Current Generation Games Analysis Technical Discussion [2022] [XBSX|S, PS5, PC]

[davis.anthony]

I do love how he provides frame rates stats and I do wish the other channels would do it.

Interesting that PS5 has the highest minimum frame rate so it's slightly more consistent than XSX, even though with the low frame rate over all it's a touch moot.

 

[Remij]

I do love how he provides frame rates stats and I do wish the other channels would do it.

Interesting that PS5 has the highest minimum frame rate so it's slightly more consistent than XSX, even though with the low frame rate over all it's a touch moot.

Nah, the Series X is more consistent than PS5.

Frame Time Counts	PS5	Series X
16.67ms	17235 (78.09%)	18666 (81.95%)
33.33ms	4803 (21.76%)	4063 (17.84%)
50ms	32 (0.14%)	41 (0.18%)
66.67ms	1 (0%)	4 (0.02%)
83.33ms	0 (0%)	2 (0.01%)
116.67ms	0 (0%)	0 (0%)

PS5 = 4836 frames that are > 33.3ms
Series X = 4110 frames that are > 33.3ms

So Series X spends more time at 60fps, less time at 30fps, and has a handful of frames which are slight stutters.

 

[davis.anthony]

Nah, the Series X is more consistent than PS5.

Frame Time Counts	PS5	Series X
16.67ms	17235 (78.09%)	18666 (81.95%)
33.33ms	4803 (21.76%)	4063 (17.84%)
50ms	32 (0.14%)	41 (0.18%)
66.67ms	1 (0%)	4 (0.02%)
83.33ms	0 (0%)	2 (0.01%)
116.67ms	0 (0%)	0 (0%)

PS5 = 4836 frames that are > 33.3ms
Series X = 4110 frames that are > 33.3ms

So Series X spends more time at 60fps, less time at 30fps, and has a handful of frames which are slight stutters.

But in the case of the minimum frame rate like I was talking about PS5 is more consistent :/

 

[Remij]

But in the case of the minimum frame rate like I was talking about PS5 is more consistent :/

That doesn't make it more consistent though. It's a few one off frames that are delivered slightly slower on Series X.. Literally just 6 frames over 50ms out of 22000+ frames..

The entire reason why he provides this data is so that people can look and see "oh, there was a one off skip which resulted in that low minimum fps.. and it wasn't some prolonged dip"

And that's because "minimum fps" is a completely TERRIBLE metric by itself. Benchmarks and games can often have these one off frames drops (when changing a scene, or popping up a menu for example) that mess with minimum fps stats which don't represent the experience at all.

 

[PSman1700]

Might be better optimized for the XSX, or because it has more grunt.

 

[dobwal]

Nah, the Series X is more consistent than PS5.

Frame Time Counts	PS5	Series X
16.67ms	17235 (78.09%)	18666 (81.95%)
33.33ms	4803 (21.76%)	4063 (17.84%)
50ms	32 (0.14%)	41 (0.18%)
66.67ms	1 (0%)	4 (0.02%)
83.33ms	0 (0%)	2 (0.01%)
116.67ms	0 (0%)	0 (0%)

PS5 = 4836 frames that are > 33.3ms
Series X = 4110 frames that are > 33.3ms

So Series X spends more time at 60fps, less time at 30fps, and has a handful of frames which are slight stutters.

I wish they would add a VRR benchmark so we could see the impact of VRR on frametimes.

 

[see colon]

That doesn't make it more consistent though. It's a few one off frames that are delivered slightly slower on Series X.. Literally just 6 frames over 50ms out of 22000+ frames..

The entire reason why he provides this data is so that people can look and see "oh, there was a one off skip which resulted in that low minimum fps.. and it wasn't some prolonged dip"

And that's because "minimum fps" is a completely TERRIBLE metric by itself. Benchmarks and games can often have these one off frames drops (when changing a scene, or popping up a menu for example) that mess with minimum fps stats which don't represent the experience at all.

Davis is sort of right, though. PS5 has has a more consistent minimum framerate assuming you remove differences that are within a margin of error (anything less than 1%). PS5 has more frames at 33ms, therefore it is more consistent. PS5 has nearly 19% more frames at 33ms than Series X.

 

[BRiT]

PS5 has more frames at 33ms, therefore it is more consistent.

