Further CS2 Input Latency Testing
ฝัง
- เผยแพร่เมื่อ 3 มิ.ย. 2024
- This video isn't sponsored or in any way associated with Nvidia, I just felt like some of the comments the first video received needed following up on using FACTS and LOGIC and Fluffykins.
See the first video on input latency here: • CS2's Input Latency
0:00 - Latency misconceptions
1:31 - CS2 VS CS:GO / CS:S Latency
2:23 - Vsync
3:43 - Gsync + Vsync + Framerate cap
5:01 - Upscaling adds latency?
5:28 - Power consumption VS Input Latency
8:10 - Nyquist's Theorem VS Slow-mo capture
8:41 - ULMB2
9:26 - Microwave VS wireless mouse - เกม
More Fluffykins in your vids please. She's a true gamer.
Cats seem very interested in CS. Especially during comp matches.
Watching this makes me realise why there were so few 3kliks videos about the cs2 beta and early release. So much work in testing a million different variables just for an update to release in the middle of testing meaning you have to start all over again. And then when all the testing is done it would only be relevant for a couple of days until it's outdated and worthless again. Makes total sense to wait until things started to settle down before doing testing like this
I think I recall Philip saying this exact thing before (probably on one of his other channels though)
@@3kliksphilipbrother, I love ya for what it counts
@@3kliksphilip sadly, with how competitive the scene has gotten for counter strike nowadays, a lot more people get way too emotional about the game. They'd rather lie to themselves and remain "right" than admit the new engine is good. Only thing I even see worth complaining about doesn't even have to do with how the game runs, it's with Valve not adding more game modes etc. I sometimes miss the days when the majority of the community didn't care so much about being the best and just wanted to have fun.
@@3kliksphilip love your stuff man, keep up the great work! Plus I love your cat
@@3kliksphilipYou have never done anything wrong with your channels, it was me, I take the blame.
The amount of hate fiction surrounding CS2 is astounding.
Ikr? Its not even the right reasons like, green checkbox not appering when accepting a match(leading to 30+ minutes of pure retries to find a match and being able to accept it) and the extremely random crashes that some specific uses( like me) are experiencing, this being related to something of either sounds or effects as this only happen when somebody dies(via headshot or grenade, also this didnt happened on the beta and is not being noted or adressed in any of the millions of updates the game is receiving[seems like a small geometrical error on an invicible mesh is more important than people being allowed to rpoperly play at all]). Lads be screaming at 0.0001ms input delay while other cant play
@@zaikolebolsh5724 I agree, it's pretty good and it also wasn't release prematurely, it has a lot of niche problems though compared to GO
My game crashed once or twice and I remember one of them, it was exactly when I jumped.
Once, scoping in with the scout lagged, like it showed the scope, didn't show it and then show it again in quick succession, and then I died because I was just startled
3 or so other times I had to restart the game because the game kept kinda switching from 4:3 to 4:3 stretched every second or so, and I saw mostly my desktop which was weird.
Also I often get stuck on random walls or ledges, I don't know why but I suspect that the maps now have way less clipping?
Also the joystick that was plugged in once in a while messed up my controls and randomly made me look up or down or spin uncontrollably but i couldn't replicate it. I didn't expect the joystick to be the problem because in CS:GO you had to enable a setting and execute a config to use a controller
Valve "fans" and competitive "gamers" often act like when they hate something there HAS to be a legit reason... when the reason is they hate change... hatred doesn't need a logic or a reason just a delusional person with a loud mouth and people who WANT to believe what they said do without question. Not just in gaming but in every aspect of life.
The amount of love for a minimum viable product is incredible. Welcome to the twilight zone. We’re going backwards.
@@3ofSpadesyou can't call cs2 minimum viable when MW3 exists 😂😂😂
Have you tried testing if your ouija board will effect your mouse latency? My house is full of them and I think the seances are holding me down in competitive.
i died laughing from this
Lmao
affect*
I think it has slightly higher latency, but the info on my opponents location from the ghosts outweighs the latency difference.
