AMD Radeon AntiLag

[masneb]

Input lag is additive. You aren't talking solely about the time between frames, but rather time in which frames can be messed up. For instance take popular serpentine movement in a FPS. When someone jukes back the opposite direction with AD strafing, you're going to overshoot. This is why buffering is really bad and input delay is a buffer. It doesn't matter how much FPS you get, because you're buffering your actions by the input delay.

FPS and Input Delay matter immensely, both in their own right. Unfortuantely very few people have talked about input delay outside of monitor testing.

The new tech looks highly promising and the demo is definitely true. Something to note is the Nvidia system is 4fps less then the AMD system, it's set to 55-56 while the AMD system is set to 59-60. You can tell they're using their own internal version of the mouse light tester with the sync at the top of the screen. It would've been more helpful and more understandable both to their audience (derpy gamers) and to people watching it with technical background if they actually showed their testing setup for context. When they turn on their Input Delay Reduction (I refuse to call it anti-lag), you can see the screen doing really weird and janky stuff with the timer, probably due to some sort of time syncing issue. This is very good, it definitely means something weird and alpha is happening.

This is first gen technology, it's only going to get better the more they focus on it. Now it's basically proof of concept and it's showing 30% gains... That's very impressive.

Nvidia has not done anything like this or even mentioned anything similar. They're just playing 'me too' with something they don't understand. AFAIK AMD has always had 1 pre-rendered frame ahead, that's why you can't change it. There used to be a registry way of changing it, but that was no longer needed.

[Stitch7]

Input lag is additive. You aren't talking solely about the time between frames, but rather time in which frames can be messed up. For instance take popular serpentine movement in a FPS. When someone jukes back the opposite direction with AD strafing, you're going to overshoot. This is why buffering is really bad and input delay is a buffer. It doesn't matter how much FPS you get, because you're buffering your actions by the input delay.

FPS and Input Delay matter immensely, both in their own right. Unfortuantely very few people have talked about input delay outside of monitor testing.

The new tech looks highly promising and the demo is definitely true. Something to note is the Nvidia system is 4fps less then the AMD system, it's set to 55-56 while the AMD system is set to 59-60. You can tell they're using their own internal version of the mouse light tester with the sync at the top of the screen. It would've been more helpful and more understandable both to their audience (derpy gamers) and to people watching it with technical background if they actually showed their testing setup for context. When they turn on their Input Delay Reduction (I refuse to call it anti-lag), you can see the screen doing really weird and janky stuff with the timer, probably due to some sort of time syncing issue. This is very good, it definitely means something weird and alpha is happening.

This is first gen technology, it's only going to get better the more they focus on it. Now it's basically proof of concept and it's showing 30% gains... That's very impressive.

Nvidia has not done anything like this or even mentioned anything similar. They're just playing 'me too' with something they don't understand. AFAIK AMD has always had 1 pre-rendered frame ahead, that's why you can't change it. There used to be a registry way of changing it, but that was no longer needed.

I think it's weird they did not show the setup. Also I hope they were using the same monitor with the same overdrive and espescially the same camera and input device. The best methodolgy in my opinion is an arduino device with an led and a m1 click. A fast engine known for input reliability and a black room. Not exactly real-world use case but much better than what they showed.

[masneb]

It was a arduino device, you can check it out in mellos link.

https://youtu.be/uPpdtXFx7gQ?t=147

I wasn't asking for their testing setup because I was doubting their results, rather because I believe a lot of gamers don't understand input lag and how it influences their game, therefore when this popped up during e3 there were a lot of people with ???? above their head in the audience.

[Stitch7]

It was a arduino device, you can check it out in mellos link.

https://youtu.be/uPpdtXFx7gQ?t=147

I wasn't asking for their testing setup because I was doubting their results, rather because I believe a lot of gamers don't understand input lag and how it influences their game, therefore when this popped up during e3 there were a lot of people with ???? above their head in the audience.

You're right I missed thst part. Didthey use a high framerate camera?
The more I talk about it the more exciting it is.
I really hope Chief will do some quality testing for this.

[open]

I wasn't asking for their testing setup because I was doubting their results, rather because I believe a lot of gamers don't understand input lag and how it influences their game, therefore when this popped up during e3 there were a lot of people with ???? above their head in the audience.