That's because the Series X has more frames at 16.67 ms.



(18666 + 4063) > (17235 + 4803)
22729 > 22038

 

[Remij]

Davis is sort of right, though. PS5 has has a more consistent minimum framerate assuming you remove differences that are within a margin of error (anything less than 1%). PS5 has more frames at 33ms, therefore it is more consistent. PS5 has nearly 19% more frames at 33ms than Series X.

lol... you're saying it has a consistently lower framerate than Series X... not that it has a more consistent overall framerate... Xbox also produces more frames overall.

 

[see colon]

That's because the Series X has more frames at 16.67 ms.

Right, Series X has a more consistent maximum framerate. Davis suggests that PS5 has a more consistent minimum framerate. As per his quote...

But in the case of the minimum frame rate like I was talking about PS5 is more consistent :/

So, assuming he meant that frametime = framerate here...
Like I said, if you throw out margin or error outliers (everything less than 1%, or .5%, or .3%, or .2%, or whatever you would consider rare enough to be a non-issue) then PS5, by the numbers, is more consistently at it's minimum frametimee. That would be 33.33ms or 30fps, which it achieves 21.76% of the time, or nearly 19% more often than Series X.

Or, if you include the margin of error numbers, PS5's worst frametime is 50ms which happens at a rate of .14% of all frames. Series X has a worse worst frametime at 83.33ms, but it only happens every .01% of the time. Therefore PS5 is more consistently at it's worst frametime. In fact it is at it's true minimum frametime 1400% more often than Series X. Therefore it is more consistently at it's minimum frametime.

PS5 is less often at it's maximum framerate. Therefore it is more consistently at it's most common minimum framerate (removing outliers). It also spends more time at it's true minimum framerate. It is therefore more consistent.

lol... you're saying it has a consistently lower framerate than Series X... not that it has a more consistent overall framerate... Xbox also produces more frames overall.

Yes, of course. That's what the numbers say. Series X spends more time at 16.67ms. It's more consistent there. Davis is saying that PS5 has a more consistent minimum framerate. That's true because PS5 spends more time there.

 

[BRiT]

This is consistently awkward wording...

 

[Remij]

Yes, of course. That's what the numbers say. Series X spends more time at 16.67ms. It's more consistent there. Davis is saying that PS5 has a more consistent minimum framerate. That's true because PS5 spends more time there.

He's said the PS5 has the higher minimum framerate... thus for that reason it's more consistent than Series X... He didn't look at frametimes at all. His argument is "PS5 didn't reach the lowest that Series X did"... but that doesn't tell the story.

If he was saying what you are.. he would say "PS5 has a consistently lower framerate than Series X"

Those are two completely different things. He's specifically using the minimum FPS stat to claim that PS5 is more consistent since it doesn't go as low... when you can look and see how often each console is at what frametimes.. you can tell Series X is more consistent to its target. It spends more time at 60fps, and less at 30 than PS5.

 

[Silent_Buddha]

Right, Series X has a more consistent maximum framerate. Davis suggests that PS5 has a more consistent minimum framerate. As per his quote...



So, assuming he meant that frametime = framerate here...
Like I said, if you throw out margin or error outliers (everything less than 1%, or .5%, or .3%, or .2%, or whatever you would consider rare enough to be a non-issue) then PS5, by the numbers, is more consistently at it's minimum frametimee. That would be 33.33ms or 30fps, which it achieves 21.76% of the time, or nearly 19% more often than Series X.

Or, if you include the margin of error numbers, PS5's worst frametime is 50ms which happens at a rate of .14% of all frames. Series X has a worse worst frametime at 83.33ms, but it only happens every .01% of the time. Therefore PS5 is more consistently at it's worst frametime. In fact it is at it's true minimum frametime 1400% more often than Series X. Therefore it is more consistently at it's minimum frametime.

PS5 is less often at it's maximum framerate. Therefore it is more consistently at it's most common minimum framerate (removing outliers). It also spends more time at it's true minimum framerate. It is therefore more consistent.


Yes, of course. That's what the numbers say. Series X spends more time at 16.67ms. It's more consistent there. Davis is saying that PS5 has a more consistent minimum framerate. That's true because PS5 spends more time there.

I see what you did there.

In other words, Davis is saying that the PS5 version consistently performs worse then the XBS-X verison.

