Debate about Multiple Mouse Tester Software

Everything about latency. Tips, testing methods, mouse lag, display lag, game engine lag, network lag, whole input lag chain, VSYNC OFF vs VSYNC ON, and more! Input Lag Articles on Blur Busters.
User avatar
Chief Blur Buster
Site Admin
Posts: 11648
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: USb 3/4+ Future USB 8,000Hz -to- 24,000Hz polling rate

Post by Chief Blur Buster » 25 Oct 2020, 15:30

schizobeyondpills wrote:
25 Oct 2020, 14:51
should've sent me one since i have lowest ram latency in the world and without doubt most stable setup ( 10k euros + in low latency equipment ;) )
If you have a good cirriculum-vitae of past experiences doing this stuff, you can contact Razer_TheFriend on Twitter. He's sending 'em out to select researchers and game developers, in exchange for good hard data. It's a big goal of getting industry ready.
schizobeyondpills wrote:
25 Oct 2020, 14:51
Great, please make it in C++ and post source code online, MouseTester is very bad and outdated using C# on 32bit platform(WoW64)
Good suggestion, however...

I'm doing it in HTML5 using PointerEvents + PointerLock. Thanks to the way Chrome does Stadia, HTML5 in some brands of browsers can now do data as accurately as MouseTester now (1000Hz+rawinput). Far easier and fancier graphs than MouseTester. Mind you, there's bugs being able to run 8000Hz, but it's perfectly fine at 1000Hz. Eventually I'll coordinate Razer+Chrome team to fix the 8000Hz Chromium bugs, as I've long trailblazed the W3C stuff to make their browsers more and more TestUFO compatible; sending lots of BugZillas and CRbugs, that fixed issues.

_______

<Why_HTML5>

BTW, this is a good time for me to go on a (temporary) topic sidetrack to explain why I am doing it in sometimes-imprecise HTML5 and not precision C++. (That said, browsers are getting amazing in some departments thanks to other influences such as Google improving compatibility with 1000Hz gaming mice in Chrome due to Stadia).

You have to realize how Blur Busters work. I'm the one who invented high-Hz-compatible "one click show-and-tell motion demos". I'm doing the same for mouse tests. My goals are different. I'll sacrifice a little bit of C++ precision to have one click convenience to massmarket. 1000-times more popular is more valuable to Blur Busters Prime Directive.

Blur Busters is a hobby-turned-startup with two lines of businesses

Blur Busters is a two-lines-of-business hobby-turned-startup. Blur Busters Media goals (advanced-mainstream microphone dropping factory) is different from the Blur Busters Laboratory (working with manufacturers). I newbie-ify my language in the Blur Busters Media department. I still write too advanced, but it's branching out over the years. Everyone tends to respect that, and sometimes it means I write Popular Science Articles instead of Harvard University Thesis. Other researchers can cherrypick my Hertz Einstein ideas and write their own Thesises or Papers like more than 20+ different groups of researchers wrote peer reviewed papers crediting me or my business (most recently, Samsung researchers in September, about a new motion blur measurement standard).

Buddy, TestUFO has almost a hundred million lifetime visits worldwide from 2012-2020, being the world's most popular one-click motion demo to silence a lot of myths (descendants of "Humans Can't See 30fps Versus 60fps"). I'm going to bring that to mouse testing. And y'know, in my bandwidth traffic, Asia region has an unusual workday traffic pattern on TestUFO. China/Taiwan/HongKong/Korea/Japan has a strong cyclic traffic indicative of many hundreds/thousands companies using my TestUFO as part of their work, a cyclic pattern found nowhere else (no other country). It's so strong that when worldwide bandwidth is averaged, the weekday workday cyclics contributed by Asia still shows up clearly in the worldwide bandwidth average.

Blur Busters is happy working in this manner; we don't give fish; we teach people to fish. That's why Blur Busters does not mass-test monitors (except Special Editions), but dozens of others use Blur Busters inventions to mass-test monitors. I don't compete against the people who use my inventions. I've taught a lot of people to fish. I dislike it when people criticize how Blur Busters functions (sorry) when people demand I mass-test monitors or do something else more researcher/scientist-appropriate. Yes, I long wish to be a YouTuber, but being a deafie -- born deaf but with a brain able to emulate a motion test (like a motion-photogenic logic in my brain, mentally looking at the rainbow artifacts, plasma contouring artifacts, GtG artifacts, persistence artifacts, optical illusions, etc) before writing the motion test -- I find it much easier to blog and show-n-tell via other means. Instead, I rather incubate the path, and any demands & disrespect will just be unceremoniously deleted. If you want to create a C++ MouseTester, be my guest. That is not the Blur Busters mission directive to give away fish. Whether it's ApertureGrille or RTINGS or TFTCentral or LinusTechTips or multiple monitor manufacturers, they're deriving a lot of their tests based off what Blur Busters freely inspired them.

