirm this yourself; assuming you have the necessary capacity to feel the difference.

    Negative and positive acceleration similarly to skipping is an issue of laser mice, and it is not a matter of quality, rather a ‘mechanical’ issue – this is how laser mice work. The most high end and newer sensors limit the problem (as users, reviewers, and the companies state) but it is still there and very alive, amidst multiple other issues. If such issues (or others) are presented to an optical mouse it is either a faulty sensor or extremely low quality.

    I see you have a Deathadder in your article’s picture here, I can assure you that from the first Deathadder to the latest one, and with all possible firmwares, they all perform equally inadequately at 1000Hz across multiple platforms. Especially the latest one that suffers from smoothing effects even more than the previous ones.

    • Chief Blur Buster says:

      Agreed, that there are definitely problems with various 1000Hz mice, that definitely affect competitive gamers, especially of the professional leagues (sponsored, paid, etc). However, several of us disagree with this:

      “500Hz used to be best for 60Hz gameplay, but this isn’t necessarily true anymore with the best/newest displays (when using the best 1000Hz mice) for more current games.”

      This statement is untrue. We used to play with CRT monitors that went up to 120Hz and run CS & Quake on 100 and 120 / 125 FPS respectively and 100 / 120 and 120 Hz respectively years before 500Hz was invented. Unfortunately, due to the hardware limitations and issues discussed earlier, 500Hz is still the king on all possible refresh rates, and all games if accuracy is to be a moderate, and above, concern. Unless, there is valid proof that 1000Hz works as intended under specific hardware or under specific circumstances (there is no such proof until now that is publicly available).

      I did say:
      “this isn’t necessarily true anymore with the best/newest displays (when using the best 1000Hz mice) for more current games.”
      I didn’t say “always true” or “never true”. There are also various metrics such as fluidity, accuracy, lag, etc, which are affected differently (and sometimes independently of each other). There are now situations where 1000Hz is definitively better (fluidity-wise), and it’s already published, and already being recommended by vendors. This would contract against your “This statement is untrue”, especially if we also account for the improved fluidity and reduced microstutters (another metric other than accuracy) especially when combined with 120Hz technologies.

      Look again — multiple parties are recommending 1000Hz with the newest display configurations.

      • NVIDIA is officially recommending 1000Hz mice for G-SYNC. They tested and found better GSYNC fluidity with 1000Hz. This is fact. It is mentioned in the GSYNC literature that came with the preview units.
      • Several competitive player sites, including http://www.eliteownage.com, now advocate 1000Hz as having noticeable benefit. This is true for 120Hz users.
      • Back in the CRT days, CPUs werent fast enough to make 500Hz vs 1000Hz stand out. Today, they now are.
      • Back in the CRT days, CRTs weren’t able to do high-def at 120Hz. Microstutters are more visible with sharper graphics.
      • Many competitive players aren’t yet playing 120Hz, 144Hz or GSYNC, where 1000Hz benefits starts to amplify more.
      • Many LightBoost users like me clearly see microstutter differences when window-dragging at 500Hz versus window 1000Hz. This is posted in several forums. (For more information, see the microstutter mathematics in the Comments section of Windows 8.1 mouse fix). 4000 pixels/second dragging/panning/turning speed, 1/1000th (1000Hz) of that is 4 pixels rounding error and 1/500th (500Hz) of that is 8 pixel rounding error (e.g. rounding errors occur when mouse report samples arrives slightly before or slightly after refresh cycle). 4-pixel aliasing creates half as much microstutter amplitude as 8-pixel aliasing. During eye-tracking situations (rather than stare-stationary-at-crosshairs situations), it is definitely visually noticeable nowadays. Also tiny 6-point text is perfectly readable at very fast panning speeds just like in the old CRT days (some models of low-persistence strobed LCD monitors were found to have less motion blur than medium-persistence CRT like Sony FW900), no longer blurring like on yesterday’s LCDs (hiding minor microstutters). LightBoost produces the CRT effect during fast panning/strafing, so motion blur no longer hides small microstutters, and it’s much easier to see.
      • Back in the CRT days, Windows wasn’t nearly as real-time-behaving as it is now. Mousereports have less latency jitter today on today’s CPUs and today’s USB controllers, to the point where if you re-attached your CRT, you will now be more able to see 500Hz vs 1000Hz microstutter differences in the window-dragging use case described above. There’s fluidity benefit, which is amplified with better-than-60Hz technologies. Utilities have been confirming this for many years.

      Agree that 500Hz is better in other cases, especially in the past, however, Blur Busters is about “Better than 60Hz” displays, where 500Hz versus 1000Hz is more amplified than it was during the 60Hz era. Yes, we know there were 120Hz CRT users, but that was back during the day when CPU’s weren’t fast enough to make 500Hz vs 1000Hz stand out, especially at lower resolutions, running under older versions of Windows that had lower-precision multimedia timers than today, and using USB controllers/drivers that had more latency fluctuations than they do on today’s controllers. Today, mouse reports arrive far more accurately (in perspective of timing of subsequent mouse report relative to previous mouse report). Whereas I could never visually see the difference between 500Hz vs 1000Hz before, I do now, because the microstutter weak links have been going down for many years; bringing 500Hz versus 1000Hz well above the noisefloor of timing fluctuations, and amplified by the recent comeback of CRT-motion-clarity displays (e.g. LightBoost) and motion-smoothing displays (e.g. GSYNC) that now makes other sources of microstutters (e.g. 500Hz vs 1000Hz) stand out FAR more than they did a few years ago.

      We are now using high-definition displays (1920×1080 and up) where a screen-width panning in half a second translates to 4000 pixels/second graphics panning motion, where a 4 pixel rounding error versus 8 pixel rounding error is MUCH more noticeable today than it would have been on your 1024×768 CRT yesterday. Lots of competitive players have not re-tested 1000Hz with their brand-new 120Hz monitor on their i7 system running speedy ultralow-jitter USB controllers. 500Hz vs 1000Hz needs retesting when upgrading to 120Hz, since the now-much-more noticeable microstutter improvement starts to outweigh some of the minor problems of switching from 500Hz->1000Hz on several of the best 1000Hz mice. One competitive gamer told me he was surprised how many competitive players don’t bother switching from 500Hz->1000Hz when they turn LightBoost ON (to figure out whether or not the LightBoost input lag is worth it or not), since LightBoost is often be very microstuttery/jittery, and any microstutter reductions help a whole deal. We are actually now at the technological point where the introduction of 2000Hz is worthwhile when such sensor technology becomes possible: It would likely increase the accuracy of the 1000Hz mode in the elite leagues, and improving motion fluidity on strobed displays even further (e.g. reducing microstutters to ~2 pixel amplitude during 4000 pixels/second mouse movement).

      For a long time, the input lag mantra has been quite strong, however, the difference between 500Hz versus 1000Hz poll latency is much smaller than SLI versus non-SLI, CrossFire versus non-CrossFire (before or after frame pacing fix), and even 60Hz versus 120Hz. Some competitive game players still use SLI anyway (And tolerate the +2ms to +3ms SLI-enforced input lag) in order to get extra frame rate to improve human reaction times. When we’re talking about low single-digit milliseconds, sometimes it’s a tradeoff between human reaction time reductions (e.g. higher frame rates, or improved motion clarity) versus tiny increased input lag (e.g. SLI enforced lag, or LightBoost strobing). Input lag is NOT always uber final word — sometimes we are happy to give up about 2-3ms input lag to get almost double frame rate, for example — look at the competitive players that have chosen to use SLI for one reason or another (e.g. cost), for example. They aren’t hypocrites if they have consciously/intentionally chosen the competitive advantage of extra frame rate, in trade for a tiny bit of added input lag. The tolerances of input lag tradeoffs vary (e.g. more casual competitive players tolerating more lag to gain more framerate/other benefits, and professional tolerating far less of tradeoff).

      500Hz is definitely the best for many situations — but also keep in mind a huge number of competitive game players aren’t yet using GSYNC or 120Hz, and many avoid LightBoost due to input lag, and they haven’t yet re-tested 500Hz vs 1000Hz as much as some of us have. I stress the last point — this mouse article is geared towards “Better Than 60Hz” users — and also geared towards both competitive and non-competitive players who is craving the best possible fluidity (e.g. people like me and many of our readers), e.g. Bioshock Infinite 120fps@120Hz strobed solo gameplay. Several of us now visually SEE a more amplified microstutter difference between 500Hz poll rate and 1000Hz poll rate (during eye-tracked screen panning motion situations) whereas it wasn’t visible on yesterday’s 120Hz CRT’s running at low definition on slower CPUs and more jittery USB subsystems.

