“Competitive” gamers, the ones who think they need 200+ fps at 200Hz+.
Aren’t you proposing that frame gen introduces unacceptable latency even though the absolute increase in latency is minimal? Does that mean you’re also in the minority?
Current Generation Games Analysis Technical Discussion [2024] [XBSX|S, PS5, PC]
- Thread starter Shifty Geezer
- Start date
“Competitive” gamers, the ones who think they need 200+ fps at 200Hz+.
Aren’t you proposing that frame gen introduces unacceptable latency even though the absolute increase in latency is minimal? Does that mean you’re also in the minority?
Charlietus
Regular
Can you please quote where I say that frame gen introduces unacceptable latency?
You claimed latency is the most important aspect of performance. You then disagreed that the impact of FG latency was a subjective issue. And also said latency can make or break the experience.
My point was that latency you find acceptable will be considered unacceptable by more “elite” gamers just like other people are perfectly fine with latency that you would consider to “break” the experience. Hence it’s subjective and latency isn’t the most important aspect of performance for everyone.
Charlietus
Regular
I'm not arguing that everyone has the same level of acceptance of latency, I'm saying that in a game, where game feel is usually more important than visuals, latency directly affects the game, and so it's the most important aspect of performance.
Lower latency gives you an advantage, even if you don't notice it.
I don't think that's a universal truth for everyone. It'll depend on player and game type. I think it also worth noting how much latency we're considering. An extra 30 ms might be imperceptible to most players on one type of game (Pikmin, say) so upscaling from 30fps to 60fps won't bother the playerbase. Whereas with other games like competitive shooters, higher input latency may matter BUT if you are already running a high refresh rate (60) and are pushing for higher (120), that additional latency might not be high enough to matter. Higher framerates isn't just about lower latency but also higher visual feedback on game state, and the smoother motion from 120 Hz upscaled might be preferential to some than the lower latency of 60 fps.
A truck load of user research is required - we don't have enough knowledge yet to know what the best average target is (for a given game style).
Charlietus
Regular
Let's remember where the conversation started: if frame generation is considered performance. It's a pretty broad topic, so unless I were to write 10000 words on the topic, I have to generalize I bit. There are so many opinions and preferences on this stuff.
But on how a reviewer should consider frame generation, it wouldn't sit well with me if he said that a game is running at 120 when it's actually running at 60 + FG. That objectively isn't the same experience that a "true" frame rate gives you, no matter if you are sensitive to this stuff or not.
If they’re reviewing hardware then I think it’s fine as long as all the hardware has frame generation enabled or it’s clearly called out where numbers are higher due to frame generation.
Agreed, 120fps with frame generation isn’t the same as true 120fps and is objectively inferior. However we don’t have a good definition of “true” in this context either.
I don’t think we can outright dismiss FG as not improving performance. It arguably improves some aspects of performance (number of displayed games) while hurting others (input latency). Not to mention IQ implications. So it’s up to the end user to understand those performance and IQ tradeoffs and decide whether they’re worth it.
Charlietus
Regular
Absolutely agree. Transparency and clarity on the data is key.
I've already discussed my view there - reviewers shouldn't talk about 'performance' that way and should clarify the different aspects that they might group under 'performance'. I feel what you've posted is tangential to that, on how people experience input and feedback. I agree with you that games shouldn't simply described as "120Hz" is they are framerate upscaled, but I disagree that input latency is of primary importance.
133ms after an early patch, which is not great, but not disastrous for a 30fps game, either.
Either way, you seem very selective over which points you think are 'relevant'. Prey on PS4 has a User Score of 8.2 on Metacritic, which would suggest that you are in the 'small minority' of people who think it's unplayable cuz of this issue. Yet you still thought it was relevant to bring up anyways.
So it looks like I spoke too soon or MS has changed stance. Starfield looks like the first 1st party game to have a better framerate mode on the X vs S. Series X is getting a 60fps mode with the May update.
DavidGraham
Veteran
Not necessarily. True 120fps can be worse if it's running with bad frame pacing, has frequent fps drops or stutters, and runs with Vertical Synchronization.
In that case, 120fps with frame gen can be objectively more superior if it's devoid of any frame pacing/fps drops/stutters issues, runs with Reflex and doesn't use Vertical Synchronization. It will achieve smoother playback as well as better useful/predictable and lower latency.
Which is why I feel context matters in these cases, it's no longer one variable, rather a multitude of them balanced in certain ways to achieve a certain level of experience.
Are any of those things related to frame generation though? Of course you can always toss in random negative things to only one side of any comparison but I don’t see the point in that.
Is that the Aloy texture issue in photomode that you're referring to? I'm not seeing anything there relating to general texture streaming issues in vram limited scenarios which seemed to be resolved in the previous patch for me.
Yea, specifically that issue.
Why would they have that in these new patch notes if you say they seem to have been solved in the last one? lol
The issue I was referring to from the last patch was how the game handles vram overflow (primarily texture streaming).
Nothing on that, specifically. You said it was already greatly improved though, right?
It's fixed as far as I've seen from my wife's playthough. No more serious drops with very high textures on my system after multiple hours of play.
Cappuccino
Newcomer
I can pretty immediately feel when FG is on latency wise, but I can’t tell the difference between Reflex on vs Reflex off. The input lag on keyboard and mouse with FG feels horrible to me.
On tick rate vs fps: I think the biggest issue is the end user doesn’t ‘see’ physics interpolations. It is very clear to me when a frame is ‘fake’ by the fact the input felt laggier and there is some artifacting. I’m not examining physics simulations in the same way.
Responsiveness imo is the big reason to run a high framerate. I did play Spiderman with FG on but that was with a controller so you can’t really feel latency the same way on controller. That game also didn’t artifact very much.
I would also say that part of the disconnect is that we’ve been conditioned to accept higher latency when fps drops. When I play at 60Hz, I expect 60fps levels of latency. With FG my eyes see basically 120, but my hands feel 60, which makes it feel ‘fake’ if that makes sense.
Similar threads
- Locked