Even many startups have been inspired by Blur Busters (Even practical nausea-free VR arrived a few years earlier than expected thanks to TestUFO, from a paid contract between myself and the Oculus Kickstarter completed long before the Facebook days). Zero Oculus or Facebook stock/options ever. Don't mind. If that means I leave lots of money on the table, nearly everybody respects that, and most businesses approaching me at the beginning knows to respect that too. Ever my lawyer remarked that it's pretty clear that it's something commendable and respectworthy, even though it is metaphorically tantamount to a cultural difference between businesses in different countries. (Yeah, I've already had to defend my way of business function. Fortunately, it's nothing involving any big companies -- they are so experienced and know people like me and work with many people like me already, and they love working with me). Whether it's about a mouse, or a 240Hz strobe backlight, or a quadruple-digit refresh rate prototype, or troubleshooting why their shutter glasses doesn't work well on a specific display, or creating a custom TestUFO for their in-house use, or teaching a class at a manufacturer across the Pacific or Atlantic, or other meritworthy contribution to the refresh rate race to retina refresh rates.

P.S. Somebody else? Sure. Please write a C++ opensource version of MouseTester and put it on github. Perhaps I'll contribute GitHub / BugZilla Issue reports, and maybe a couple of source code commits.

P.P.S. My Special Edition article (at least Part #1) is focussed on mainstream photographic-proof microphone dropping to stop the 8000Hz-laughing audience before next steps -- before more scientific results in Part #2 onwards. The Blur Busters Way is our forumula. Thus, there is little MouseTester-style results in Part #1; just to temper expectations.

</Why_HTML5>
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
schizobeyondpills
Posts: 103
Joined: 06 Jun 2020, 04:00

Debate about Multiple Mouse Tester Software

Post by schizobeyondpills » 25 Oct 2020, 16:21

Chief Blur Buster wrote:
25 Oct 2020, 15:30
>I'm doing it in HTML5 using PointerEvents + PointerLock. Thanks to the way Chrome does Stadia, HTML5 can do data as accurately as MouseTester now (1000Hz+rawinput). Far easier and fancier graphs than MouseTester. Mind you, there's bugs being able to run 8000Hz, but it's perfectly fine at 1000Hz. Eventually I'll coordinate Razer+Chrome team to fix the 8000Hz Chromium bugs, as I've long trailblazed the W3C stuff to make their browsers more and more TestUFO compatible; sending lots of BugZillas and CRbugs, that fixed issues.
but why? only people who are ever going to be concerned with good mouse frequency and input are those who need it for more than casual daily use aka competitive players (its different compared to displays/monitors) and what you test ingame or on OS is as far off as possible from being tested in a browser (that gets slower every version) for being more accurate and precise especially when u take context into account.

dont believe me? check this out https://github.com/krausest/js-framewor ... issues/683
not to mention how mass amount of installed spyware(full site access to read all data) in extensions( multiple js unoptimized scripts)/virtualization/security patches will all cause performance variations

Image

now lets check latest update ( https://github.com/krausest/js-framewor ... -706343090 )
Image
and this is just 1 OS just for benchmarking chrome without anything, imagine the variation when you mention all of the things i included above, not to mention OS and other apps (electron aka chrome) running on the system!!!!!!!


