I think this looks good.

40 minutes ago, steve310002 said:

My theory is tired and dead? Jeez, relax a little.

Personally, I don't see anything but logic in what I said. They are not going to upstage the Xbox version with better clouds in the PC version.

Why your theory is a dead one and why @Krakin is right, is because they are already upstaging the Xbox with various sliders that are absent from XBox, but are only available for the PC version of MSFS. DLSS is another example of a feature that is available for PC, but not for XBox.  With DLSS alone, the PC version of MSFS outshines XBox, especially if your graphics card is weaker on a PC.

With DLSS, and all the other sliders that are available on PC but not XBox, Microsoft/Asobo have made a conscious decision to branch the PC version from the XBox version for graphics, and allow the PC version to get more advanced as new/faster hardware appears for the PC in the coming years.

15 minutes ago, abrams_tank said:

Why your theory is a dead one and why @Krakin is right, is because they are already upstaging the Xbox with various sliders that are absent from XBox, but are only available for the PC version of MSFS. DLSS is another example of a feature that is available for PC, but not for XBox.  With DLSS alone, the PC version of MSFS outshines XBox, especially if your graphics card is weaker on a PC.

It doesn't have to be black and white. MS are looking for a middle ground. Sure, of course they will give PC users extra sliders etc. that we are used to but they will not give us visibly better clouds, which are such a fundamental element to how the sim looks.

DLSS gives better performance, it does not make the sim look better. If you were to compare a maxed out MSFS on PC with the best Xbox version, apart from draw distance and frame rate, you'd struggle to notice the difference (or at least the average person would). Give us better clouds on PC, and hey presto, you have a much better looking sim as it is 'the' major element affecting how the sim looks (combined with lighting of course, the two being integral anyway).

I am not so arrogant or bold to say that my theory is 100% true and foolproof because unless MS or Asobo tells us that, absolutely no one can say for certain what the definite reasons are, but when you look at the evidence and coincidence, it does point to it being the most likely explanation. 

12 hours ago, MrBitstFlyer said:

The clouds in a new X-Plane YouTube video may shed light on what we can expect as far as resolution is concerned in MSFS.  The clouds in this video certainly look softer than what is in MSFS now - with two modern flight sims using a similar technique to produce the clouds maybe we can't hope for a quick increase in resolution.

Well this tells us a bit about how great one of the planes are going to be in X-plane 12. Not sure what it tells us about X-Plane clouds or MSFS ones though.

46 minutes ago, steve310002 said:

It doesn't have to be black and white. MS are looking for a middle ground. Sure, of course they will give PC users extra sliders etc. that we are used to but they will not give us visibly better clouds, which are such a fundamental element to how the sim looks.

DLSS gives better performance, it does not make the sim look better. If you were to compare a maxed out MSFS on PC with the best Xbox version, apart from draw distance and frame rate, you'd struggle to notice the difference (or at least the average person would). Give us better clouds on PC, and hey presto, you have a much better looking sim as it is 'the' major element affecting how the sim looks (combined with lighting of course, the two being integral anyway).

I am not so arrogant or bold to say that my theory is 100% true and foolproof because unless MS or Asobo tells us that, absolutely no one can say for certain what the definite reasons are, but when you look at the evidence and coincidence, it does point to it being the most likely explanation. 

So regarding the clouds, as it has been mentioned time and time again in this thread, we have the same type of clouds that we had on the release of MSFS.  It's just that we don't have them in live weather now.  You can get the same clouds you got at the release of MSFS, if you go and customize a predefined weather for your flight.  The loss of variation of clouds is for live weather; live weather, when it was integrated to work with METAR, lost the ability to show all the different variation of clouds. This is now being added back gradually, and SU 10 is a huge step in adding back the variation of clouds for live weather.

As for better looking clouds, I think we should wait for the variation of the clouds to be added back in live weather first, and then we can push Asobo to make the graphical quality of the clouds better.  To be honest, I think the graphical quality of the clouds is excellent, and it's one of the biggest strengths of MSFS.  It's the variation of clouds lost in live weather after the METAR integration, that is the problem.  

I just need to point you to the graphical quality of the clouds in the latest XP 12 preview (the XP 12 Citation preview from yesterday): https://youtu.be/XyBp78sdJXs?t=82. By far, the graphical quality of the clouds in MSFS (not the variation, mind you), is miles ahead than XP 12.  Personally, I am in no rush to push Asobo to fix the graphical quality of the clouds in MSFS.  I am vocal (and have voted on) for Asobo to add back the variation of clouds that they had though, before the METAR integration with live weather.