Regards,
SB

 

[Silent_Buddha]

"Which one is PS5 and which is Series X?"

Does anyone want to tell him that if you 'right click -> Open in new tab' it tells you the resolution of the images so it's easy to know which one is XSX as it's the image that's slightly higher in resolution than the other

https://twitter.com/i/web/status/1538192154442534914

What a joke...

x frames after movement cut to allow TAA settle. Which one is PS5 and which is Series X?

You mean take away the one extremely major area what distinguishes a good temporal implementation from a not so good temporal implementation, motion, to determine whether they are similar in quality? Is this a joke? Does this guy actually have any technical knowledge?

Does this guy only play games with his hands off the controller so that the visuals never show any motion? HIs perfect game must be the one that runs at 1 frame per day since that should produce the best visuals when not in motion.

I'm absolutely flabbergasted that he actually posted something like that to try to justify his analysis.

Regards,
SB

 

[pjbliverpool]

What a joke...



You mean take away the one extremely major area what distinguishes a good temporal implementation from a not so good temporal implementation, motion, to determine whether they are similar in quality? Is this a joke? Does this guy actually have any technical knowledge?

Does this guy only play games with his hands off the controller so that the visuals never show any motion? HIs perfect game must be the one that runs at 1 frame per day since that should produce the best visuals when not in motion.

I'm absolutely flabbergasted that he actually posted something like that to try to justify his analysis.

Regards,
SB

He's actually admitted in a more recent tweet on that thread that his point was to demonstrate how similar the images are. I've asked him to clarify to what end he's demonstrating that and whether it has any bearing in his recent RE2 video.

 

[Flappy Pannus]

He's actually admitted in a more recent tweet on that thread that his point was to demonstrate how similar the images are. I've asked him to clarify to what end he's demonstrating that and whether it has any bearing in his recent RE2 video.

Um, isn't showing "how the sausage gets made" literally the point of these videos?

https://twitter.com/i/web/status/1540677975477207040

Like if your video is just about reviewing the patches for a particular platform and how well they pull off any upgrades, then subjective opinions are fine - especially if you state you're unsure on certain aspects of how they're delivered. But a significant chunk of his videos are about platform comparisons, which includes the differences in implementations and what effect that has on performance. Ridiculous.

 

[BRiT]

I guess no one should bother watching any of the videos then since mostly all game output looks the same...

 

[Remij]

Maybe he should just stick to twitter polls then

 

[dobwal]

He's said the PS5 has the higher minimum framerate... thus for that reason it's more consistent than Series X... He didn't look at frametimes at all. His argument is "PS5 didn't reach the lowest that Series X did"... but that doesn't tell the story.

If he was saying what you are.. he would say "PS5 has a consistently lower framerate than Series X"

Those are two completely different things. He's specifically using the minimum FPS stat to claim that PS5 is more consistent since it doesn't go as low... when you can look and see how often each console is at what frametimes.. you can tell Series X is more consistent to its target. It spends more time at 60fps, and less at 30 than PS5.

Agreed.

It makes little sense to direct attention to those frame times below 33.3 3ms. Let's focus on the <0.25% of the frames while ignoring the 18-22% of frames that don't hit the target frame time.

It takes ~5 1/2 minutes to run through 20,000 frames with a perfect frame time of 16.67 ms for every frame. The sub 33.33 frames on the PS5 would add about 0.8 seconds to that time while the sub 33.33 frames on the Series X would add 1.7 seconds. However, those 33.33 frames on both consoles add roughly 2 minutes as both the Series X and PS5 take almost 7 1/2 minutes to spit out 20,000 frames.

It makes his statements nonsensical.

The most relevant thing to me is that both the Series X and PS5 will have you spending about 20% of playtime dealing with higher input latency.

 

[pjbliverpool]

Um, isn't showing "how the sausage gets made" literally the point of these videos?

https://twitter.com/i/web/status/1540677975477207040

Like if your video is just about reviewing the patches for a particular platform and how well they pull off any upgrades, then subjective opinions are fine - especially if you state you're unsure on certain aspects of how they're delivered. But a significant chunk of his videos are about platform comparisons, which includes the differences in implementations and what effect that has on performance. Ridiculous.

Yep I figured thats where he was going with it. I've asked if that means he'll be using DLSS in future performance comparisons where it's available.