Chief Blur Buster wrote:
25 Oct 2020, 15:30
>You have to realize how Blur Busters work. I'm the one who invented "one click show-and-tell motion demos". I'm doing the same for mouse tests. My goals are different. I'll sacrifice a little bit of C++ precision to have one click convenience to massmarket. 1000-times more popular is more valuable to Blur Busters Prime Directive.
then please do take in consideration (a warning or a disclaimer for your research/tools being mass audience rather than eSports focused) that your research/tools/findings will be used as arguments in environments that you did not consider/create them for and where different results would be measured (ingame vs bloated browser - bloated browser vs dedicated C++ native application optimized in ASM)
Chief Blur Buster wrote:
25 Oct 2020, 15:30
>Buddy, TestUFO has almost a hundred million lifetime visits worldwide from 2012-2020, being the world's most popular one-click motion demo to silence a lot of myths (descendants of "Humans Can't See 30fps Versus 60fps").
thats nice but quantity is nothing without applying quality evaluation on it. 10 billion broken transistors dont make a CPU, quality does.
Chief Blur Buster wrote:
25 Oct 2020, 15:30
>I'm going to bring that to mouse testing.
why? people can use displays without any effort in interacting with them, mouse usage to saturate even 125hz needs to be in a proper environment that needs it, aka eSports, casuals care for display quality, for mice not so much and theres nothing wrong with that, their use case differs. bringing mass market to mouse testing will do more harm than good to those who actually need it.
Blur Busters is a two-lines-of-business hobby-turned-startup. Blur Busters Media goals (advanced-mainstream microphone dropping factory) is different from the Blur Busters Laboratory (working with manufacturers). I newbie-ify my language in the Blur Busters Media department. Everyone tends to respect that, and sometimes it means I write Popular Science Articles instead of Harvard University Thesis. Other researchers can cherrypick my Hertz Einstein ideas and write their own Thesises or Papers like ]more than 20+ different groups of researchers did on Google Scholar. We're happy working this way; we don't give fish; we teach people to fish.


That's why I don't mass-test monitors, but dozens of others use Blur Busters inventions to mass-test monitors. I've taught a lot of people to fish.
"The society that separates its scholars from its warriors will have its thinking done by cowards and its fighting by fools."

you should measure monitors so that consumer-exploiting companies stop and actually create something good (u know without RGB, 5 different ports, 2w speakers, falsified response time claims, locked/hidden firmware settings etc..). you have the necessary reasoning and research knowledge to expose their dirty tricks. i honestly dont know how you can feel great about your achievements and then remember every single monitor display out there is labeled with marketing tricks on response time? (not to mention 10++ other tricks and skew'd claims)
Chief Blur Buster wrote:
25 Oct 2020, 15:30
>If you want to create a C++ MouseTester, be my guest. That is not the Blur Busters mission directive to give away fish.
why are you measuring devices not meant to be used for browsing the internet but for competitive eSports inside a browser then?

on topic of
Thanks to the way Chrome does Stadia, HTML5 can do data as accurately as MouseTester now (1000Hz+rawinput)
1. written in C# (not compiled ahead of time - huge performance loss)

2. written in 32 bit (WoW64 emulation on 64 bit)

Image


3. uses old outdated/unoptimized JIT runtime of .NET which received many "native"/performance oriented improvements in .NET Core
3.0/.NET 5.0 (tiered compilation/Span/Unsafe class/native transition/Ryu JIT optimizations - base class libraries rewrite )

4. C# has a huge penalty calling into native from managed code and especially on 32 bit emulated app running very old .NET Framework version. ( https://medium.com/@jarl.gullberg/an-in ... 08e4da54db )
Image


5. uses converted from C++ tutorial about raw input on msdocs into C# and hence uses a generic raw input that queries things for abstract RAW_INPUT procedure call rather than mouse one.

6. calls into native world multiple times per mouse packet which is very bad doing just once due to things mentioned in 1,2,3,4 (all combined) multiplied by ~ 3 or 4 such calls = jitter/latency/innacurate/imprecise measures.

7. sets the priority of itself to real time which causes it to be above csrss which is responsible for delivery of ALL messages from kernel to user mode apps (ouch) - source : https://superuser.com/questions/555560/ ... -the-mouse
The csrss in csrss.exe stands for client/server subsystem. It is the part that talks to win32k.sys which - after this part got moved into kernel mode (used to be in user mode for NT 3.51) - is responsible (as the name implies) for much of the Win32 subsystem's "GUI" functionality (including window messages).
i wouldnt be very satisfied being as accurate as MouseTester, because its not accurate at all :lol:


User avatar
Chief Blur Buster
Site Admin
Posts: 11648
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: USb 3/4+ Future USB 8,000Hz -to- 24,000Hz polling rate

Post by Chief Blur Buster » 25 Oct 2020, 18:32

schizobeyondpills wrote:
25 Oct 2020, 16:23
[...]
You miss the point. It doesn't matter if you are right. There is merit to both my and your approach. My goals are just different than yours.

Perfect is the enemy of Good Enough. Our goals are different. Browsers are now reasonably high performance animals under the hood, it's not like 1990s where you can't run bitcoin mining & emulators (Internet Archive). The venn diagram of measurable mouse performance is now sufficiently useful to mainstreamize this data. Even C# inaccuracies (e.g. garbage-collect overheads) can be statistically filtered, with outliers rejected, such as steering garbage-collect events outside critical windows, as well as pre-allocation of statistics-recorder arrays, etc. You have seen me write Tearline Jedi in C#.

I Already Do Microseconds in HTML5 and C#, Even Defuzzing Meltdown/Spectre

Check this thread out of me doing microseconds in C#. I program in C# about 10x faster than I can program C++, since it is a Rapid Application Development language to me. A program that exists is always better than no program at all. I can do C++ but I leave most of it to other people.

What is not mainstream today may be mainstream tomorrow. Remember how 120Hz used to be elite stuff like 4K displays? Now 4K is a $299 Walmart special. Even high-Hz will be commoditized. And all smartphones will have 120Hz within a few years (even Apple, even though they bumped 120Hz to iPhone 13). Now even Dell is going to add 120Hz to their generic office monitors approximately year 2025-2029 and most 4K HDTVs have a 120Hz feature built into them now. By 2030, a 60Hz screen will be hard to purchase much like 720p is hard to purchase today. I'm bringing mouse testing to the mainstream and the 10-second-attention-span generation, whether you like it or not. :)

For this thread, replies in in Priority Class 9 of a Priority scale of [1..10]. Please expect my gentle rebuttal reply later this week. Further replies to this thread will further delay my replying, so I suggest leaving this thread fallow for a few days. I would like someone to see do C++, be my guest.

After 2 days, depending on the continued discourse, this thread may be moved to OffTopic Lounge as it is fun/useful but mostly irrelevant nitpicking to me in the near-term since the error margins of GPU-accelerated web browsers is powerful enough to be still comfortably beyond the Vicious Cycle Effect at least on a properly optimized computer and I can still autodetect when the browser is too errorprone (see how I remove the green READY automatically when I detect insufficient browser framepacing precision). Screenshots will have embedded approximate data-quality indications.
schizobeyondpills wrote:
25 Oct 2020, 16:23
[...]
I Already Do Microseconds In HTML5, Despite Meltdown/Spectre Anti-Precision Moves

Please allow me to microphone-drop browser precision:

Image

(Random screenshot I just took now. Some systems do even better than this)

See? 100us framepacing precision in a browser on a 3-year old computer, sufficient for 8000Hz statistics. 1000Hz mice is child's play for HTML5 now via PointerEvents+PointerLock API combo. And I can still self-detect the amount of error/fuzz/etc, and tell the user (how much error pollution is in the data) or invalidate (disappear the green READY).

Know WHY scientists/researchers trust TestUFO? On about roughly 80-90%+ of GPU-accelerated desktop computers (Privacy Disclaimer, already covered in Privacy Policy already publicly disclosed: Google Analytics records TestUFO data. Including computers that does a successful green READY indication -- so I know how many % of my audience is reliably running TestUFO, so I can optimize TestUFO better if certain tests are failing a large percentage of computers) -- my heuristics detects when TestUFO stutters, and immediately invalidate results that have a framedrop. My color coded result invalidator, a famous trademark of TestUFO trustworthiness, has an advanced heuristics engine under the hood.

There are certainly defective browser configurations out there, but no other HTML5 motion test elsewhere on the Internet is as framepace-trustworthy as TestUFO is.

I may be a smalltime developer, but people respect my crazily temporally-accurate HTML5 and C# skillz. (if you haven't clicked yet, see my C# microseconds sheninigians too). I cut my teeth on 6502 assembly language, and I know how to spew near-assembly equivalent of JavaScript at 100 words per minute into Notepad++ by hand for certain TestUFO algorithms (like the VRR emulation algorithm) that performs fast enough that it even works on semi-older iPhones/Android/iPads, as long as it's at least a weakly GPU-accelerated browser. TestUFO is completely handcoded, uses no third party frameworks (except Google Analytics), zero jquery, zero Angular, zero anything. Yes, yes, yes, C++ is nice and higher performing, but I'm all about one-click-awayness these days.

I still haven't yet posted my larger rebuttal yet.

I don't drop microphones. I golf them 400-yards into holes-into-one.

If TestUFO Mouse Tester produces publicity that spawns off a boom of improved C++ mouse testers, then achivement unlocked.