PS. For anybody reading this comment, to me, graphical quality of clouds <> variation of clouds.

Edited August 3, 20223 yr by abrams_tank

Firstly, TL;DR: to achieve more detail, etc., in cloud depiction and quality may entail a more complete code rewrite, which isn’t in the best interests of the development team or the lower-spec simmer.

Now for the full (best I can do) summary/thoughts/opinion:

We should probably curb our enthusiasm about future (significant/meaningful) improvements in the likes of cloud detail and depiction. I could have tried to do this without bringing up the Xbox, but it's worth mentioning.

Firstly, MSFS was designed, developed and launched with the highest-end consumer hardware readily available and likely to be purchased at that time. The highest-end home PCs, for those who could afford them, included some variation on the RTX2080 card or AMD equivalent. The 30xx GPUs or i9 CPUs weren't around during the design and development period. The Xbox Series X was effectively on parity with the highest-end home PCs, too, i.e. AMD custom GPU equivalent to a RTX2080 Super or thereabouts. The console's only real limitation was RAM.

Here's the deal with why a console usually has lower RAM than you can put in your own PC: obviously it's cheaper to have less RAM, but it's shared RAM that's usually as fast as GPU VRAM. The console's custom architecture means that having shared system RAM is more efficient and faster, tied with faster component bus connections than your home PC has, and in the case of Xbox a faster drive system than your average home PC. The console is all about moving data faster than your average home PC and therefore relying more on streaming,  therefore not needing so much RAM as your home PC because it doesn't need to hold so much data in memory at any given time. Also, the OS is highly streamlined, it uses specific optimised drivers, etc. -- all things that don't eat up as much RAM as your home PC will in the background. It’s an all-round more efficient piece of hardware.

But of course low RAM works great for your average game. A complex simulator like MSFS is more of a struggle. It's even more reliant on higher RAM (at least, it was at launch).

What Asobo did with the Xbox build (different build; DX12 for example before DX12 was even a glint in the PC version's eyes) was move more over to streaming, as stated by them in an interview somewhere (I can't remember where), including more things streaming from the Azure servers than they originally did in the PC build.

They didn't need to dumb anything down at all on the PC. They just had to get more "stuff" to stream rather than queue in memory on Xbox. Not an easy task, but when they did it they probably decided to employ the same techniques in the PC build to find headroom for future features (features which might not appear on Xbox -- like water wakes, which I believe aren't going to be in the Xbox version, if I read the recent SU10 notes correctly).

And with that said, I’ll go back to the point about hardware during development time and at launch. MSFS was sold with minimum and recommend specs. These are based partly on the home PC consumer market: what hardware do most PC gamers have right now? Minimum was something like an i5 CPU and an older GTX card. With the quality sliders in the sim, obviously the minimum spec is the PC that can pull off running the sim on that low-end hardware with acceptable frame rates and still being a decent experience. The recommended spec was an RTX2080 and an i7 variation (hello Xbox Series X with your parity specs). So that’s the upper limit the simulator was designed for, to get the best overall experience and visual representation with the sliders up or near the top (taking into account your choice to adjust them depending on whether you prefer higher frames over visual fidelity), while keeping the simulator stable under highest slider settings on the best hardware (of the time).

Now, if you buy a RTX3080, which wasn’t around when it was developed (but maybe hit the market just as or after the simulator launched), that’s fine. You’ll see gains with that higher-than-recommended-specs GPU, because it’s more scalable than a CPU. Plug it in and it pretty much runs away with itself, pumping out more frames faster, meaning you can start ramping up your screen resolution and still get stable frame rates. That’s largely what new iterations of GPUs are all about at the moment. But a CPU – which runs everything, instructs the GPU, and does all the arithmetic -- isn’t so “plug and play”.

So, you have headroom in your CPU, yes. You’ll probably have a lot in your i9 or AMD equivalents. But it doesn’t mean the sim will happily skip along to the beat of your new CPU, unfortunately, because the developers have to do some serious, and complex, coding work to properly and fully utilise multi-core architecture. For one, it would be concurrent code instead of sequential. And most game code is still designed on sequential principles. It’s efficient, easier, far easier to debug, friendlier to compilers (I think compilers still haven’t quite caught up with multi-core hardware, and neither have the programming languages themselves; compilers like to rearrange the execution of instructions because “they know better “, for example, which you have to keep in check, and I think they still favour sequential execution and presume it, so you need to get off-the-shelf linking tools or develop them in-house to get everything working properly on multiple cores concurrently).

