• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

Digital Foundry: The Finals - PS5/Xbox Series X/S - DF Tech Review - Destruction Physics at 60FPS

adamsapple

Or is it just one of Phil's balls in my throat?



Sporting some of the most impressive destruction physics seen so far this generation, The Finals pushes the Unreal Engine in an exciting direction. It's a free to play game - but surprisingly backed by robust mechanics that hark back to the days of DICE's Battlefield Bad Company 1 and 2. And the best news is it runs at 60 frames per second on console.



00:00:00 Introduction
00:01:46 Tech Overview
00:05:26 PS5 vs Xbox Series X/S Graphics Comparison
00:08:41 Xbox Series S Frame-Rate Test
00:10:57 Xbox Series X Frame-Rate Test
00:12:11 PlayStation 5 Frame-Rate Test
00:13:26 Conclusion
 
Last edited:

adamsapple

Or is it just one of Phil's balls in my throat?
- Built on Unreal Engine 5
- Lots of physics based destruction with a lot of vertical scalability in environments
- RTGI is included on all 3 console versions (PS5/SX/SS)
- Quality is lower than on PC and it can take a few seconds for the light probes to solve themselves (can take 10~ seconds after destroying a ceiling for example)
- All consoles target 60 FPS out of the box with DRS.

Visuals:

PS5 / Series X:

- 1440p to 2160p DRS with up-scaling to reach 4K.
- Visual features are a match between the two

Series S:
- 720p to 1512p DRS
- Drops terrain quality compared to SX, but this is a minor issue
- Texture filtering is lower, which is more notable, at oblique angles
- Texture asset quality is the same as Series X tho.
- Shadow quality and filtering uses lower pre-sets

Performance:

- Series S:

-The visual cut-backs justify the means.
- 60 FPS very closely locked in DF's testing
- Only able to see minor drops when shifting camera from sky to ground (trigger highest DRS)

- Series X:
- Water-tight 60 FPS 99.9% of the time with only one-off drops in physics heavy explosions

- PS5:
- Much like Series X but when it drops, it drops harder and longer stretches than Series X
- Drops also don't resolve back to 60 FPS as quickly as on Series X
 
Last edited:

DenchDeckard

Moderated wildly
So ps5s narrow approach is not enjoying this destruction in this game. Unreal engine 5 with this game must prefer a wider slower approach for the destruction etc.

Xbox drops like 1 fps in total while ps5 is dropping to 50fps.
 

Darsxx82

Member
It is an impressive work by Embark in all its versions.

The XSS version magnificently maintaining 60fps while RTGI and a decent resolution in UE5 deserves credit.
XSX (probably the version with the least optimization time compared to the other two) is really strong in all situations.

I want to start seeing UE5 games with target 30fps and all its possibilities (Nanite, Lumen/RT Lumen).
Until now (although really 2023 has been practically the first year with UE5 games), the majority are created with 60fps in mind and with that standard and coupled with the fact that UE5 is still in the optimization process... it is difficult to see graphical exponents and visuals on these consoles.
 

amigastar

Member
You would think that 16x Anisotropic Filtering would be standard on todays games but Series S doesn't look like it has that.
 

adamsapple

Or is it just one of Phil's balls in my throat?
1440p->4k, RTGI, 60FPS, UE5, and tons of destruction? These developers are to be commended.

The Series S version also looks and runs pretty damn good.

The video talks about a notable anisotropic filtering difference between the two, but honestly, it looks perfectly fine in stills.

These devs are pretty competent.

COMP-3_3j89A5O.jpg
 
Last edited:

Mr Moose

Member
Xbox drops like 1 fps in total while ps5 is dropping to 50fps.
It drops to 58 (which is still impressive) using the bar, there's something wrong with the number in the corner (wtf is the point in .0?).
58:
Screenshot20240113at.png

59:
Series-X-FPS-2.jpg

Though drops are quite rare on all machines, they did a pretty good job on it. Hadn't even heard of the game until a week or so ago.
 

DenchDeckard

Moderated wildly
It drops to 58 (which is still impressive) using the bar, there's something wrong with the number in the corner (wtf is the point in .0?).
58:
Screenshot20240113at.png

59:
Series-X-FPS-2.jpg

Though drops are quite rare on all machines, they did a pretty good job on it. Hadn't even heard of the game until a week or so ago.

I just went by the number in the top right.

Ps5 had some serious lurches when shown and dropped all the way to 50fps.

Still though. Great on all really. Just that xbox takes the win on this one, when comparing launch versions.

Its funny how games waver back and forth with the wins. Really interesting seeing what games prefer what architecture.
 
Last edited:

twilo99

Gold Member
A locked 60 FPS on a UE5 DRS 4K game with RTGI is pretty impressive for the current consoles tho :p

Series S is practically 100% % locked 60, even non UE5 games barely manage that.

True, devs have done a good job. This should be the norm for all GaaS shooters in terms of consistency across the available boxes, but still, 120FPS should be the standard at this point.

Its funny how games waver back and forth with the wins. Really interesting seeing what games prefer what architecture.

I think by "games prefer" you mean the devs having different skill sets and understanding of the provided APIs, which then results in difference in performance..
 
Last edited:
True, devs have done a good job. This should be the norm for all GaaS shooters in terms of consistency across the available boxes, but still, 120FPS should be the standard at this point.



I think by "games prefer" you mean the devs having different skill sets and understanding of the provided APIs, which then results in difference in performance..

After so many years it's not just down to APIs at this point. Each system has its own set of advantages over the other. But neither is vastly more powerful than the other which is why the differences are small.

Only time I can see a big difference is when they are compared to PCs or the Pro models.
 
Last edited:

Vergil1992

Member
What I don't like about DF's analyzes is that now they don't even compare the same scenario. First they were extensive side-by-side comparisons, then "selected" areas to show how each worked, and now finally in very different scenarios and situations.

I'm not saying DF doesn't have credibility, they're probably right, especially in terms of frame rate, but the truth is I like seeing more than believing. In that comparison, PS5 clearly has a worse framerate, it drops more frequently and the drop is worse (in the mid-50s), but it is a completely different scenario and different situations. Admittedly, that behavior suggests that under stress XSX appears to perform better, but I want to see it, not hear it. I insist, I don't think they lie. The comparison seems valid to me and I don't think there is bias, but I think they are going easy. The analyst has probably tried both versions, he has drawn the conclusion, but he has not wanted to make the "effort" to match versions because it is much more difficult and requires more time.
 

SomeGit

Member
What I don't like about DF's analyzes is that now they don't even compare the same scenario. First they were extensive side-by-side comparisons, then "selected" areas to show how each worked, and now finally in very different scenarios and situations.

I'm not saying DF doesn't have credibility, they're probably right, especially in terms of frame rate, but the truth is I like seeing more than believing. In that comparison, PS5 clearly has a worse framerate, it drops more frequently and the drop is worse (in the mid-50s), but it is a completely different scenario and different situations. Admittedly, that behavior suggests that under stress XSX appears to perform better, but I want to see it, not hear it. I insist, I don't think they lie. The comparison seems valid to me and I don't think there is bias, but I think they are going easy. The analyst has probably tried both versions, he has drawn the conclusion, but he has not wanted to make the "effort" to match versions because it is much more difficult and requires more time.
Logistically it's extremely difficult. Even on PS5 the readout is 99% of the time 60FPS, so the side-by-side areas are probably a flat 60 or close to it.

Unlike BRs or other MPs, you can't just die and spectate another player/team and have a perfect side-by-side so the only way to record similar area would be to play the game with someone else and stick very close to him which in a game like this would be a nightmare. It's just not worth it.
 

Vergil1992

Member
Logistically it's extremely difficult. Even on PS5 the readout is 99% of the time 60FPS, so the side-by-side areas are probably a flat 60 or close to it.

Unlike BRs or other MPs, you can't just die and spectate another player/team and have a perfect side-by-side so the only way to record similar area would be to play the game with someone else and stick very close to him which in a game like this would be a nightmare. It's just not worth it.
But it doesn't even seem like the same map, or at least the same area. I can understand that it is not a side by side, but at least try to make the situations and scenarios similar. For example, in 12.43 you see a scenario that is having drops to 50fps on PS5, and there is nothing dynamic on the screen (there are no alpha effects, there is no destruction, the framerate drop occurs simply by looking at the map), so difficult is it to go to that area and see how XSX behaves?

I understand that in a multiplayer game it's a pretty heavy lift to do a side-by-side comparison, but I don't think it's that hard to go into that area and see how the Series S/X performs.
 

SomeGit

Member
But it doesn't even seem like the same map, or at least the same area. I can understand that it is not a side by side, but at least try to make the situations and scenarios similar. For example, in 12.43 you see a scenario that is having drops to 50fps on PS5, and there is nothing dynamic on the screen (there are no alpha effects, there is no destruction, the framerate drop occurs simply by looking at the map), so difficult is it to go to that area and see how XSX behaves?

I understand that in a multiplayer game it's a pretty heavy lift to do a side-by-side comparison, but I don't think it's that hard to go into that area and see how the Series S/X performs.
The building in the front in collapsing, top left.
 

Vergil1992

Member
The building in the front in collapsing, top left.
Good observation, but I think the framerate drops start before they start to collapse. But well, I'm not sure. In any case, as always, very even versions. On this occasion small victory for XSX.
 

tr1p1ex

Member
I recently bought an XSX and found out yesterday I could still play this even though my free trial for whatever Xbox service ran out.

Sweet game. Seems very well done.
 