@@ieatthighsöffect"
I do find it so fascinating how fast computers work.
same
I suggest looking a bit more in depth about how they work, so you can be even more fascinated. It's completely insane how fast calculations are done nowadays.
What’s even more fascinating to me is how fast they work, and yet how displeased we are when they still aren’t working “fast enough”.
Oh how spoiled we have become with new technology lol
yep that's what electricity going through some rock does for you apparently
who asked you? you find computers fascinating? you need bedded
Having tearing appear only near the bottom of the monitor with G-Sync on and VSync off makes sense, as the fps limiter is *mostly* doing its job, and just sometimes letting frames get rendered slightly too early, i.e. just before the current frame scanout finishes. And the scanout always goes from top to bottom.
Also, the Nyquist-Shannon sampling theorem doesn't apply here because the signal you are sampling, i.e. the pixel response of the monitor, is not a band-limited signal. For example, if you imagine a perfect monitor with instant pixel response times, and the image just changes between black and white on every frame, the result is a square wave, which is well-known to have infinite bandwidth. Intuitively: If you capture one sample where the pixel is perfectly black, and on the next sample it's perfectly white, there is no way for you to know at which exact instant between these two samples the change happened. And even if the theorem did apply, the result simply means that you can reconstruct the signal perfectly via interpolation, which is not being done here.
You only have 35 likes cuz u be lying out yo ass. U dont know shit. Sir bluffington pussy
I've been playing CS2 since the big beta release just before it went live. I was playing when the shooting felt off, when the movement felt wrong, when the play models were broken, I even continued playing when people were randomly being vac banned, and I'm still playing it now. I don't think I'm that bad at the game either at rank 16k.
I can happily say that cs2 has reached a point in which input latency is either not noticeable, or non existent and if people are still complaining about it are just doing so to find any reason as to why they couldn't hit a shot.
The game input latency was never an issue in the first place.. It's always
- Frametime pacing
- Poor lag compensation (Absurd Peekers advantage )
These major issues are still needs to get much better
Yeah everything apart from some movement feels perfect for me
input latency is always gonna be noticable for a specific, smaller set of players. Theres evidence of players doing double blind trials being able to detect 7(!?)ms of input delay difference with a high statistical significance (i.e passes hypothesis testing).
Why people can't hit a shot is probably because of the first shot bloom and the massive deagle nerf.
@@Vinicantstopcrying The peeker's advantage is super weird in CS2.
Basically the server is trying to do some advanced ping compensation for all players so that it's "fair".
Games nowdays do the hitreg confirmation on server side, so basically whoever has their shot registered by the server first wins the gunfight, which gives an advantage to people with lower pings.
To compensate CS2 tries to wait for the response of the other player who might be playing with a higher ping. So if you have 10ms ping and a guy you are shooting at has 150 ping, trying to kill him will feel like you are playing with 150 ping yourself.
The high peeker's advantage is probably a result of this system being poorly implemented.
The irony is that the 128 tick solution that existed in CS:GO before was much better than all this BS combined, and not only would it have been cheaper to implement, it probably wouldn't cost much more to run, because CS2's packet's are way larger than CSGO's so by the time you add everything up the internet bandbwith usage is probably not much different (And internet bandwith is what's really expensive. Not the servers)
something interesting is that in the rhythm game osu!, they have a newer version in development which has faster latency, and because of that people actually have issues playing it, because it feels off. it's too quick, too responsive, not what people are used to, and many people fall back on the issue of blaming it on extra latency, when it's actually the opposite. perhaps something similar is going on here
i can vouch for that, i hopped off lazer because it felt too responsive and im playing older maps anyways, like image material. mania feels REALLY good in lazer though.
I think for a new player especially it's a no brainer that Lazer feels better than stable but it feels off when you're so attuned to the game
your videos are so technical but I always learn something with them, I have maxxed out the performance of several PCs I've had over the years just watching your content... + the background music is so good I even bought your music kit for CS, it's soooo calming. Love from Chile❤ don't ever stop making videos
Thank you for the work you do in researching and educating the fanbase. Even when parts of that fanbase are exasperating. Been watching your videos for years, even when taking a break from the game, and you've always been one of the best resources for information about CS.
