Видео

I know, right? I also expected them to be lighter than TBSPD...

I did not! There are so many different combinations of settings that I tend to concentrate on the most popular ones...

I also don't have the same quick reflexes I had as a teenager :(

I take that back. It still drops a frame or 2 causing stutters. Not as much though but still present.

​@@alexandercarlsson Yes, unfortunately that's its only drawback :(

1) Of course you can! I don't use any special tool/software... just Intel PresentMon: github.com/GameTechDev/PresentMon/releases [PresentMon-v2.2.0.msi] 2) In your case, the best combo of settings would _probably_ be: Refresh Rate at 120 Hz + Latent Sync at 75% Input / 25% Display and ~ 40 FPS as Framerate Limit.

Thank you so much :D If 1% Low ≥ 116 FPS, then your combination is literally perfect! If 1% Low < 116 FPS, then I'd round down and set that as the limit. For instance: 1% Low = 97 FPS? Set the RTSS Framerate Limit to 90. In short: G-SYNC (NVCP) + Vsync (NVCP or In-Game: the latter may include other optimisations) + FPS Limit (NVCP or RTSS) set to a value you can always maintain :)

What tool should I use to measure 1% lows? When I use RTSS overlay with a RTSS frame lock, I get never changing 1% lows (0.1% lows still fluctuates) with flat frametime graph. However, with the same RTSS frame lock using other overlays (like Nvidia), 1% lows still fluctuate and frametime graph is never flat.

@@mindinfinity3714 The "flatness" of the frametime graph depends on the combination of "Framerate limiter" & "Frametime calculation point" (as they are called in RTSS): - If you use "Framerate limiter: async / back edge sync" + "Frametime calculation point: frame start"... the graph _will_ be flat. Instead, if you select "Frametime calculation point: frame present"... the graph _won't_ be flat. - If you use "Frametime limiter: front edge sync" + "Frametime calculation point: frame start"... the graph _won't_ be flat. Instead, if you select "Frametime calculation point: frame present"... the graph _will_ be flat. So my guess is that the Nvidia overlay uses "frame present" as its frametime calculation point (but you can't change it, afaik). To record 1% Lows, my suggestion is to simply use the "Benchmark" function in RTSS: easy and you already have it installed :)

@oakenglass I see so it isn't an inaccurate benchmark I am seeing with RTSS overlay versus Nvidia overlay just a different methodology of benchmarking. That being said from the methods you suggested is there any you prefer to use for benchmarking over another? Perhaps, CapFrameX for more accuracy?

@@mindinfinity3714 Exactly :) Besides RTSS, I also use "vanilla" Intel PresentMon. But since CapFrameX is based on the latter... you can definitely use it!

Nope, just use NVCP + RTSS :)

I am... mamma mia!

Low Latency ON is mandatory for Low-Lag Vsync ON / Fast Sync, so I didn't test them without it. As for HAGS, I always keep it disabled as it tends to interfere with my OBS Studio recordings... In short: I don't know, sorry :(

same question

found it u need to click the . button on the letters side of the keyboard between the ? and :

@@mohammedamine4298 yeah i did that but when you set up it is retrieving the previous number

I just press . on my keyboard. Make sure that RTSS is updated to its latest version ;)

@@oakenglass thank you for the reply regardless u need to drop a quick tutorial on rtss decimal points ur comment didn’t helped me

In this video, Low-Lag Fast Sync -0.010 was tested w/ Async [Tier 2]... not w/ Back Edge. The only sync method tested here with Back Edge was Hybrid Scanline Sync [Tier 1] :D

@@oakenglass Thank you very much for the clarifications

You bet!

Thank you so much :D

I'm very glad!

