Blur Busters Forums

lyrill wrote: ↑

30 Mar 2021, 00:37

yes...i'm aware of that github link and I didn't see any listing more extensive than nvidia website last time, do you have a specific link that lists that or?

There's a listing of 67 mice total on that Git here, if that's what you mean? I think the list is still periodically updated:
https://github.com/NVIDIA/Reflex-Latenc ... Github.csv

lyrill wrote: ↑

30 Mar 2021, 00:37

Today I just got told that nvidia reflex is literally a toy. if you are interested here's a read, you might need translator.

https://www.acfun.cn/a/ac17965268

Boiling down latency testing to a simple, all-in-one user-friendly process is not fully viable, so of course Reflex would be limited in this respect.

ah ok nice, it doesn't have viper8k, as expected. they added so many already might as well let every mice support it already? I mean every mice that is 1khz lol.

what does reportExtensionId perClickLatency polledActuation each mean?

those adjustedAvgLmbLatencyUs stats are surely interesting. DAV2s wired (14ms) and the dark core RGB Pros wired (6ms) and gpx wired (8ms) ranked fastest.

lyrill wrote: ↑

31 Mar 2021, 03:56

what does reportExtensionId perClickLatency polledActuation each mean?

The descriptions for those columns is on the same Git page I linked to you in my first reply:
https://github.com/NVIDIA/Reflex-Latenc ... components

jorimt wrote: ↑

31 Mar 2021, 09:15

lyrill wrote: ↑

31 Mar 2021, 03:56

what does reportExtensionId perClickLatency polledActuation each mean?

The descriptions for those columns is on the same Git page I linked to you in my first reply:
https://github.com/NVIDIA/Reflex-Latenc ... components

oh wow that's cool, seems like the only mouse that has those features is the dav2. beefed to 2x the actual mouse poll, the optical switch detection laser pulses 2khz. I'll bet it's the same on their best all purpose mouse BAV2 ...and maybe per the same logic it's even 16k pulse on the Viper 8K? It would be lame that the whole package polls 8k total signal transmission but only the sensor exceeds that by 3x+ but the switches are 1/4....The wire being denser is surely not just extra shielding and extra transfer duty, but extra electricity drain requirements too?

Hello,

For years, I have configured my settings on my FPS games according to the recommendations of experts:

G-Sync On + Nvidia V-Sync On + Limit my FPS by -5 FPS relative to my screen (in my case I limit to 138 FPS).

Now, Nvidia has arrived on Apex Legends and Overwatch so I'm reviewing the mining I need to configure but I'm reading different opinions on the internet (I saw the Battle(non)senses video in particular but I was even more lost.).

As I understand it, enabling G-Sync is a good thing with Reflex Enabled + Boost. Most of what I've read recommends to keep using Nvidia V-Sync on "ON" (not Fast)

But there are three things I can't decide:

1. Why do you still need to limit your FPS if Reflex does it automatically?

2. For games like Overwatch, Apex or Battlefield, is it better to cap your FPS at 138 (I'm using a 144Hz monitor) or is it better to not cap it?

3. In the Battle(non)sense video, he says that Nvidia Reflex is useless when doing G-Sync +V-Sync Nvidia + limit FPS. I didn't understand why. Don't you have to activate it in this case of configuration?

In conclusion:

In the case of competitive FPS, is it better to disable G-Sync and only use Reflex Boost without limiting my FPS?

If I want to use G-Sync + V-Sync in addition to Reflex (in case Reflex remains useful), should I limit my FPS to 138 or not cap it?

(I always limit my FPS directly in the game when possible. If there is no limiter like on Apex or Battlefield, I use the Nvidia one).

Sorry for my questions that may seem silly. I admit that I am not an expert and that I am lost when I read contradictory information.

Thank you.

so they added the

"Corsair CHAMPION SERIES SABRE RGB PRO 1b1c 1b79 0109 wired none FALSE TRUE FALSE 0 0 1000 573 285 573 NVIDIA"

looks more and more (or less, is there a term for this situation? no I'm making it one) like Nvidia heavily shilling with peripheral brands honestly.

well at least there is no lie that any of these some how runs 8000hz with reflex.

But I guess the takeaway is still that any reviewer that shills " oh wow lowest latency top ranks all corsair so good corsair best brand best kb best mouse get super good price" well, you know what the deal is.