^
Microwaves have decent enough isolation (seen this mesh in the window and wondered what it's for?) to not affect most cases, unless you're trying to study stars and get random sygnals out of nowhere and it turns out that someone was turning microwave on in the kitchen on the other side of the building at the same time sygnals were detected (true story)
Or listening to music on bluetooth
My microwave sends some heavy static to my headphones and it’w funn every time
or trying to create a resonance cascade.
that's more a electric insulation issue, probably a ground loop somewhere creating static discharges. Any appliance sold in USA/UK/EU should not produce interference if they passed certifications
Джо?! Внезапно... А так-то у меня микроволновка замедляет вайфай почти до 0, хотя сигнал полный показывает. Где-то читал что это из-за слабого передатчика в устройствах, т.е. они получают вайфай сигнал, но не могут ничего передать
They explicitly state that the microwave didn't cause interference while running *only* when employees opened it before the timer ran out, which happened often.
Optics/Electrical Engineering grad student here, glad you got to hear more about the Shannon-Nyquist theorem! Just a small nit-pick on your verbiage, the sampling theorem is still accurate, but because pixels don't update at the frequency of your refresh rate, the fundamental frequency you are trying to sample is larger. Hence, to obtain un-aliased measurements, you must sample at a frequency twice as fast as the frequency of the pixel updates to be able to fully measure how the monitor is displaying information.
This is the essence of what you said, but the theorem doesn't have an issue, it just that the frequency components of the signal from the display updating exceed the monitor refresh rate because the monitor doesn't update globally.
Sorry for being pedantic, just thought I would put a bit more precise language in a comment. I always love what you publish and admire the care you take to make accurate and informative content. No joke, your content has really inspired me with my academics because of how relevant the content I study is to the topics you discuss. It is very rewarding to see you apply concepts I see in textbooks, and you do it so well when (I assume) you are self teaching. You and your channel really are a gem Phillip. I can't thank you enough for the effort you take to analyze the technical aspects of gaming and how you clearly communicate this information to an nontechnical audience. Since your earliest mapping videos you have given so much to this community. Cheers!
I don’t really understand how Nyquist-Shannon applies when re-sampling already time-discrete values. Ignoring the (interesting but more complex) actual refresh behaviour, and assuming constant frame times, why wouldn’t a camera frame rate of exactly the monitor refresh rate (or anything above) suffice? Above that, there can never be two display refreshes happening in between two camera captures, so no information would be lost.
I've really found this video entertaining and informative. Well explained and I really like the data you provided, presentation is enjoyable.
Thank you 🙋🏻♂️
im one of the five people that would appreciate an in depth video about vsync
also, i find your research about latency and efficiency interesting. i'm about to grab a 4090 and 7800x3d myself so i might want to dabble in this stuff too.
4:38 battlenonsense is correct here, the tearing at the bottom happens because framelimiters don't always delay the next frame long enough to meet or exceed your monitors refresh duration. even fps limited you will still get some frames coming in more quickly than your monitor can refresh a single frame, leading to the shown tearing at the bottom. vsync is needed to ensure it doesn't happen, which would lead to that single frame that was produced more quickly being shown a tad bit later, effectively waiting for the monitor to finish refreshing the previous frame. this also causes some small latency penalty when it happens
I love the fact that your cat always looks mad.
Your cat looks like a sad old man with mumps in his eyebrows. I love it
I love and really appreciate your long videos
Great video as always, I just love the way you present things & talk.
The microwave thing is quite interesting, I could see how a microwave could interfere with a mouse that uses wifi/other signals that are 2.4GHz. But I think with modern enough microwaves that prevent leakage, along with newer wireless technologies have mitigated this issue.
@@3kliksphilip my 2.4ghz wifi almost dies near the microwave but still shows full signal. I guess my phone can receive data but can't transmit anything in response
You can test to make sure if a microwave is properly sealed, unplug it, disable wifi on your phone, place it in the microwave, call that phone. If the phone rings, you have a leak, if the phone does not ring, it is perfectly* sealed.
Edit: I suspect microwaves with improper seals would interfere with some connection methods
@@TruCloudGaming cell service has a different frequency and will get through most microwaves, i'll test it with speedtest on wifi
@@TruCloudGaming results: cell showed 1 bar, wifi disconnected, bt headphones only worked if i put them very close to the door
@@0ka354 You're right, I got that backwards, I am running on very little sleep right now
I love technical breakdowns like this but the cat really pushes it into holy knowledge territory.
I'd love to see a video on the fast vsync topic! it's very unexplored and having someone like you properly test it would be great
lol'd at "electrical infetterence" nice reference
I don't know how but your cat somehow looks both really happy and really sad at the same time. Still, absolutely adorable.
Oh, and good video too. Very informative.
Your cat is adorable 🥰 that ending shot especially lol Great vid!
I really like your cat Philip! And microwave bit was hilarious.
I gotta say I love your mad scientist approach to figuring shit out.
I waited the entire video for the bit with the microwave and you did not disappoint
Your cat is an absolute angel, I always love seeing them pop up.
Thank you for featuring Fluffykins, it is much appreciated. I have not had a cat in a while so I enjoy seeing other's cats. Genuinely fills me with so much joy.
Also, yeah, I was wondering if that box was tied to the game or just your system. Because, since it is just a box, it seems disconnected to everything else.
So happy to see Fluffykins in another video 😊
You won't believe it, I accidentally enabled that option last night, and thought it was a visual bug on CS, and now you've uploaded a video and explain what that box is/was HAHAHAH that's actually crazy.
Afaik the gsync+nvcp vsync+fps cap thing originates from the blurbusters. They reccommend it as their optimal settings and in their faq they explain why.
Gsync does require vsync to eliminate all causes of tearing. Vsync does something blurbusters dubbed frametime compensation. You can try slowmo recording your screen in CS2 while using +forward;+turnright and you'll probably still observe tearing in the lower third of your screen with gsync on while is vsync is off. That is also true for other variants of VRR.
I can't believe Flip microwaved his cat
current title: CS2 - Microwave Gaming? AND MORE!
Why do you think the title will change?
@@bluesides8323 A lot of youtubers change thumbnails and titles to see how it impacts 1st day analytics, it's quite common practice. Some even do it over a week period, as I have seen on the channel Veritasium.
@@bluesides8323 because it changes most of the time
i fucking lost it when the microwave was suddenly chilling on your desk xdd
Please include Fluffykins even MORE in your next videos. I freaking love cats.
the latency testing with the microwave being next to the monitor was such a 3kliksphilip thing to do haha love it mate
The amount of effort put in testing this kind of stuff is simply astounding! Thank you!
I think I caught a glimpse of a possible dell inspiron 7577 laptop in one of your videos, could it be possible if that is so, to ask you just how cs2 runs on it? Mine is pretty bad, on 720p with low everything I get mostly sub 60, around 47 or less fps and am wondering If maybe it is just my machine thermal throttling or something, since a 4 core 8 threads and a 6gb 1060 should I believe still offer a bit more performance in cs2 even though it is a 7 generation intel paired with a laptop 1060 maxq. With 32 of ram also. Maybe some else has the model and they see my comment and can reply with their experience. Anyway, thank you very much for the amazing content and for the lengths you go through to keep the community informed!
I believe it is the cat that negate the microwave interference with the mouse 😆😆
Love the content as always. Since it seems like you have a decent way of testing latency now, would you be interested in making a video on latency optimizations? Testing which work and which don't, etc.. Especially the ones in that one YT video titled "THE ONLY WINDOWS PC OPTIMIZATION GUIDE YOU WILL EVER NEED IN 2023".
Fun test with the microwave, however it depends on microwave model, and the wireless connection method, I used to have one that when turned off would slice my wifi signal, noticeable when watching videos and it suddenly started buffering when the microwave was on.