Front Edge Sync trades tearline stability for higher latency... AFAIK, it can't _technically_ match the lower latency of Async & Back Edge Sync D: NB: I am talking about "Low-Lag Vsync ON" scenarios, as shown in this video of mine: ruclips.net/video/_uo2BgakZkI/видео.html

@@oakenglass Thanks I will check it out. Also since I commented yesterday I have been further researching into perfect 30 fps and I found that the latest riva tuner statistics server has built in nvidia reflex frame limiter and I started testing it. There is an insane reduction in input lag even on 30 fps. I tried it at 60 and 120 and those are butter. But I was genuinely shocked at how responsive the game became at 30 fps. It was so responsive that the mouse turns so fast that 30 fps is not able to keep up so I has to reduce the mouse sensitivity 😂. My settings are vsync on in nvidia control panel, gsync on, low latency mode off, riva tuner cap at 30 with nvidia reflex frame limiter and inject nvidia reflex sleep call set to ‘before frame presentation’ for a smoother image. After frame presentation reduces input lag like crazy but doesn’t look as good at 30 fps but it’s amazing on higher fps. If you have the time please let everybody know about it and if it’s a good way to get even more superior experience to consoles or if it has any drawbacks. Cheers 🥂

@@srinivasbharadwaj95 I've already tested RTSS Reflex in that video, but admittedly in "Low-Lag Vsync ON" scenarios and not G-SYNC ones... I'll keep your suggestion in mind ;)

@@srinivasbharadwaj95 Why low latency mode off?

I'm still a bit overwhelmed by my last video on this topic, which was a pita to make... but I'll keep your suggestion in mind, thanks :)

If your monitor doesn't support VRR... yes, Low-Lag Vsync ON is one of the best options.

Frame Generation with just 75 fps? Does it work well?

I believe so!

@@oakenglass thank you for producing these high quality videos!:)

Noted!

I haven't tested this yet... but I will as soon as I can and report back!

RTSS back edge sync gives better experience regarding latency, but fails to hide occasional tearing even if set precisely and i guess amd windows drivers are complete trash as i experience random slow downs and speed ups which gives me headache. windows is terrible indeed, but also it is worse than terrible on amd hardware

I didn't know about HoloISO... thank you for suggesting it! Yes, Back Edge trades lower latency for occasional tearing.

If you have a G-SYNC monitor, take advantage of it... by all means!

@@oakenglass bro i cant understand , nvcp vsync+ingame 142cap gives 18ms. but nvcp vsync+ingame 142cap+gsync gives 10-14ms. no tearing both. Doesnt gsync increase latency?

​@@skkdkdkdkdkd I suggest you watch this video of mine for more info on latency: ruclips.net/video/WSpbYM_TNbY/видео.html

Yes, G-SYNC is OFF!

Fast Sync requires very high FPS to work properly (3X your Refresh Rate).

Try this combination: G-SYNC ON + NVCP Vsync ON + In-Game FPS Limiter set to a value that your system can ALWAYS maintain.

@@oakenglass this game stutter even with gsync they never fixed the vrr support it seems. But I'll try again as you say thank you 👌🏻

True! It's a choice between micro-stuttering (LL Half Vsync) and occasional tearing (SSync X/2). However, if properly calibrated, both synchronisation methods can minimise their side effects.

Yep... I have not deleted any comments!

It's "Low-Lag HALF Vsync ON V2". I've explained at 02:40 its meaning ;)

Because he is Italian.

Because it's a me, Mario! Mamma mia!

What method are you using for 40fps lock?

Well it depends.. If you have gsync, always use that. If not, I prefer special k 1/3 latent sync on games that are well optimized. Latent Sync eliminates the tearing completely, it's that good. But one big disadvantage is its incompatibility with DXVK mod. So low lag vsync is your second best option in that case.

Those extra 10 frames are 33% increase which is quite significant

Hi there! I agree: if you can always maintain 40 FPS on a 120 Hz monitor, you should probably use "Latent 1/3 Sync" (SK) or "Low-Lag 1/3 Vsync ON" (RTSS + NVPI). Apart from the partial compatibility of Special K with DXVK (DX9 games work, DX11 games do not), it should be noted that you can't use it in multi-player games. However, I have to admit that the target of 30 or 40 FPS probably implies that one only wants to play single-player games. Or at least I hope so, for their sake!

​@tvrtkocro28 Well yes. If you can, then go for it. 33% increase but simultaneously big increase in fluidity and latency, there's that too. With PCs, we can squeeze more performance by changing settings unlike consoles.

I cannot help but respond with this famous quote: "you're breathtaking!"