      That said, it’s worth mentioning of 500Hz being better in many competitive situations (especially when playing in the rareified leagues & on 60Hz displays). I can amend the article to add a paragraph about this, as well as mentioning the existence of problem 1000Hz mice, even though 1000Hz is confirmed (by multiple sources) to be the best motion fluidity on strobing/on GSYNC.

      After all, NVIDIA is now recommending 1000Hz for GSYNC.

      • zldd says:

        Fantastic discussion you’ve got going here. Could you elaborate on a few things I still don’t quite get, please.

        Is it correct to assume that it’s better to use 1000Hz for games with varying frame rates, and 125Hz/500Hz for games with a locked frame rate or when using VSYNC? Also, does it matter whether it’s double or triple-buffered?

        More importatly, how to tell apart these games? Simply googling whether a certain game has a locked frame rate, internally or visually, seems unreliable. Can you give a specific example with two very popular games at the moment – CS:GO and Titanfall – both of which employ some sort of frame locking.

        In your comments you didn’t tell specifically what is a modern system and which are the best mice. For example, does it matter whether you connect the mouse to a USB 3.0 controller, and does it matter if it’s a native Intel controller or 3rd party? Is it okay to connect another peripheral (in addition to the mouse) in the adjacent USB port? Is it safe to assume that all platforms from Sandy Bridge and up are ‘modern’ enough and offer proper USB controllers?

        Thank you for your time.

        • Chief Blur Buster says:

          Is it correct to assume that it’s better to use 1000Hz for games with varying frame rates, and 125Hz/500Hz for games with a locked frame rate or when using VSYNC?

          If you’re using 125fps@125Hz (e.g. Quake Live), you do not have as much aliasing effects between framerate and mouse pollrate. So it’s possible that the improved accuracy of the lower poll rate may outweigh the fluidity-enhancing benefits of the higher mouse Hz rate. It is a personal preference that you will have to test out, and many people can’t tell the difference. That said, some of us (myself) feel the difference when using LightBoost during fast mouse panning movements.

          However, VSYNC ON at 144Hz or 120Hz still creates a motion aliasing effect (microstutter), so 1000Hz still makes a noticeable fluidity improvement over 500Hz during fast mouse panning while tracking eyes on the pan, during framerate capping when the framerate is not an exact multiple of the mouse pollrate. You won’t notice the difference if just staring at the crosshairs stationary, but the 500Hz-vs-1000Hz difference is observed during 144Hz/LightBoost/GSYNC when tracking things panning past your screen (eyes away from crosshairs)

          Also, does it matter whether it’s double or triple-buffered?

          Double buffered is synchronized to refresh rate (e.g. framerate caps out).
          Triple buffered is unsynchronized to refresh rate (e.g. framerate can fluctuate).
          The exact answer likely depends on the current refresh rate (which may determine framerate, if VSYNC ON), and the mouse pollrate if it’s an exact multiple of the refresh rate. Even when the frame rate is a multiple of the mouse poll rate, the multiple is not always perfectly exact, so if you run at a lower poll rate, you may have intermittent latency effects (caused by the frame rate slowly slewing against the nearly-but-not-quite-exactly-matched mouse poll rate). You may not be able to feel them at all, but it may wreck aiming ability. I leave my mouse on 1000Hz at all times, as it never degrades motion smoothness/fluidity (when that’s the goal above all else).

          A bigger effect is the game’s synchronization to the mouse poll rate, rather than synchronization to a frame rate cap. For example, in Quake Live, the frame rate (125fps) is designed for minimum latency against the mouse poll rate (e.g. game engine synchronized to the mouse poll), so latency is designed not to degrade when using a lower poll rate. This is a different situation than, say, for Counter Strike:GO which is not synchronized to the mouse poll the way Quake Live is. So going 500Hz vs 1000Hz can make a bigger motion quality difference in CS:GO than for Quake Live, for instance.

    • Xyriin says:

      “If you are talking about SC / DoTA / LoL players, they both do not have the necessity for such a level of accuracy nor have the necessary skills (in terms of sensitivity) to realize the difference, as accuracy in both games is not as significant as it is on a FPS game.”