Metaphor Of Goal: Easier Mousetrap, not Better Mousetraps

Others can build better mousetraps than I can do; but I love to invent the easier mousetraps. Puns intended.
schizobeyondpills wrote:
25 Oct 2020, 16:21
but why?
schizobeyondpills wrote:
25 Oct 2020, 16:21
why?
schizobeyondpills wrote:
25 Oct 2020, 16:21
why are you[...]
I will create a bigger reply later, assuming the gentleperson etiquette of keep threads turn-based (AMA etiquette is that no one person can hog all the questionasking to a VIP) -- give me time to focus back on this thread later to do fully replies. Appreciated!

[Resumes other more important tasks]

</MicrophoneGolfed>
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

howiec
Posts: 183
Joined: 17 Jun 2014, 15:36

Re: Debate about Multiple Mouse Tester Software

Post by howiec » 25 Oct 2020, 20:44

The way I see it is that while yes, there are numerous system factors, both HW and SW, that ultimately result in perceived stutter, jitter, and increased input-to-photon latency, any improvement for any significant factor is welcome in this regard.

Like Chief said, the output is only as good as the weakest link although some dramatic improvements in other areas may "compensate" or offset other deficiencies (e.g. 1kHz refresh rate + 0.1ms GtG response times probably make make things visually feel great even while other parts of the system remain at today's inferior standards - yes, inferior per my high expectations).

So 8kHz+ mice are certainly desirable because the benefits are clear and in order to innovate the industry, you have to start somewhere even when the rest aren't caught up.
It's always scary to be first to innovate because there can be compatibility issues and there is a development cost but at the same time you can be 1st to market and bolster your brand.

Hopefully as 8kHz+ mice become more prevalent and ultimately the standard, MS, Intel, and AMD will properly support this area and at least provide options for the performance-oriented customer.

Saying no to improved hardware just because the software (OS, etc.) aren't up to snuff is simply choosing to remain stagnant especially if others have the resources to innovate.

Meowchan
Posts: 40
Joined: 17 Jun 2020, 02:06

Re: Debate about Multiple Mouse Tester Software

Post by Meowchan » 25 Oct 2020, 20:56

Chief is making a website, so having a stand alone tool doesn't help him much. I've seen the call to make a replacement tool on Overclock.net and no such tool has been made yet. If someone wants to be my guest.

I'm not sure having a better MouseTester (C++ / HTML or otherwise) is what the future of mice benchmarking is. Mice testing endgame probably looks more like this:

phpBB [video]

phpBB [video]


And Overlock.net click latency compilation.

Edit: and if you want to observe packet consistency then you can use a logic analyzer https://www.aliexpress.com/item/1005001 ... hweb201603_
phpBB [video]
Last edited by Meowchan on 25 Oct 2020, 21:25, edited 2 times in total.

User avatar
Chief Blur Buster
Site Admin
Posts: 11648
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Debate about Multiple Mouse Tester Software

Post by Chief Blur Buster » 25 Oct 2020, 20:56

howiec wrote:
25 Oct 2020, 20:44
So 8kHz+ mice are certainly desirable because the benefits are clear and in order to innovate the industry, you have to start somewhere even when the rest aren't caught up.
Crossposted from another forum:

The diminishing curve of returns is why Blur Busters isn't a big fan of incrementalism (144Hz->165Hz, or 240Hz->280Hz).

Just like our standard Average User upgrade recommendation is doubling your refresh rate minimum to remain noticeable during gaming (assuming sufficient GPU horsepower). 60Hz->120Hz->240Hz->480Hz->960Hz

60Hz to 144Hz = 2.4x upgrade
144Hz to 360Hz = 2.5x upgrade

Not everyone can tell 144Hz->240Hz immediately (without spending time getting used to it), or 240Hz->360Hz immediately (without spending time getting used to it), but a lot more can clearly see 144Hz->360Hz quicker when versus'd against each other.

For better human-visible upgradefeel to the Average User (i.e. less experienced than OCN.net people), resolutions needs to increase geometrically until retina'd out (240p->480p->1080p->4K->8K), refresh rates needs to increase geometrically until retina'd out, wherever GPU horsepower allows.