Thank you so very much... and welcome aboard! I've never been to Brazil: I hope to go there one day :D

First UV to reduce power consumption and temperatures, then OC (if there is room).

Scanline Sync X/2 is part of my next video (09/23)! Yes, X/2 sets a frame limit that's half your refresh rate :)

I think it's because they use the same injection method. Read here: wiki.special-k.info/SpecialK/Local I will get back to you if I find a solution ;)

​​@@oakenglassThanks. A bummer that it's not compatible. A big disadvantage for Special K, as stuttering issues for a lot of games can be resolved with Vulkan API.

​@@shayanali8771 Apparently, they can only work simultaneously in DX9 games... Read here: wiki.special-k.info/en/SpecialK/Tidbits ("Special K’s Vulkan Bridge" section).

​@@oakenglass👍

Thank you so very much, and sorry for the late reply! As you can see in the "Combination of Settings" chapter, G-SYNC is indeed + Vsync + Framerate Limiter... because that's the correct way to set it up :) Sadly, I don't know if the "G-SYNC" results on my G-SYNC Compatible display (AW2723DF) would be the same on a "simple" VRR display, as I don't have one. I would be inclined to say yes... but without hard evidence, this is just my speculation. As for the choice between SK and RTSS: since I think (but I'm not sure) that SK "Normal" is equivalent to RTSS "Front Edge Sync", I would choose the latter simply because you can use RTSS in (most) multiplayer games. It should also be noted that the minimal latency difference between SK and RTSS (i.e. 2.37 ms) was obtained with SK "Low-Latency (VRR Optimised)" and RTSS "Async" limiters... not with SK "Normal" and RTSS "Front Edge". Had this been the case, I would have expected an even smaller difference. To sum up, in your case I would use VRR + Vsync + RTSS "Front Edge Sync" limiter :D

Really?! What an honour! The world is definitely scary... but we can face it :)

You're welcome!

Thank you so much! I always strive to do my best :)

I really, really appreciate it! I do still enjoy making these videos, although it's incredibly hard without a steady income... but hopefully I will find a solution soon :)

I didn't know that as I'm not particularly religious D: I've learnt something new today, so thank you very much!

Generally, no, because Fast Sync is designed to work best with games that have a framerate at least 3 times the refresh rate of your display...

​​@@oakenglassso for displays that are not gsync compatible? Can we imply that Gsync + vsync + RTSS cap is the best sync technology that provides the smoothest low latency tear free experience? Are there scenarios where others would work better. You have covered like SSync, Hybrid, low lag, special k etc

@@shayanali8771 I have put together a ranking here, at the end of this video: ruclips.net/video/WSpbYM_TNbY/видео.html According to my data, the best sync method for G-SYNC monitors is G-SYNC + Vsync + Special K "Low-Latency (VRR Optimised)" Framerate Limiter. For non-G-SYNC monitors, the best sync method is Special K "Latent Sync" (with 75% Input & 25% Display). Just remember that you can't use SK in MP games!

​@@oakenglassThanks a lot 🙏

14Hz? Try it again: www.testufo.com/refreshrate

As many of you are interested in the "30 FPS console experience" on PC, I'm going to test different methods in my next video (09/23). To be honest, at the moment, I don't know the best way to achieve this... since I've always aimed for 60 FPS as a bare minimum D:

​@@oakenglassI would say 40fps on 120hz is a more reasonable target. You get way less input lag and it feels much smoother if proper motion blur is there.

​@@shayanali8771 True! However, many people have asked me how to achieve the "30 FPS console experience"... I think it's because they're still playing on a 60Hz monitor D:

exactly! i am looking forward for that video because 30fps on console feels so amazing but on pc... even with rtss i feel the choppiness. i would be so happy for that video

@@erkam2926 Don't worry, it's coming ;)

Don't worry... I'm just experimenting with the algorithm, as it doesn't like my channel very much D:

Just an idea, test doing these videos on more popular games, u may confer the same information and reap the benefits of the algorithm ​@@oakenglass

@@11ELFs I switched from Counter Strike 2 to Hollow Knight because I can't use Special K in multi-player games... If I had to choose a popular SP game, I would have to redo all these tests to compare new results with previous ones D: