Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

LatencyFleX: A new approach to game latency reduction

ishitatsuyuki.github.io

Post reply on HN
Print view View original

61–69 of 69 posts

Re: LatencyFleX: A new approach to game latency reduction

#61

Post by Melatonic »

terafo wrote: post #42
eertami wrote: post #36

Earlier quoted context omitted.

There's definitely diminishing returns the higher we go with refresh rates. 60hz to 240hz for example is like playing a completely different game. But going from 240hz to 360hz, even in CSGO it's a lot harder to notice a difference. Personally I believe the newly announced 300hz 27" 1440p monitors[0] are going to be the perfect sweet spot for the foreseeable future. I imagine it will be a long time before technology…

There are definitely diminishing returns with increase of refresh rates. But nonetheless your comparison is unfair, since you are comparing quadrupling of refresh rate with a mere 50% increase, which is similar to comparing 60 to 90, not 60 to 240. And with advent of VR demand for refresh rate increase of display panels will only grow, since it's much more noticeable while using headset.

Not sure I understand exactly what you are saying

As for VR I think that's an excellent illustration of my point - we know that many people don't do well at 60FPS per eye in VR due to motion sickness. Move up to 90FPS per eye though and there is a massive improvement that I have seen first hand others. By the time you get to 120FPS the experience feels pretty damn smooth and while I would of course like to see more frames I am not convinced going a ton beyond 120 is really worth it performance wise considering you have to render that twice and the extra compute could be instead spent on the new shiny like ray tracing.

Luckily when I was working at a VR startup I was one of the few people that never seemed to get motion sickness so I became the test dummy for everyone's work - they would throw me in something they hadn't optimized at all yet that was only getting 40fps on a system with dual Xeons and quad Nvidia top of the line workstations cards and while it felt a little weird it for some reason never bothered me :-D

Re: LatencyFleX: A new approach to game latency reduction

#62

Post by TacticalCoder »

snovv_crash wrote: post #45
Melatonic wrote: post #43

Earlier quoted context omitted.

I think people were saying 60fps was the limit but still I agree That being said in the quake days I dont think monitors could go over 60hz anyways so even at 120FPS you were not gaining a similar advantage from what we have today. From what I remember however there were other advantages to high FPS in games like Counterstrike as well in terms of player movement - the monitor might have "smoothed" the motion back dow…

My dad had a big old 19" Nokia CRT that did 90Hz. Of course resolution was capped because this was all being controlled from VGA.

I'm pretty sure I was playing Counter Strike at, as reported by the game, 99 fps on my Sony Trinitron back then (2000?). We'd use "low poly" mods (simplified 3D characater models) to speed everything up and reach these speeds.

There are still messages on message boards from this era where several people mention reaching that same 99 fps.

It was two decades ago but I'm pretty sure it was 1024x768 @ 99 fps (19" Sony Trinitron). I may be mistaken on the monitor though.

Re: LatencyFleX: A new approach to game latency reduction

#63

Post by mrob »

eertami wrote: post #36
beebeepka wrote: post #34

I just want to say I am glad pro gaming took over. Back in the day it was only Quake players advocating for 120 FPS (for various reasons, including Q3 physics being somewhat broken), 125hz mice and stuff like that. I am talking 20 years ago. The number of lost souls parroting the old "human eye can only see 30 fps" has gone down considerably over the years. The last 10 years were fantastic in that regard, despite the…

There's definitely diminishing returns the higher we go with refresh rates. 60hz to 240hz for example is like playing a completely different game. But going from 240hz to 360hz, even in CSGO it's a lot harder to notice a difference. Personally I believe the newly announced 300hz 27" 1440p monitors[0] are going to be the perfect sweet spot for the foreseeable future. I imagine it will be a long time before technology…

There are diminishing returns, but 360Hz is still too low to display sharp-looking motion without strobing. 360Hz strobing is visible as phantom array effect whenever you move your eyes. If you are sensitive to this artifact and instead want motion that looks like real life, you need more like 1000fps/1000Hz. There is no hardware capable of this, but at high enough frame rate you could probably get away with interpolating the frames along motion vectors, e.g from 500fps to 1000fps, with very minor latency/artifacting.

See: https://blurbusters.com/faq/oled-motion-blur/

Re: LatencyFleX: A new approach to game latency reduction

#64

Post by PudgePacket »

zamadatix wrote: post #16
moffkalast wrote: post #13

Earlier quoted context omitted.

I'm sure Godot will integrate it in version 4.1... when it comes out in 10 years.

I'm all for poking fun but this just seems excessive as Godot isn't even 10 years old.

https://godotengine.org/article/godot-history-images :)

Re: LatencyFleX: A new approach to game latency reduction

#65

Post by bschwindHN »

duped wrote: post #18
snarfy wrote: post #8

Earlier quoted context omitted.

There are a variety of factors that add up to many ms of lag. USB and HDMI are a few that come to mind. USB is incredibly complex when compared to e.g. ps/2. If you are building an embedded board like rasberry pi and you want to add ps/2, you just add a couple pins. If you want to add usb, you buy an extra chip.

Many CPUs have USB support. Pin count also doesn't correlate to latency, what matters is the minimum buffer size and transfer speed. USB can beat PS/2 for latency without issue.