Small upgrades aren't human-visible to most, except power users or esports professionals. As soon as a refresh rate becomes cheap/free it becomes the popular mainstream recommendation. Professionals and people with lots of money can leap at the small incrementals, but for the Average User to benefit, needs a big upgrade jump (like the 60Hz iPad to the 120Hz iPad, as well as the future 240Hz tablets). Doubling Hz is needed to noticeably halve scrolling motion blur (e.g. web browsers) on sample-and-hold displays.

...First official 1000Hz mice were approximately 2005-ish back in the ~60Hz LCD dregs.
...First official 120Hz LCD gaming monitor came out in 2009 (ASUS VG236H, Samsung 2233rz)

Back in the mid oughts, that was an 8x pollrate upgrade (125Hz->1000Hz) which was quite human-visible then even at 60Hz with less mouse jittering and better mouseturn smoothness, especially if you were still on a CRT (many were). Even though most saw 500Hz was a sweet spot for 60Hz monitors in less optimized games back then.

Today, we're finally seeing a new 8x pollrate upgrade (1000Hz->8000Hz) which is human visible in the 360Hz+ era. But most will probably initially set it to 2000Hz or 4000Hz at first until games/refresh rates keep going up.

Just like we often settled at 500Hz in the 60Hz days in 2005, we'll probably settle at 2000Hz or 4000Hz at 360Hz in 2020, but start pushing the needle to 8000Hz when 1000fps&1000Hz becomes an esport reality in the 2030s. Keeping mouse pollrate a healthy margin above refreshrate, to prevent mouse jittering.

At small poll-vs-display Hz separations (only 3x difference or less) the jittering goes above the human-visibilty noisefloor. 2000Hz pollrate is already a human-visible improvement for 360Hz monitors already, 8000Hz headroom is a bonus.

Certainly, subject to different user priorities, is whether the "benefit" is is cosmetic pleasing visibility -- or latency benefit -- or other performance benefit -- or all the above -- will be for others to decide, but already one checkbox benefit of worthiness already is immediately apparent to a few of us: Human visible benefit at Windows desktop. Much like if you were there in 2005-2007, got your first 1000Hz mouse, circled the cursor or window drag, and then immediately noticed the smoother cursor at your 60Hz Windows desktop 125Hz to 500Hz/1000Hz.

Gaming mice was stuck at 1000Hz for so freakishly long (15 years) that a pollrate upgrade was already clearly worthy to society when 360Hz monitors finally came out.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

User avatar
kurtextrem
Posts: 41
Joined: 05 Mar 2017, 03:35
Location: Munich, Germany

Re: Debate about Multiple Mouse Tester Software

Post by kurtextrem » 26 Oct 2020, 06:28

What's also important to be said about testing: As long as the system & method stays the same and the variance is minimized, you will always be able to produce good-enough results (as they're relative to each other). The only thing you might or might not be able to do is compare yourself with others (as results are absolute to each other).
But I think this is fine, as this is the job of hardware reviewers: Buy & collect, benchmark, share. Based on that you can inform yourself about how much better or worse your device is compared to others.
Acer XF250Q, R6 competitive player

User avatar
schizobeyondpills
Posts: 103
Joined: 06 Jun 2020, 04:00

Re: Debate about Multiple Mouse Tester Software

Post by schizobeyondpills » 26 Oct 2020, 17:11

kurtextrem wrote:
26 Oct 2020, 06:28
What's also important to be said about testing: As long as the system & method stays the same and the variance is minimized, you will always be able to produce good-enough results (as they're relative to each other). The only thing you might or might not be able to do is compare yourself with others (as results are absolute to each other).
But I think this is fine, as this is the job of hardware reviewers: Buy & collect, benchmark, share. Based on that you can inform yourself about how much better or worse your device is compared to others.
problem is very simple
1. as u scale up frequency to 1000+ the impact of 1st,2nd,3rd,4th,5th decimal (depends on display/mouse rate/cpu hz, they all SCALE TOGETHER COMBINED) is what causes problems, not first decimal or 999Hz vs 1000Hz. so add up all the factors which increase perception/temporal resolution (high OC cpu, high ref rate mon, high freq mouse, optimized OS) to see impact of latency/jitter of 1000.X, or 999.X Hz (or more 999.XXX/1000.XXX etc) .
2. to measure such things you need atleast 100x better resolution, so you need to be able to measure 4th,5th,6th++ decimal with ACCURACY, PRECISION, CONSISTENCY, WITHOUT JITTER.
^ dont forget you also one more level of optimization and baremetal performance due to difference in system load vs ingame load, especially due to duration of that gaming session + temperatures.