It was 100% replicatable every time, but only ever with that specific microwave, the one I owned before and after never had any issues. Could be down to poor shielding, frequencies from the induction, I dunno. All I know is that the commenter pointing out it may be an issue can be correct, but probably most the time you'll be fine.
Came for microwave, stayed for cat
i love the amount of Fluffykins in this video
"Infetterance" what a brilliant reference. Loved that!
Love the ending. 😂 Good stuff mr 3kliks!
I see Fluffykins -> I click!
Literally fell in love with this old cute cat. such a cute boy (:
Another great vid thanks philip :)
Thank you for testing the microwave latency, this is what we can to see!
your videos are so relaxing. i like to get stoned and listen to you talk about source engine. you and your cat are very cool, and i like cats. good video your delivery is so chill and when you go blah blah blah and leave mistakes in it makes it so human and you are a cool dude phillip. your content is nice to come back to because you are humbly just talkin about meaningless shit
lol, the microwave banter was great. I have had a hard time finding microwaves that interfere with any wireless communications in a hot minute, but I have had microwaves that fizzled my wifi long ago. Can't imagine there are any more produced too much nowadays since the majority of the world is online in some way.
Thank you for answering my question!
very nice and constructive video Philip :3
Cat in the end is gold 😂
THANK YOU. I have been struggling to play in cs2 and when I was watching this video I saw that you had less latency when you had Nvidia low latency mode off so I turned it off. Now I can finally play like before and Bunnyhop again. And I also noticed that I had less latency when I turned on the G-sync on my monitor and disabled it in Nvidia.
AMD 5 2600
GTX 1060 6gb
I was one of the people who wrote that I use Gsync + Vsync + Frame cap (alternatively nvidia reflex). From my understanding Gsync by itself does not prevent tearing, it only syncs the refresh rate to the framerate. Tearing can still occur because of individual frametime variances. Vsync should introduce little to no latency when gsync is used and the framerate is below the monitor's refresh rate. I assume you've read the blurbusters gsync article, but if not, I recommend giving it a look. But I do have to say this whole debate about the optimal gsync setup is very confusing and seems that everyone says something different. That's why people often refer to the blurbusters article, because they are quite a trustworthy authority on this, at least compared to anecdotal evidence and random Reddit comments.
this videois subtitles are quite funkyÖ like i am fascinated by how they happened
Thank you for making videos.
Enjoyed the music in this one
Hahaha! Microwave testing! Need more of it! XD
Though the only case I can relate to that is when I turn on my electric heater, it actively reduces my phone wi-fi signal
Microwave testing answers questions we all had
great video once again
Cant wait for Macrowave gaming
4:36 - As far as I'm aware, that bottom of the screen tear it's there with G-Sync On and V-Sync Off (NV CPL), while with both G-Sync and V-Sync On, that bottom of the screen tear disappears, but adds to the latency.
this is the background music I prefer!
i was expecting this video to be about cs2 playable through an LED microwave
I own the same microwave! I'm glad to see it's gaming certified!
I love your cat, they're so adorable.
Classic 3kliksphilip move wearing a bath robe at all times.
8:13 there is other theorem that states sampling any analog signal should be done 10x the frequency of that signal, so other way around should work too.. heck its even used in every sound card/setting (max hz that ppl hear is around 19-20khz, so thats why there is setting lets say in windows, that lets you select from 44k to 192k Hz sample rate, with higher the better quality)
I’ve heard that Reflex Boost runs your GPU at max clocks all the time to minimize latency, which would explain the higher power consumption.
but the power consumption literally went down? 8:00
well you heard wrong
@@MaksKCS It substantially increases it in the 360 fps cap scenario, from 92W to 124W
@@PPedroFernandes “Boost Mode increases GPU clock speeds in CPU-bound scenarios, allowing the GPU to submit frames to the display faster, thereby improving Reflex’s effectiveness.” -NVIDIA
@@jeffreymalone2079 yhea. Boost mode. But boost mode is not reflex.