Also what's the point of using the term RGB in naming, like, it's not even funny for meme effect---like, can you ACTUALLY believe the RGB version that simply adds nothing else but RGB and cost 5 bucks more AND heavier, somehow is faster and actually the fastest reflex mouse (that would be, for the last time, all limited to running 1khz A WHOLE HALF YEAR PAST 8KHZ ANNOUNCEMENT), when their OWN new tkl 8khz kb has a tournament switch that CLOSES all led for max perf......facepalm...

edit: static red, sure, so tell me why did you think Cherry was being faulty with going standard on mx 6.0 again? Corsair? Remind me when Corsair actually makes a standard keyboard of all things. No hate on mmo gamers btw (does any still rely on extra macro keys on kb? they seem extinct whether one brand makes them or not) And why always cheaping out on OPX versions in the releases? not even sending them to all the generic reviewers you paid for? (honestly this industry is dead and stale EXACTLY because EVERYWHERE all the money spent in the wrong places, paying for these people who don't even know or care 8khz isn't gonna be wireless supported and keep asking for wireless, like why do you even review 8khz let alone feel entitled to be given a copy YOU DON'T EVEN WANT OR APPRECIATE let alone feel justified to have an arbitrary BLOATED AND WORTHLESS counts of followers that know and care even less) It's not even like 1mm actuation is anywhere short and fast enough going by 2019 industry standard set by apex pro, and this is just pairing for 1khz. pairing 1mm with 8khz in 2021 vs pairing 0.4mm with 1khz in 2019 is like pairing my dad with Kobe vs pairing Lebron with a retired Jordan. That's how it is..

Oh and what's the point of 4khz scan but 8khz poll? it's like less meaningful than having 200fps and 400hz monitor.

I used to run G-Sync On, V-Sync On, FPS set to -3 below refresh rate, and Low Latency set to 'On' in NVCP.

But now Reflex automatically sets it further below the refresh rate. According to Reflex Latency Analyzer, it is slightly quicker in input lag than the old method (because Reflex), but I did like the increased FPS with the old method which meant smoother movement of the mouse.

Is there any way to change that setting through any program? To get it to limit higher, closer to refresh rate?

Also, what should average system latency and pc+display latency be on a modern mouse?

7Z7 wrote: ↑

07 Apr 2021, 06:13

1. Why do you still need to limit your FPS if Reflex does it automatically?

You don't "need" to, but if you want to use G-SYNC + V-SYNC + Reflex + an in-game or external FPS limit, you have to limit a few frames below the auto Reflex limit to override it.

7Z7 wrote: ↑

07 Apr 2021, 06:13

2. For games like Overwatch, Apex or Battlefield, is it better to cap your FPS at 138 (I'm using a 144Hz monitor) or is it better to not cap it?

You mean G-SYNC + V-SYNC + 138 FPS vs. G-SYNC off + V-SYNC off + uncapped? If so, it's entirely subjective...

There's no "better" here in every respect; G-SYNC off uncapped can reduce average input lag over properly configured G-SYNC by a couple more ms (the reduction/advantage of which decreases the higher the max refresh rate) due to tearing, assuming 1) your framerate is at least 2x above your refresh rate, and 2) said framerate doesn't max out your GPU-usage (or you have Reflex on with uncapped G-SYNC/V-SYNC off).

However, on the flipside, G-SYNC + V-SYNC, within its range, eliminates tearing, and syncs the framerate to the refresh rate, making frame delivery smoother and more consistent at the cost of tearing removal, which prevents the framerate/frametime from exceeding the refresh rate.

7Z7 wrote: ↑

07 Apr 2021, 06:13

3. In the Battle(non)sense video, he says that Nvidia Reflex is useless when doing G-Sync +V-Sync Nvidia + limit FPS. I didn't understand why. Don't you have to activate it in this case of configuration?

I don't specifically know what comment of his you're referring to, but I'd assume he means if you can sustain your FPS at your set limit with G-SYNC + V-SYNC, and your GPU isn't maxed at said framerate, Reflex can do nothing further to reduce render queue input lag unless your framerate drops below your limit (in which case it can be left on as a fallback).

7Z7 wrote: ↑

07 Apr 2021, 06:13

In the case of competitive FPS, is it better to disable G-Sync and only use Reflex Boost without limiting my FPS?

Absolute lowest input lag will always be G-SYNC off + V-SYNC off + uncapped + GPU not maxed, but tearing/micro-stutter is the trade-off. Again, it's subjective; There's no one "right" answer between the two.

7Z7 wrote: ↑

07 Apr 2021, 06:13

If I want to use G-Sync + V-Sync in addition to Reflex (in case Reflex remains useful), should I limit my FPS to 138 or not cap it?

Again, depends on if your FPS ever drops below your limit; if it does due to your GPU maxing out, use Reflex as a fallback, if it doesn't, Reflex won't do anything further to reduce input lag.

drmcninja wrote: ↑

20 Apr 2021, 18:08

I used to run G-Sync On, V-Sync On, FPS set to -3 below refresh rate, and Low Latency set to 'On' in NVCP.

But now Reflex automatically sets it further below the refresh rate. According to Reflex Latency Analyzer, it is slightly quicker in input lag than the old method (because Reflex), but I did like the increased FPS with the old method which meant smoother movement of the mouse.

Is there any way to change that setting through any program? To get it to limit higher, closer to refresh rate?

No.

drmcninja wrote: ↑

20 Apr 2021, 18:08

Also, what should average system latency and pc+display latency be on a modern mouse?

Completely depends on the given mouse + monitor + system + game + testing method combo. There is no set number, hence the need for testing.