Last edited:
It is getting very simple that Xbox Series require more time to optimize due to Series S. It is a catch 22 phase for developers so either they invest time on optimizating ,therefore, sure they will first of course,
the game 3rd Party PC ported to PS5 consume time and get it best possible way to done due to PS5 has more than 80% market share and than time remaining is done on both Xbox Series S and X. If X was the MS console
this gen than you would see Xbox Series X beating PS5 in majority of titles.
 
No 120FPS in a competitive shooter?

Courtesy of those lovely Zen2 CPUs I’m sure
Like I’ve been saying the biggest travesty of the console isn’t their rt capabilities or the gpus it’s that they didn’t use zen 3 hopefully the pro fixes this
 

twilo99

Gold Member
Like I’ve been saying the biggest travesty of the console isn’t their rt capabilities or the gpus it’s that they didn’t use zen 3 hopefully the pro fixes this

Sad state of affairs but it’s been somewhat working out so far I guess.

The real kicker is that we would still be playing on Zen2 in 2025/26 .. that’s like what, 3.. 4 generations behind?

 
The Series S version also looks and runs pretty damn good.

The video talks about a notable anisotropic filtering difference between the two, but honestly, it looks perfectly fine in stills.

These devs are pretty competent.

COMP-3_3j89A5O.jpg
Doesn't hold back the other 2 consoles though?
 

tommib

Gold Member
I played this for like 1 hour just to check it and I have to say it’s one of the most impressive looking games out there on consoles, specially with that performance.

The destruction physics are incredible and the overall image is very clean. If I was into free to play shooters this is were my immaculate taste would end up sinking hours.
 
I play on PS5 sometimes and honestly it feels pretty damn smooth all the time. If there are drops they are rare and not bad at all. I wouldn't worry about that.
 

Ozriel

M$FT
Like I’ve been saying the biggest travesty of the console isn’t their rt capabilities or the gpus it’s that they didn’t use zen 3 hopefully the pro fixes this

Price and availability in time for a holiday 2020 launch. These things are designed well ahead of time and production begins early enough to have stock.
Waiting for Zen 3 would have ended up backfiring with the semiconductor shortage situation and missing the holiday launch period.

Sad state of affairs but it’s been somewhat working out so far I guess.

The real kicker is that we would still be playing on Zen2 in 2025/26 .. that’s like what, 3.. 4 generations behind?


Yes? That’s every console generation. Fixed specs, for 7-8 years. That’s never been a problem.

Not even sure why there’s so much CPU handwringing for a title that’s hitting 60fps most of the time.
 

Inviusx

Member
Huh, I played this on PS5 for a few hours without realising it was UE5 with RTGI. It looked very last gen to me.
 

twilo99

Gold Member
Yes? That’s every console generation. Fixed specs, for 7-8 years. That’s never been a problem.


Not even sure why there’s so much CPU handwringing for a title that’s hitting 60fps most of the time.

Why only 7-8 years? They can push that to 10-12 years and devs would be forced to still optimize games for whatever fixed hardware the boxes have on offer.

Forcing ancient hardware for so long is not the right thing to do... that's the real "holding back gaming" if you think about it
 

onQ123

Member
It is getting very simple that Xbox Series require more time to optimize due to Series S. It is a catch 22 phase for developers so either they invest time on optimizating ,therefore, sure they will first of course,
the game 3rd Party PC ported to PS5 consume time and get it best possible way to done due to PS5 has more than 80% market share and than time remaining is done on both Xbox Series S and X. If X was the MS console
this gen than you would see Xbox Series X beating PS5 in majority of titles.
Wrong and loud SMH.
 
Sad state of affairs but it’s been somewhat working out so far I guess.

The real kicker is that we would still be playing on Zen2 in 2025/26 .. that’s like what, 3.. 4 generations behind?

If your referring to the pro I just don’t believe the zen 2 leak
 
I didn't realise this was done by ex dice devs. First thing that came to mind when I played this game, was this looks like a crisper version of bad company 2. Was on a snowy map. It looks very nice on the S and X anyway and feels smooth.
I hope they add some more modes to this game, Bad Company 2 Squad death match was fantastic a mode in this game similar to that could be brilliant 4 teams of 4 competing for kills in buildings fully collapsible.
 

adamsapple

Or is it just one of Phil's balls in my throat?
I didn't realise this was done by ex dice devs. First thing that came to mind when I played this game, was this looks like a crisper version of bad company 2. Was on a snowy map. It looks very nice on the S and X anyway and feels smooth.
I hope they add some more modes to this game, Bad Company 2 Squad death match was fantastic a mode in this game similar to that could be brilliant 4 teams of 4 competing for kills in buildings fully collapsible.

Imagine a single player game running on this tech. They seem to have really good grasp of UE5 already, just need to put it in a formal shell.
 
Top Bottom