And boost mode most of time does shit all. Normal reflex is enough
Wireless mice can frequently be interrupted if you have a router close to the dongle that has a powerful wifi signal. I had that issue with my superlight because I plugged the usb receiver directly into my pc, which is close to a router. I was getting tracking that was skipping and getting interrupted. Plugging the usb receiver in to the lightspeed dongle and putting the dongle further away from the router fixed it.
It's not even an unusual scenario. In some apartment buildings for example there will be dozens of routers of varying strengths all blasting into your apartment. You will often notice wifi being so unreliable in that scenario you are forced to use a network cable to get playable gameplay. That same interference will potentially lead to your wireless mouse having a similar issue as the completely verifiable one I mentioned.
i love the arc caboosing is on, literal gaslighting
If you're cpu limited you could try disabling hyperthreading, making each thread faster, though halving your thread count.
The Nyquist theorem is the minimum needed to reconstruct a sine wave. For anything else like e.g. a square wave you need a lot more as the frequency contents are higher.
The theorem states that if a signal is sampled at a rate that is at least twice the highest frequency in the signal, then the original signal can be perfectly reconstructed from the samples. This is true regardless of the shape of the signal, as long as it is band-limited. The theorem is often explained using the example of a sine wave, because it is a simple and intuitive way to illustrate the concept of sampling and reconstruction. However, the theorem is more general and does not depend on the signal being sinusoidal. Any arbitrary waveform can be decomposed into component sine waves using the Fourier transform, and thus any band-limited waveform can be perfectly represented by samples at >2x the band limit under Nyquist
As someone who uses microwave as a mousepad, I appreciate your last test a lot ty
Another good video sir!
Love the Operation Flashpoint Easter Egg ❤
i do find it so fascinating how fast 3toesphilips monitor work
2 things that makes me think people might miss in latency is that internet speed/packets are set at a standard settings instead of peak like performance either in game or in windows.
Also windows might overlap mutiple inputs in a single sub tick of the game, and cs2 averaging it out.
Heyo king of the klik empire, awesome video as always but I do have this wish list request.
Did you happen look at if running CS2 using Vulkan rather than DX11 makes any difference to your findings? i made this switch before the reflex fixes were done as it improved my perceived performance (yet placebo is one hellavofa drug).
I'm mostly interested to see if im hindered myself at all doing this, i should note that im just running uncapped no reflex atm w/ a 3070ti + 6900HS laptop.
It can be forced on windows using the -vulkan start up arg.
I just love the cat cameos haha
Thanks for clarifying!
cool video, but now I need more of your cat in the next one too
great video fluffykins is hella weird tho, they look like a human trapped in a cat
In my experience, Fast Vsync works well in almost every DX11/DX12 game.
DX9 games usually don't work with it, and require in-game vsync to be enabled, which is usually not a huge deal as games of that era are more responsive than current gen anyway, mostly due to animation techniques and double buffering being the norm.
In my experience, Fast sync only works in fullscreen DX9/10/11 games and does nothing in DX12.
input latency lore is expanding by the second
2:21 love this song.
Your cat is adorable
I will say that a *poorly shielded* microwave once disrupted my 2.4ghz wifi A LOT. Like knocked it offline every time. Switching to 5ghz seemed to help a lot, and replacing it with a properly functioning microwave made the problem disappear. But on the off chance someone has a malfunctioning microwave, it certainly could interfere with 2.4ghz wireless connectivity.
Damn, I never knew about the 'fast' option. Seems a no brainer indeed.
Fluffykins? I love Fluffykins!
Excellent video! Could you please make a video about the few cases of auto overwatch banning people with high DPI? It is easily reproducible, and your attention/video might make valve notice.
The cat is the true latency. They distract you with snuggles while you are trying to game. You should do a test to see how you do in a match without the kitty vs how you do with the kitty!
Mr. Fluffykins appears to have a very long input (headpats) latency.
Oh thats good, cut to the confused cat wondering why there is a microwave in his favorite place to annoy you
I unironically used to play with my microwave beside me on my table, interesting information.
I love your cat, he is a treasure.