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

Or, they stop posting videos.

If you don't care for this level of minutiae, don't watch. For others, these are the only differences being highlighted in their hobby of comparing different tech and whatever differences, no matter how small, need to be identified as the requirement of that hobby. And yes, it's also ammunition for console warring, but it's not a weapon being manufactured for the purpose; just a 'tool' being misused in laughable West-Side Story face-offs between kids. Instead of chains, screenshots. Instead of pipes, DF quotes.

 

Well-said i think. I dont feel that DF is into any warring territory or trying to feed it either. They all have their favourite platforms ofcourse but their reviews and analysis are almost always with a neutral point of view.

 

And this is not like Digitalfoundry told in video or DF direct than resolution difference is minor this gen. I have much more confidence in Digitalfoundry than other site because if they can do iso comparison between consoles and PC they do it. They are much better than el analista and NXGamer and they don't try to give reason for difference of performance if they don't know or if they don't have any indication coming from dev. Out of DF my favorite one is VGTech because they don't do any analysis out of giving some number, no interpretation or bad comparison with different settings and so on just cold fact...

 

VGTech comparison of Stranger of paradise shows a very noticeable difference between the 2 big consoles because of the almost lack of AO on XSX. Framerate seems about the same though. Slighty higher mean framerate on XSX (+0.11fps), but slighty better frame-time stats on PS5. A wash.

But visually...

PS5:


XSX:


I also noticed a small difference of CBR implementation on the vegetation. PS5 seems to have a bit more details resolved at the cost of more artefacts.

 

Seems XSX is slowly getting the upperhand more and more. What some did predict would happen.

 

Nothing about that game should be used to make any comparisons between the consoles. It's a mess. After all...

 

yeah something not right with this game, res mode min on ps5 1944p and on xsx 1512p, ps5 targeting in this mode 60fps with mean 53.5 and minimal 24fps, xsx targeting 30fps with mean 29.9, min 26fps

 

It's crazy to think how Team Ninja used to be one of the more technically proficient studios out there. Recent times have not been kind.

 

Speaking of which, today's DF Direct Weekly touched on this today, from addressing "Why doesn't QA catch this?", noting the Ascent dev's experience in particular, to also explaining why the 'get all the PSO's cached beforehand dummies!' best-practice is not exactly as simple as it sounds with UE (still should be done to get the majority of them of course).

So yes, DF can indeed highlight things the QA team can miss - because they're not always having the same experience. Also these types of conversations further highlight the massive gulf between this kind of analysis vs a framerate counter video.

 

Lol wut QA teams? Sorry but what exactly am I reading? Does the Ascent dev team being understaffed equate to all dev teams being understaffed? Honestly, no offense but your whole post reads like a bad joke. Without even checking Digital Foundries video releases for the year, I can tell you that they've probably covered less than 1/10th of all the games released on PC, switch, ps5, & Xbox talk less of old gen and mobile. So who's ensuring those games have good performance? The lol wut qa teams you were slagging in your posts.

The fellating of Digital Foundry is beyond comprehension. In your mind, somehow a five man team stands on guard for consumers to defeat the evil of unoptimized games. Let's ignore the fact that there are thousands of hard working individuals in the industry employed for the sole purpose of qa. Somehow, this five man team that doesn't even cover all games is the reason performance has improved gen on gen. Not the improved hardware from manufacturers, not the improved Middleware tools, not the growth of the industry, not the devs who have pride in their work, not the QA teams but, Digital Foundry? Absolutely ridiculous!

 

My apologies for the delayed response. I'll address the text in bold first and the comment on the rest of your post.

1)ELAnalistaDeBits doesn't have to prove anything as far as I'm concerned. They've grown to 500k+ subscribers without a face and they continue to grow. It's clear that the audience enjoys things being presented in a sort of impartial way. Could they replace DF? I'd argue that they could as i expect there to be some overlap between both audiences.

2)Why were they unable to see difference in framerate and resolution? Is it because Digital Foundry didn't show it to them or does it have more to do with proliferation of improved display technology at an affordable price? In my estimation, it's the latter.

3)I'd argue that the driving force behind progress has little to do with DF or consumer demands and more to do with rapidly improving technology and economies of scale.

With regards to improvements in audio, imo, its driven by the low adoption of high quality audio devices. People have shown time and time again that they're okay with average audio so, why should resources be spent developing something that a small segment of the user base will enjoy? I own high quality audio equipment and it's bothers me that it's underutilized but, I also understand why.

Anyway, like I've said earlier, I enjoy Digital Foundry's work but I cannot attribute much of the progress made in the industry to their work. They have a role to play but it's certainly small imo. Digital Foundry has a million subscribers? So, what about the other 150+ million people who play video games on consoles that have never heard of DF or the hundreds of millions on mobile? As far as I'm concerned, the role they play is statistically irrelevant.

 

Series-S power draw is actually poor compared to the other two machines.

 

That's somewhat hard to say without being able to isolate how much power is drawn by each component of the machine. WRT the XBS-S compared to the XBS-X, most of the component power draw will be the same. CPU, system board, NVME drive (different capacities but the power draw for the controller chip which is the main consumer of power for an SSD is the same), etc. There will be a minor difference in power draw for the RAM. The only real major difference will be in the power draw of the GPU.

Depending on what the power draw of the non-GPU portion of the system is, it could be either relatively good or not so good. So, as examples using one point in the video where XBS-S is 86.07 and XBS-X is 194.36 and since I'm lazy I'm just going to call it 86 versus 194.

If the rest of the system (everything except GPU) is using...

• 20 watts then the GPU difference would be 66 versus 174 watts. About 2.64x more for XBS-X.
• 40 watts then the GPU difference would be 46 versus 154 watts. About 3.35x more for XBS-X.
• 60 watts then the GPU difference would be 26 versus 134 watts. About 5.15x more for XBS-X.

That is all, of course, very rough napkin math as the non-GPU power consumption will be slightly higher for XBS-X than XBS-S, but not very much higher except when the optical drive is actually in use.

From benchmarking on PC, we also often see that the most power efficient GPUs in terms of perf/watt aren't the smaller GPUs but the larger GPUs and thus we see something similar here. Basically there's a fixed amount of power consumed by the non-GPU components which pushes overall perf/watt efficiency towards more powerful GPUs where the percentage of non-GPU power consumption becomes a smaller percentage for the whole device. There are obviously exceptions, like if the higher power GPU is being push far beyond it's knee of the power curve while the smaller GPU isn't but even then it's not uncommon for the larger GPU to still have a slight perf/watt lead.

So, I'd say the XBS-S is about where one would expect it to be, but that from a perf/watt perspective of the overall consoles the more powerful console has, as expected, a higher perf/watt.

Regards,
SB

 

You don't say

 

Not really.

 

Oh wow, you definitely got me with that response.......

 

XSX is around 2.2x the power draw of XSS but yet XSX has 3x compute, 60% more RAM, DVD drive, larger NVME drive....

So yea, for me XSS is disappointing from a power consumption point of view.

 

ram , dvd drive and nvmes don't use much power. I also don't see why a larger nvme drive would use more power unless your using transferring larger files

https://www.eurogamer.net/xbox-series-s-review-digitalfoundry

Gears 5 peak power was 82.5w on the series s and 210w on the series x. It used less than half the power. The series x actually used 2.54 times the power

 

Yes, as I noted, the XBS-X has a much better perf/watt than the XBS-S, that's totally to be expected and basically falls in line with what you would expect. So I'm not sure that I'd call that disappointing as exactly what it should be.

If we take a look at GPUs for example.

Graphics Card Power Consumption Tested: Which GPUs Slurp the Most Juice? | Tom's Hardware (tomshardware.com)

The highest perf/watt devices aren't the smaller GPUs. That generally only changes if the larger GPU is pushed greatly beyond the knee of the power curve. You can see this with, for example, the Radeon 6800 compared to Radeon 6800 XT where AMD pushed the power envelope much more with more aggressive clocks on the 6800 XT.

Throw in the rest of the system needed to actually run games and that moves even more towards the larger GPUs.

As explained that's because both XBS-X and XBS-S share similar (like the NVME drives) or identical (like the CPU cores) non-GPU related power costs that you can't just engineer away. And unlike PC GPU comparisons neither are clocked aggressively.

For a consumer that is unaware of these things, sure, it's disappointing that the smaller/less powerful device doesn't consume 1/3 (4 Tflops versus 12 Tflops) of the power of the more powerful console, but it falls exactly in line with where one should expect it to when compared with each other if you consider the console as a whole rather than only looking at the GPU. Thus considering it is still using less than 1/2 as much power as the larger console isn't anything to sneeze at.

I mean a rather perverse different way to look at it is that it has the same CPU and same CPU speed so the power consumption should be the same between the XBS-S and XBS-X, right? So it's incredible that it consumes less than half the power of the XBS-X! I mean if we only look at the GPU differences then it should be OK to only look at the CPU differences, right?

When comparing the XBS-S to the XBS-X, I find it neither impressive nor unimpressive. It's just what would be expected.

Regards,
SB