I hope this draws attention to the matter. If it really does shave off 10ms of input lag then some people may be able to see the difference 10ms can make. Alot of times we quantize things vs lets say something like human reaction time ~200ms+. But simply having the feel of the mouse movements closer to the screen updates changes how we can aim. People who have experimented with refresh sync techniques to reduce input lag know this very well.

[Notty_PT]

Isn´t this suppose to just recude input lag on 60fps cap conditions tho? Because I seen the demonstration with those conditions.

[open]

It might be able to scale with frame time. So 144hz would be almost 5ms. 240hz would be 2.7ms. Still welcome. Another part of a good setup. Every ms counts.

[Notty_PT]

It might be able to scale with frame time. So 144hz would be almost 5ms. 240hz would be 2.7ms. Still welcome. Another part of a good setup. Every ms counts.

Yes but from what I seen it only works on 60fps Vsynched situations, we have to wait to see what it does when the reviews are out on 7th July

[Stitch7]

It might be able to scale with frame time. So 144hz would be almost 5ms. 240hz would be 2.7ms. Still welcome. Another part of a good setup. Every ms counts.

Yes but from what I seen it only works on 60fps Vsynched situations, we have to wait to see what it does when the reviews are out on 7th July

Notty I think this might be the decision maker for me. I really do hope it is worth it. If it is Navi 20 I'm coming for you.

[masneb]

I wasn't asking for their testing setup because I was doubting their results, rather because I believe a lot of gamers don't understand input lag and how it influences their game, therefore when this popped up during e3 there were a lot of people with ???? above their head in the audience.

I hope this draws attention to the matter. If it really does shave off 10ms of input lag then some people may be able to see the difference 10ms can make. Alot of times we quantize things vs lets say something like human reaction time ~200ms+. But simply having the feel of the mouse movements closer to the screen updates changes how we can aim. People who have experimented with refresh sync techniques to reduce input lag know this very well.

I do too. I hope normal tech review websites start picking this up and add it to normal testing they do for video cards and other hardware in gaming. It's very crude and rudimentary right now with the mouse button test, AMD definitely had a better setup, but even that can be refined further. There are plenty of unique tool makers on the market, I'm surprised someone doesn't have a solution for this.

Yes, also anyone who attempts to sum up humans in some sort of rigid number is blatantly talking out of their bum. As esports become more recognized and more money pours into them as far as testing and lifestyle (sports coaches, nutrition specialists, psychology), I'm sure this all will become much more streamlined and substantiated.

Notty I think this might be the decision maker for me. I really do hope it is worth it. If it is Navi 20 I'm coming for you.

I agree, I might downgrade my GPU throughput performance (2080) to upgrade my input delay performance. This is a big deal and heavily understated how big of a impact this is to gamers.

Isn´t this suppose to just recude input lag on 60fps cap conditions tho? Because I seen the demonstration with those conditions.

No! I read the articles and watched the video. While Scott (retired from Techreport) was talking about it having diminishing returns the higher you get in FPS, it isn't because input delay goes away, rather AMD has less control over what is being bottlenecked. They showed a 60fps scenario (in my mind) because it shows exclusively a input delay bottleneck, not a frame throughput bottleneck (typical FPS), and definitely not a CPU bottleneck.

If they're streamlining and improving the GPU stack including their drivers and a game becomes CPU bound, there isn't much AMD (graphics) can do to improve that. In a complete Ryzen system that means AMD COULD potentially tackle this from the USB port all the way through your GPU. Obviously this technology is still in its infancy and we may eventually see a whole OS implementation or specialized input stack eventually derived for peripherals with ways for games to address it exclusively in the long run. Judging on how slow Microsoft is to helping gamers, I doubt we'll see them doing anything regarding this and whatever they attempt to do is just going to ruin whatever Nvidia/Intel/AMD come up with.

This isn't the same as refresh timing and talking about the difference between cycles in a monitor. There isn't a diminishing return in input delay.


Just going to reiterate that input delay is not happening in parallel to frame draw, it's additive. It's in addition to whatever bajillion high FPS your monitor and computer is getting. The moment something can't be buffered in your mind because it acts unpredictable, it doesn't matter how much FPS you get, you'll miss your shot.

Also AMD really, really, really needs to rebrand this technology like Input Delay Reduction tech - IDR or something. Anti-lag is so stupid and covers soooo many things.