      I’m curious you make such an assumption. Obviously you’re a dedicated FPS fan but SC2 in particular requires extreme precision in battles in order to micro individual units and at a highly rapid rate which vastly exceeds even the most hectic FPS. Even DoTA and LoL vastly differ from the default auto attack and click attack mechanics a new player to the game experiences. I think you may be confused due to the appearance of the game if you’re watching a stream. If you watch a CSGO match you get the actual player view as the observer mode cycles between players. For something like SC2 watching the players screen is next to impossible because it’s simply moving too fast. What you see if you observe a match is a 3rd party slowly scanning around and not the frenetic pace of the actual player.

      Similarly a SC2 player may need to make as many precision clicks in a single game as a CSGO player needs to make in an entire match. To discount an entire group of professional players due to limited perspective is disingenuous at best.

  7. SS4 says:

    Very good info and discussions going on here.
    I’d like to add that laser and optical are not the proper them, it should be infrared laser diode or light-emitting diodes . Anyways all are optical since they use optical instead of mechanical means to track the motion:P

    Now infrared/laser/optical becomes confusing as the terms are used all over the place but look for the best sensor available with no such things as positive acceleration and crap like that. Ultra high DPI and the term laser are just used as a marketing ploy and im pretty sure those dual infrared/laser mouse are the same and won’t bring anything really worth it.

    If you want some of the best sensor on the market, try Logitech G400, AM Zowie, Microsoft intellimouse, CM storm spawn, cm storm recon, razer deathadder, razer abyssus to name a few with great sensors.

  8. iksde says:

    I see this article is not very new, so is 2000 Hz mouse tweak, etc.

    However, want to mention something that could have been overlooked. It could be the frametime introducing those stutters, why not give an example how to avoid them?

    F.e. I mostly play Counter-Strike 1.6 and it is for =100 fps (for steamed set fps_max 99.5, non-steamed not sure), setting monitor 100 hz gives fluid frametime, when mouse input gets involved, a good idea is to set it to match frametime for it as well, such as 500 Hz, 1000 Hz are multiples of integers either one or another. Personally I set 500 Hz as software I used for testing mouse showed more faults with 1000 Hz. It is fluid and really nice that way.

    It goes 500 Hz (mouse refresh) / 100 FPS, Hz (monitor refresh) = 5 so each frame has five mouse refreshes processed, with it being divided and equal to integer it is not going to drift and skip at some interval as it is synched.

    This is just a theory, though.

    • Chief Blur Buster says:

      There’s actually 3 frequencies involved, mouse refresh rate, monitor refresh rate & frame rate. There are unsmooth effects even if only one of the three deviates. Mouse poll rate is not generally locked to refresh rate nor framerate. It’s easier to lock frame rate to refresh rate (VSYNC) but that causes lag issues. And mouse sensor reads aren’t at the exact same instant as GPU frame render times. So there can still be dis-aliasing effects. Also, depending on what you’re trying to do 100fps is not always exactly one-fifth of 500Hz. For example, the poll rate may end up 500.2171356Hz in more mathematically exact terms, and frame rate may actually be 99.97136fps. In this situation, you’ll potentially have beat-frequency stutter effects and/or slowly slewing input lag that varies from +0s through +1/500s, as the slightly different numbers goes through the “beats” effect. Fortunately at 500Hz and 1000Hz poll rates, the beat frequency issues aren’t generally a problem at all. It’s more of a problem at non-gaming poll speed of 125Hz, where all kinds of stutters from beat-frequency effects are much more visible to the human eye.

      Either way, the best way to mitigate any stutter effects caused by any frequency mismatch (pollrate, refreshrate, framerate) is even higher poll rates combined with higher frame rates (and/or synchronized to refresh rate, as in GSYNC/FreeSync or the laggy VSYNC-ON). As long as accuracy doesn’t begin degrading — some competition gamers still swear by 500Hz instead of 1000Hz — or 1000Hz instead of 2000Hz presumably as certain mice occasionally becomes becomes slightly less accurate/reliable when pushed to its very limits.

      Mind you, for most, this is mostly simply nitpicking or slicing hairs, but there are indeed situations (the rareified leagues of well-matched paid professional competition gamers) where a millisecond of advantage matters — in terms of the Olympics crossing-the-finish-line-first effect — of being the one who shot the other first — in the “identical-human-reaction-time-but-my-hardware-was-1ms-faster” situation — in a gaming deathmatch to win a hundred-thousand-dollar tournament prize — and similar situations where a single millisecond of advantage can actually matter.

Add Comment