SlimySnake
Flashless at the Golden Globes
the windows editor caused a lot of bugs so i had to settle for the nvidia control panel limiter.
-
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: Star Wars Jedi Survivor - DF Tech Review - PS5 vs Xbox Series X/S - Ambitious But Compromised
yamaci17
Member
if games are going to be this much cpu bound and hit 800p 1080p on consoles to barely get 30 40 FPS, I'd say they should keep building games on rasterization and add ray traced effects as a bonus. but problem is maybe ray tracing helps them save huge amount of money. in that case we'd be getting a kind of okay but borked experience at the expense of their relaxation opportunity
having ray tracing is good and all and will always look better than rasterization but sacrifices are too harsh and kind of diminisihes the point of ray tracing due to the poor experience ( at least for consoles )
on PC you still get stutters but midrange computers also kind of get erased. even if you make ray tracing standard for your game, a huge bulk of people do not even want to enable ray tracing on ray trace capable hardware (2060 super to 6700xt to 3070 3060 and beyond). then most of these folks paired their gpus with the likes of 3600 5600x and so on. even 5600x is largely problematic in the game. it is just alineating a big userbase
as you said cone reflection errors etc. will make me / force me to play this game at freaking 1080p. some places work okay with 8 gb vram at 4k but some others... simply don't. only at 1080p i get a resemblance of frame stability due to vram with ray tracing. and its too big of a sacrifice. I was like you with Hogwarts, I didn't give a care about if i couldnt enable ray tracing in that title because graphics looked gorgeous with raster and I opted to get stable framerates and kind of high resolution (4k dlss balanced). now i cannot even go to 1440p if i want to play with ray tracing. and of course.. the stutters
but yeah game is really great. no loading screens seamless huge open world experience with great visuals etc...
DenchDeckard
Moderated wildly
This game looks fuxking incredible. I don't think the vid does it justice.
Also that zoomed in 300 percent shot of the opening is nothing like that xbox pic for me. It looks identical to the ps5 version for me.
Also that zoomed in 300 percent shot of the opening is nothing like that xbox pic for me. It looks identical to the ps5 version for me.
yamaci17
Member
by the way check your dedicated gpu memory for game and see how much vram it uses (task manager, not afterburner). there's still possibility you're vram bound
reason:
the game, at max, will allocate 7-7.1 gb "dedicated" gpu memory to itself on 8 gb. and out of that 6.8-7.2 gb, the afterburner per game usage usually shows 6.3-6.4 gb vram (never goes above this value, like, never). in practicality, I'm VRAM bottlenecked at 6.4 gb per app vram usage on 8 GB. (%80 rule, similar to how spiderman "behaved"). for 10 gigs, this is 8 gb (and with some extra dwm usage, yours might be 7.5 gb) I will be VRAM bottlencked but per game vram usage will never go above 6.3 6.4 gb. this is super super similar how hogwarts + spiderman behaved.
here's what i mean
actually you can verify this. try to make your dedicated per game vram usage go above 7.5 or 8 gb. it won't. game has a hard %80 set fixed rule. it simply won't use more than that. total vram usage can be very misleading also (due to background clutter).
in task manager, dedicated gpu memory never goes above 7-7.1 gb for me. so that's the max game can "allocate". and out of that allocation, game will never go past 6.3-6.4 GB VRAM.
I've observed this weird mad rule with spiderman, reported to DF and nixxes, and they upped the allocation + usage percentages a bit and that helped.
I think there's simply a big miscommunication / problem / stuff that even devs are not aware. this is just resources left unused for no good reason at all.
so yeah, you're most likely vram bottlenecked while seeing 7.5-7.7 GB VRAM usage. becase by my calculations, that is where you would usually land on game's logic of using available VRAM
(spiderman also VRAM bottlenecked around 8 GB on 10 GB cards and around 6.4 GB on 8 GB cards).
it is time devs address this problem. them cards already have super duper limited resources as is. by putting such "relax" allocation percentages, they're actively hurting us for no good reason
this stupid rule, behaviour whatever it is, also misleads people as you think your problems have nothing to do with VRAM, whereas you do. you just see empty vram space and think the game is tame in vram consumption. it is super super misleading.
even if this behaviour is required for game operation, then I deem all 12 10 gb cards also obsolete. this means for this game, your 12 gb gpu is simply a 10 gb gpu. this is just too much limitation.
Last edited:
saintjules
Member
More DF coverage/impressions for about 35 minutes (timestamped).
Last edited:
DenchDeckard
Moderated wildly
I guess this is the grate every one was getting bent up over? I took a screen as soon as I got off the ship.