Ben Eater has a good video on this:

https://youtu.be/wdgULBpRoXk?t=1765

Re: LatencyFleX: A new approach to game latency reduction

#66

Post by ksec »

beebeepka wrote: post #34

I just want to say I am glad pro gaming took over. Back in the day it was only Quake players advocating for 120 FPS (for various reasons, including Q3 physics being somewhat broken), 125hz mice and stuff like that. I am talking 20 years ago. The number of lost souls parroting the old "human eye can only see 30 fps" has gone down considerably over the years. The last 10 years were fantastic in that regard, despite the…

>I just want to say I am glad pro gaming took over.

Yes! I have been crying about latency for nearly 10 years [1]. Computing has always been optimising for throughput. And before Pro Gaming, there just hasn't been a marketable incentive for companies to work on / minimise latency. Now we finally do!

Even in the best case scenario, the lowest latency is still 25ms, and in most cases we are still above 50ms. I think it is worth posting [2] Microsoft Research on Input latency. It would be nice if we could get average system end to end latency down to sub 10ms level. Which is what I hope work on VR will bring us.

[1] https://news.ycombinator.com/item?id=6422632

[2] https://www.youtube.com/watch?v=vOvQCPLkPt4

Re: LatencyFleX: A new approach to game latency reduction

#67

Post by ksec »

hffftz wrote: post #5
ishitatsuyuki wrote: post #4

Earlier quoted context omitted.

Input lag has nothing to do with network ping. Input lag is the time between you perform the action and the computer shows that on screen. It depends on your frame rate, refresh rate, and peripheral polling rate, as well as how good the game schedules things (which is what LatencyFleX tries to optimize). Network ping on the other hand is often hidden away. Whether you are on 2ms ping or 100ms ping, the bullet always…

How can you get input lag of 20ms? do you have to slow down your code? I only implied it was network lag because it was so high.

Those are End to End System latency. Your USB Input Devices and Monitor Refresh Rate, excluding anything in between, could easily had 20ms delay in the worst case scenario. A 60Hz monitor would have had 16.6ms of worst case latency alone.

Re: LatencyFleX: A new approach to game latency reduction

#68

Post by formerly_proven »

Melatonic wrote: post #44
eertami wrote: post #36

Earlier quoted context omitted.

There's definitely diminishing returns the higher we go with refresh rates. 60hz to 240hz for example is like playing a completely different game. But going from 240hz to 360hz, even in CSGO it's a lot harder to notice a difference. Personally I believe the newly announced 300hz 27" 1440p monitors[0] are going to be the perfect sweet spot for the foreseeable future. I imagine it will be a long time before technology…

I am not convinced we need to go nearly that high - 300hz puts a crapload more stress on your system performance wise for not much gain. 90hz is in my opinion already such a massive improvement over 60 that I do not see mind blowing results even going to 120 or 144. And many pro gamers were using 120hz monitors over 144hz at one point. Realistically I think the two sweet spots are 120hz and 240hz - not necessarily be…

> And many pro gamers were using 120hz monitors over 144hz at one point.

This was done solely in order to enable strobing as 144 Hz panels at the time were too slow to support strobing which requires scanout speeds equivalent to ~200 or so Hz at 144 Hz.

And strobing is solely used because S&H displays have too much transition and motion blur at 1xx and 2xx Hz.

> Realistically I think the two sweet spots are 120hz and 240hz - not necessarily because they are the best of the best but because they are each divisible by both 24 and 30 (the most common FPS of films and television)

Principally I agree that 120/240 Hz are more suited to general purpose use for this reason [1], but on the other hand this really has nothing to do with the hardware and is purely so because of software limitations. Really what one would like to see is that video playback causes the variable-display refresh to adjust to a multiple of the exact video frame rate instead of janky ad-hoc frame-rate conversions.

This is a common theme; hardware is generally much more capable than what the software/drivers allow everywhere you look.

> All of this however is nothing compared to the improvement that a true HDR display brings

Many people would probably already be quite happy with something that doesn't turn shadows into a foggy, cloudy mess like all IPS panels do, and VA panels as well (but less so).

[1] though 120 Hz does not solve the 50p problem, as content produced by broadcasters in 50 Hz countries, which is basically all of the world that isn't the US, cannot easily be converted at playback time. 25p can just be handled like 24p with 1:1 playback, basically nobody notices the slight speed-up / slow-down and that's how films have been shown in television in 50 Hz countries since always.

Re: LatencyFleX: A new approach to game latency reduction

#69

Post by formerly_proven »

badsectoracula wrote: post #52
Melatonic wrote: post #43

Earlier quoted context omitted.

I think people were saying 60fps was the limit but still I agree That being said in the quake days I dont think monitors could go over 60hz anyways so even at 120FPS you were not gaining a similar advantage from what we have today. From what I remember however there were other advantages to high FPS in games like Counterstrike as well in terms of player movement - the monitor might have "smoothed" the motion back dow…

> I forget how refresh rate worked on CRT's though - maybe those could higher than 60? Yes, even the standard VGA 13h mode (320x200x8) is 70Hz and many CRTs could do 85Hz. By Quake 3's time CRTs that could do 120Hz and above were very common. Personally i have such a CRT as well as another that can do 160Hz. Also FWIW the refresh rate is only part of the story - CRTs have practically instant "response time" so 120Hz…

A large reason why CRTs are rather excellent in regards to motion is because they're not sample and hold displays resulting in very low duty cycles (dominated by the phosphor's fall time of somewhere between 200-1000 µs, as rise time is basically instant at <20 ns or so and video bandwidth is well above 100 MHz). That's the main reason why a 240 Hz LCD using BFI for strobing (D=0.5) can't compete with a 120 Hz CRT (D~0.05 or so).
Post reply on HN