3. you cannot do that in a browser because you have to strip away alot of bloat for those digits past the dot, its exponentional. to be able to accurately measure microseconds and beyond you need to work in nanoseconds using raw assembly/C.

theres no problem of having multiple mouse testing tools, its just that there needs to be a clear label (on both sides), one is being made for casual audience that need reviewers and other is made for those past the "dot", for top 0.9% not 99% and below.

otherwise there will be a huge issue of submission to authority whenever those top 0.9% outcasts speak up about something being flawed or not performing as advertised. (and they will because their use case is deeper than average joe which implies more research/more impact of "good enough" engineering penalties, he doesnt care about RGB heating up problems, mouse sleeping, usb controllers; at best he will run web mouse tester and see its up in 980Hz then assume thats normal and move on with his gaming session)

and to make things worse they will be silenced by two replies (as always)
1. "some reviewer X said its ok" (using authority to argue rather than objective arguments)
2. "it works on my machine" (using self as authority rather than observing the context for objective properties)

do i dislike browser mouse tester? no
do i hate casuals ? no
do i think casuals should care about bleeding edge esports? no

i just want a clear label that theres a group of people beyond casuals for whom browser mouse tester results mean only that its plugged in, the real information is gathered from multiple digits past the dot of whatever Hz is on the product box, and for that you need a proper software architected solution in ASM/C. not in browser, not in C# (unless a wizard steps in).

"RESULTS/FINDINGS OF USING THIS TOOL ARE NOT AIMED AT TOP ESPORTS PLAYERS OR OTHER VERY LOW LATENCY SENSITIVE INTERACTIVE ENVIRONMENTS" ( not so hard, like ????????????? )

if i have a 144hz/240hz/360hz display i can perceive far more due to lower temporal latency, now mix it with 5ghz constant clock CPU with all power saving things disabled, debloated OS, locked GPU clocks, OC'd ram.

its the multiple digits past the decimal seperator that matter for high ref rate displays under everything i mentioned above combined (aka eSports optimized PC)

User avatar
Chief Blur Buster
Site Admin
Posts: 11648
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Debate about Multiple Mouse Tester Software

Post by Chief Blur Buster » 26 Oct 2020, 18:04

schizobeyondpills wrote:
26 Oct 2020, 17:11
[...]
Chill, please. I already know to post a disclaimer. In a simpler way though. Already part of the plan.

Also, you may underestimate the useful precision I can pull out of a 1000Hz mouse in a web browser, and how often Chrome team uses TestUFO to improve Chromium. Many mentions of “testufo” in the issues tracker system that were used to fix past issues.
Meowchan wrote:
25 Oct 2020, 20:56
I'm not sure having a better MouseTester (C++ / HTML or otherwise) is what the future of mice benchmarking is. Mice testing endgame probably looks more like this:
Very good comments....but:

You just spoiled the surprise of my upcoming article :D — it also involves a mechanical rig. But I don’t want to reveal more in this thread yet. The TestUFO mouse tester is not part of the 8000Hz tests, but as a separate parallel way popularize one method of mouse tests — a parallel “mouse splash” tie-in, per se, for lesser mouse testing since browsers aren’t 8KHz-reliable yet anyway (under 10%).

Just like TestUFO cannot be used for all display tests either, it’s just some possible tools in the toolbox. I already put a small testing-limitations disclosure and I am also already planning to put disclosures in the upcoming TestUFO mouse tester. But I’ll be attacking the ease angle (zero-additional-equipment angle) too, even if it is but just mere one way to test a mouse. But by popularizing one zero-equipment-zero-download mouse test, interests people into advanced tests. The reviewers who originally tested links now buy equipment to do better tests like GtG heatmapping etc. Free razors, meet blades. It’s also like how my handwave smartphone pursuit camera can lead someone to later purchase a mechanical camera rail.

So many people here miss the point of how Blur Busters operate... That is partially my fault, but over time that is being remedied with a 2021 clearer separation of Blur Busters Media (Mainstream/easy tests/show tells) and Blur Busters Laboratory (Advanced/engineering/equipment/consulting). It already operates that way but they really need their separate websites soon.

Just because there a specialty torx screwdriver/wrench/drill in the toolbox, doesn’t eliminate the need for Simpler tools like a pencil/hammer/ruler in the same toolbox. TestUFO is an easy tool that introduces workshop newbies to more advanced tools.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

Post Reply