Looks the same as ps5 to me?
Looks the same as ps5 to me?
Mr Moose
Member
Was this with the new patch?I guess this is the grate every one was getting bent up over? I took a screen as soon as I got off the ship.
Looks the same as ps5 to me?
Edit: Looks like the new patch came out a few hours ago and fixed it. And if the performance has increased on all now... The video is even more pointless than it was the other day. Was hardly a "VS" to begin with,
Last edited:
DeepEnigma
Gold Member
Textures still look lower res a bit.I guess this is the grate every one was getting bent up over? I took a screen as soon as I got off the ship.
Looks the same as ps5 to me?
DenchDeckard
Moderated wildly
Mr Moose
Member
Edited my post, but yeah seems the new patch has fixed that issue. Time for another DF video lol. I wonder if they will actually do a proper VS this time instead of this shit.
PlayStation 5 and Xbox Series X/S (arriving 2nd May, aleady on PC)
- Multiple crashes fixed across PlayStation and Xbox Series X/S and various areas of the game.
- Fixed crashes that were tied to skipping cinematics.
- Performance improvements across PlayStation 5 and Xbox Series XIS.
- Fixed an issue with dynamic cloth inside the Mantis.
- Fixed various rendering issues.
- Fixed an issue with registered Nekko colours not saving.
- Fixed an issue with registered Nekko disappearing from the stable.
- Fixed issues with cinematic dialogue overlapping.
- Fixed various collision issues.
- Fixed an issue with enemy Al remaining in T Pose during photo mode.
- Fixed a freeze that occasionally occurred while talking to Doma.
- Fixed a bug where the BD-oil VFX did not properly render.
- Fixed an issue where players were getting stuck inside the Chamber of Duality if you didn't save after leaving the chamber and die.
Last edited:
DenchDeckard
Moderated wildly
Yeah looks pretty damn amazing on series x so hopefully it just gets better.
DenchDeckard
Moderated wildly
If I was digital foundry I'd lose my mind. You do a video and then 3 days later it's null and void.
Or did he simply not suffer the same LOD load or did he capture after loading all the LOD?
Whatever it is, I agree, a more in-depth analysis would be interesting to find out how much performance has improved on the different platforms... although I doubt that miracles have been achieved with a single patch days after launch
adamsapple
Or is it just one of Phil's balls in my throat?
Mr Moose
Member
Different videos showed the same thing, so now it seems it's fixed with the latest patch.Or did he simply not suffer the same LOD load or did he capture after loading all the LOD?
Whatever it is, I agree, a more in-depth analysis would be interesting to find out how much performance has improved on the different platforms... although I doubt that miracles have been achieved with a single patch days after launch
Yup, one that shows a bit more than this one hopefully. This one was shite.
DenchDeckard
Moderated wildly
Or did he simply not suffer the same LOD load or did he capture after loading all the LOD?
Whatever it is, I agree, a more in-depth analysis would be interesting to find out how much performance has improved on the different platforms... although I doubt that miracles have been achieved with a single patch days after launch
Maybe there is a way for me to capture a video clip to show you it looks like that at all times no lod loading.
There is no need
It is simply known (there are videos even of the PS5 version) that sometimes the LOD load was late and it was not unreasonable to think of that possibility in that place. Even more so when in others places these differences did not occur. PS . From what I'm reading, it seems that the patch hasn't improved performance much. At least not the gaming experience.
onnextflix5
Member
Funny thing is everyone predicted this, after the first game, yeh they didn’t disappoint lol. Ue4 must be double edged sword, easy to work with, hard to optimise.
Kataploom
Gold Member
Should this work if drivers are just partitions in the same nvme? I don't have the game but it should help in other UE4 games
winjer
Gold Member
No.
DenchDeckard
Moderated wildly
It's amazing that they can get a patch out within 3 days that fixes a lot of the issues. Need to get performance sorted ASAP.
What do you know....turns out it wasn't the xbox hardware that was the problem once again...
What do you know....turns out it wasn't the xbox hardware that was the problem once again...
winjer
Gold Member
If they had delayed the game a couple of weeks, it would probably be enough to sort out most of these performance issues, and not give the game such a bad from day one.
At a time when most of the marketing and distribution is digital, these things are much easier to do.
It's like these companies are run by morons.
At a time when most of the marketing and distribution is digital, these things are much easier to do.
It's like these companies are run by morons.
saintjules
Member
As per DF, they believe it would need a few months worth of patching to get the game to an acceptable level of performance.
I might cave in and buy it since Amazon dropped the price by $10.
winjer
Gold Member
Get a EA Play Pro subscription for one month. You get Jedi Survivor and a bunch of other games for 15$.
Tsaki
Member
Yeah what the hell? I was watching the video and was like "Bro it really is not that different". Such a drastic drop in resolution and fps for miniscule differences.All that laughing at the PS5 version in the other thread, and DF confirms that XSX performs the same with the same drops in heavy areas and similar resolutions.....
Let warring die for one goddamn game people and come together to call out the industry who is happy pitting the most passionate fanbases against each other while they ship last gen trash for $70. This is just a very poorly optimized game. The game fancies the highest end CPUs like the new 7800x3d and 6.3 Ghz Intel CPUs that run the game flawlessly even on 16 tflops 6800 and 17 tflops 3070. Clearly, the game favors higher clocks and doesnt do a good job multithreading.
The RT comparisons are infuriating. They barely make a difference even when zooming in and DF acts like they are comparing a victoria secret model to honey booboo. The non-RT mode even does a great job illuminating the halo around bright neon lights. In the open world, there is literally no difference in the GI. Turn that shit off. RT reflections use SSR anyway in most spots. This is so stupid. These guys are completely detached from reality.
SlimySnake
Flashless at the Golden Globes
DF is too worried about attracting fanboys and controversy to actually do their job.
The Anal de bits guy has been doing some amazing work with way better head to head comparisons of ALL modes, but people here got him banned because he got some pixel counts wrong. Something DF, NX Gamer and well everyone is guilty of.
We need to talk to the mod team about allowing him again because his comparisons are leagues ahead of everyone else out there. Pixel counts are just one aspect of the comparisons. The framerate comparisons, and other details that we can see with our eyes do still count and should be allowed. He goes out of his way to find the worst performing areas in games like RE4 and Hogwarts instead of just the first mission which might not always be the worst performing level.
Maybe an asterisks can be added whenever his comparisons are posted to ignore the pixel counts. He already has a very good analysis of this patch out while everyone else is tweeting about having to do more work.
Lysandros
Member
He was rightfully banned by the mods because he was caught using faulty comparison tools/methods and manipulating data more than once to favor one specific platform (being the XSX).
He still continues to be proven wrong constantly by reliable outlets such as Vgtech, NXgamer and Digital-foundry. But he is surely the best place to go to see XSX performing the best and at higher resolutions nearly every single time, i'll give you that. Even when he doesn't have anything to back it up he he will throw something like "XSX has higher resolutions in all modes" like in the recent case of Resident Evil 4 again contrary to all other reputable outlet's findings. The thing he is truly the best at is creating reality bubles. You may still personally think that he is the best and outstanding, that is you. I 100% support the mods decision on the matter.
Last edited:
SlimySnake
Flashless at the Golden Globes
Like i said, ignore or ban his pixel counts. But his framerate comparisons speak for themselves and he chose RE4 areas that showed the xbox dropping down into the mid 30s while the PS5 remained in the 50s. No one else ran comparisons in the village rain sections.
DenchDeckard
Moderated wildly
But nxgamer has consistently got resolutions wrong and made wild claims to why something is performing a certain way...the game then gets patched and he looks silly. He also paints the ps5 in a strong light in every opportunity and we know he got his entire ps5 press pack for free including the console etc. Something no other manufacturer did for him....
....so I think it's OK to say he shows some bias too. Not sure if he is banned here or not?
01011001
Banned
he used a tool in the past that is known to not work correctly.
it's not clear if he still does, but almost everyone that isn't VG Tech, DF and NXG use this tool (TRDrop)
said tool is especially unreliable as soon as a game exhibits tearing, and Jedi Survivor literally tears every time it drops below 60fps (or 30 in resolution mode)
Last edited:
01011001
Banned
I don't think he gets resolutions wrong all that often, not more than anyone else when a game has a highly dynamic res scaler or ridiculously muddy TAA/reconstruction.
what he does do tho is defend bad PC ports like Spider-Man, or say really weird shit like how he thinks the combat and animations in Star Wars are top tier... which is... like... wow
the one thing I personally will never forget tho is his claim that RE3 on current gen doesn't use Checkerboard rendering... and I even argued with him on this site about it, making screenshots and using matching screenshots from the last gen comparisons to showcase that both version exhibit the exact same artifacts... and last gen was confined checkerboarding.
and this after literally every RE Engine title used Checkerboarding before that as well... like... why would they suddenly switch to a different reconstitution method that then also has the same artifacts?
his "proof" for that was that the Interlacing mode on PC looked different to the console reconstruction, completely ignoring that that was true in the past versions and games as well.
...also suddenly we're back to Checkerboarding again in the RE4 remake
makes sense of courseso he has is super weird takes sometimes for sure.
Last edited:
Says this who continually talks about DF's bias in favor of XBOX and its low credibility, who sets NXG as an example of credibility and impartiality.....
The reality is that the results of "elanalista" have sometimes been wrong in favor of PS5 as well, but in those cases you were not there to discuss its credibility. What's more, here sometimes there are people who use screenshots of their analysis as a means of proof even despite being a banned source. Surprisingly we don't see you go out and censor their use when they are used to confirm advantages of the PS5 version...
In my opinion, no source should be banned. That it has continuous errors does not mean that its analyzes always lie or are not useful to be able to confirm or not non-concrete points of other analyses.
But the decision to ban or not is not mine and I have to respect it.
LOL. good invented That is false. WTF
In facts, the mistake in RE4 was that he did not mention the clear IQ advantage of the XSX version compared to the PS5 version. He said resolution was The same on both
But hey, it's normal for you. Complain when you don't like what a comparison says but not When You like.
Last edited:
Thirty7ven
Banned
Think DF should keep doing this “generally the same” type of communication unless differences between versions are very significant. But they have to stay consistent on that…
Mr Moose
Member
It's his fps readings that have been very wrong in the past, multiple times.
His pixel counts are probably better than Tom and Tom 2.0s though lol, both of them (Tom and Tom 2.0) miss details that anyone could see too it seems and they normally show it right after saying there's no difference which I always find funny. I would only ever use his videos for the side by sides but not fps.
As
01011001
said, he uses trdrop, looks like he is using the latest version now but still... Wouldn't trust those readings.
SlimySnake
Flashless at the Golden Globes
It shows 8.3 GB in the task manager and 7.5 GB in MSI afterburner. weird.by the way check your dedicated gpu memory for game and see how much vram it uses (task manager, not afterburner). there's still possibility you're vram bound
reason:
the game, at max, will allocate 7-7.1 gb "dedicated" gpu memory to itself on 8 gb. and out of that 6.8-7.2 gb, the afterburner per game usage usually shows 6.3-6.4 gb vram (never goes above this value, like, never). in practicality, I'm VRAM bottlenecked at 6.4 gb per app vram usage on 8 GB. (%80 rule, similar to how spiderman "behaved"). for 10 gigs, this is 8 gb (and with some extra dwm usage, yours might be 7.5 gb) I will be VRAM bottlencked but per game vram usage will never go above 6.3 6.4 gb. this is super super similar how hogwarts + spiderman behaved.
here's what i mean
actually you can verify this. try to make your dedicated per game vram usage go above 7.5 or 8 gb. it won't. game has a hard %80 set fixed rule. it simply won't use more than that. total vram usage can be very misleading also (due to background clutter).
in task manager, dedicated gpu memory never goes above 7-7.1 gb for me. so that's the max game can "allocate". and out of that allocation, game will never go past 6.3-6.4 GB VRAM.
I've observed this weird mad rule with spiderman, reported to DF and nixxes, and they upped the allocation + usage percentages a bit and that helped.
I think there's simply a big miscommunication / problem / stuff that even devs are not aware. this is just resources left unused for no good reason at all.
so yeah, you're most likely vram bottlenecked while seeing 7.5-7.7 GB VRAM usage. becase by my calculations, that is where you would usually land on game's logic of using available VRAM
(spiderman also VRAM bottlenecked around 8 GB on 10 GB cards and around 6.4 GB on 8 GB cards).
it is time devs address this problem. them cards already have super duper limited resources as is. by putting such "relax" allocation percentages, they're actively hurting us for no good reason
this stupid rule, behaviour whatever it is, also misleads people as you think your problems have nothing to do with VRAM, whereas you do. you just see empty vram space and think the game is tame in vram consumption. it is super super misleading.
even if this behaviour is required for game operation, then I deem all 12 10 gb cards also obsolete. this means for this game, your 12 gb gpu is simply a 10 gb gpu. this is just too much limitation.
I did turn on resizable bar so going to test that out.
Also confirmed that resolution scaling goes from 100 to 87 in the windows editor file once i switch it from epic to high. But i swear it looked exceptionally muddy today even at Epic. I checked and it showed 100 resolution scaling so I am guessing the game is bugged. Looks better in FSR Quality mode.
P.S DWM is fucking 800 megabytes while the game is running. EA app is 200 MB. wtf is this nonsense.
Lysandros
Member
I personally wouldn't thrust him with anything out of principle. Once a fraud always a fraud.It's his fps readings that have been very wrong in the past, multiple times.
His pixel counts are probably better than Tom and Tom 2.0s though lol, both of them (Tom and Tom 2.0) miss details that anyone could see too it seems and they normally show it right after saying there's no difference which I always find funny. I would only ever use his videos for the side by sides but not fps.
As
Last edited:
Iamgoat100
Neo Member
We have been told (or fed) that due to newer upsampling techniques that internal resolution will no longer matter. Games can run at a lower internal resolution and have little to no perceptual difference to the output resolution. By doing this, the devs are saving resources, and games should perform better. Why is this not happening? A game that is dropping down to sub 720p internal resolution should be able to hold a solid 60.
I am speaking strictly from a console perspective. I know their are difference between FSR and DLSS and other upsampling techniques.
It's the same with Variable Rate Shading. We shouldn't perceive much of, if any difference, and games should perform better due to saving resources. We all remember the Dead Space debacle. The PS5 VRS was broken, and I believe the devs then disabled it on both PS5 and Series X, with no real hit to performance, if memory serves me correct. So what was the point of using it in the first place?
So are devs just not utilizing these techniques properly? Are they really not capable of achieving what they claim? Is it all buzzwords and pipe dreams?
You also have Xbox that doesn't even seem to be using a lot of the "Next Gen" features they have incorporated in their machines to improve graphics and performance. I would think this most likely comes down to Xbox not putting out a lot of big AAA exclusives and third parties not bothering to incorporate these features. I guess we will have to wait and see what the Coalition has in store, and Ninja Theory with Hellblade 2 for that matter.
I am just disappointed that this many years into this console generation it doesn't feel like we have seen these supposed performance saving features and techniques put to good use. Maybe it's just another case of hype, overpromissing and under-delevering, or maybe they will come in time... who knows...
I am speaking strictly from a console perspective. I know their are difference between FSR and DLSS and other upsampling techniques.
It's the same with Variable Rate Shading. We shouldn't perceive much of, if any difference, and games should perform better due to saving resources. We all remember the Dead Space debacle. The PS5 VRS was broken, and I believe the devs then disabled it on both PS5 and Series X, with no real hit to performance, if memory serves me correct. So what was the point of using it in the first place?
So are devs just not utilizing these techniques properly? Are they really not capable of achieving what they claim? Is it all buzzwords and pipe dreams?
You also have Xbox that doesn't even seem to be using a lot of the "Next Gen" features they have incorporated in their machines to improve graphics and performance. I would think this most likely comes down to Xbox not putting out a lot of big AAA exclusives and third parties not bothering to incorporate these features. I guess we will have to wait and see what the Coalition has in store, and Ninja Theory with Hellblade 2 for that matter.
I am just disappointed that this many years into this console generation it doesn't feel like we have seen these supposed performance saving features and techniques put to good use. Maybe it's just another case of hype, overpromissing and under-delevering, or maybe they will come in time... who knows...
ChiefDada
Gold Member
This is becoming the most annoying misconception being repeated over and over again. There is much more to gpu performance than resolution. All pixels aren't created equal.
Iamgoat100
Neo Member
I get that. I know there is more to it. But having to lower the internal resolution to that extent is to help keep a performance target. Are devs being too ambitious in other areas and not able to hit the target on current hardware? Is it just poor optimization on many releases? Is it the need to push it out the door to meet a deadline? Do these performance saving features actually work in real game scenarios?
My rambling on was more about performance saving features that we have been told about that are not quite hitting the mark right now and asking why.
To me, it seems like the gaming industry is so secretive about certain things. Why is game "A" performing better on PS5? Why does game "B" have a higher resolution on Series X? These are just generic examples. We can have our theories when certain things occur and speculate, but it would be great for devs to do a sort of deep dive and explain exactly why from a hardware standpoint, such things take place.
I don't know if there are NDA's involved, or they just don't want to paint the less performant console in certain instances in a lesser light to keep in good grace. It would just be nice to get some concrete transparency.
Tsaki
Member
This. I still haven't seen a satisfying answer to your questions. Software Lumen (which runs on shaders) is still Global Illumination, yes? And it changes dynamically with it's environment, correct? It has a certain resolution/quality to it, keep it the same but run it on the hardware acceleration that is created specifically for this kind of work. Why the hardware accelerators wouldn't complete the same load faster than the shaders is not clear to me.
Tell me you have not played Cyberpunk overdrive without telling me you have not played Cyberpunk overdrive