In multi-core programming, you want to switch to concurrent task execution, and if not done well, you can end up with worse performance than sequential tasks done on a single core, because you’ll get bottlenecks if you’re not careful. So you have to start looking at eliminating data dependencies, which isn’t efficient or in any way easy for a game or simulator (let’s say fancy game) design.

Now, Asobo were on the back foot from the start: Microsoft said welcome aboard, now here’s the old code from the last iteration of Flight Simulator that you must use and, by the way, this is the launch date we’ve set. Go, get it working! And that old code was designed for single cores architecture. Entirely sequential execution. When the Xbox version was gearing up (well, probably since launch of the PC build), Asobo took the time to start rewriting the core code from scratch, so that it did start to make use of multiple cores, and we started seeing the performance benefits when that update launched alongside the Xbox version.

However, to get the kind of detail, resolution, whatever, gains in cloud structures along with everything else the sim is now doing (and everything still on the back burner to add in), it might take another major rewrite of the code that absolutely depends on, let’s say, 8 cores to get that true power. And perhaps an almost complete switch to concurrent rather than sequential code. And that’s where you start leaving behind the older hardware for good. Then the people with their i5 CPUs can no longer run the simulator at all, but they purchased it on good faith they would always be able to, based on the minimum specs at the time of purchase. If that happens, queue major complaints and bad publicity for Microsoft...

So, to gain more, more, more might necessitate a complete (or near) rewrite of the code base, and that is going to be at best really low priority (because it’s hard to program and debug), and at worst they do it and annoy the minimum spec simmers.

That’s where an MSFS 2030 would come in: a complete new base code that now has a minimum spec requirement of, say, i7 and recommend spec of i9, etc., so the consumer can choose whether to invest the money in building a new PC to run it before they put their money down for the software.

Edited August 3, 20223 yr by March Hare

One of the DirectX12 new features is smoke simulation.

https://gamingbolt.com/gameworks-flow-tech-in-directx-12-with-real-time-fire-and-smoke-simulation-showcased-by-nvidia-in-new-video

This could be useful for cloud simulation?

1 hour ago, March Hare said:

Firstly, MSFS was designed, developed and launched with the highest-end consumer hardware readily available and likely to be purchased at that time. The highest-end home PCs, for those who could afford them, included some variation on the RTX2080 card or AMD equivalent. The 30xx GPUs or i9 CPUs weren't around during the design and development period. The Xbox Series X was effectively on parity with the highest-end home PCs, too, i.e. AMD custom GPU equivalent to a RTX2080 Super or thereabouts. The console's only real limitation was RAM.

...

...

That’s where an MSFS 2030 would come in: a complete new base code that now has a minimum spec requirement of, say, i7 and recommend spec of i9, etc., so the consumer can choose whether to invest the money in building a new PC to run it before they put their money down for the software.

Yes. I think MSFS was targeted at older PC hardware from 3 years ago, which matched the XBox hardware of today.  To get a meaningful improvement in the quality of the cloud graphics, a rewrite of code may be necessary.  MSFS 2, in 2030, would make a lot of sense then.  Microsoft gets another huge injection of revenue if they come out with MSFS 2 and make us buy it, and that's where the code rewrites could come in.

Either way, I'm happy with the graphical quality of the clouds in MSFS at the moment, only because the clouds in the competing simulators are that much worse.  I think there are other areas they should focus on, for example, the quality of the photogrammetry (if it's possible to improve the quality of the photogrammetry, that would be a huge boost to the overall graphics of MSFS).

1 hour ago, Kassu62 said:

One of the DirectX12 new features is smoke simulation.

https://gamingbolt.com/gameworks-flow-tech-in-directx-12-with-real-time-fire-and-smoke-simulation-showcased-by-nvidia-in-new-video

This could be useful for cloud simulation?

I was thinking it would probably be too FPS intensive right now, for something as wide-scale as a cloud.

I suspect you might see it first in things like chimney smoke, tire smoke, and wing effects like vortices....

Maybe things like fires, and volcanos as well...?

Though to be honest, such particle effects have been around for years and years through things like Nvidia PhysX. Its just that none of the extant sim creators at the time were interested in being tied to any proprietary technologies. (I did ask, back in the day)

5 hours ago, steve310002 said:

I would say it's kind of obvious.

There are so many differences in MSFS between the PC and XBOX version, that set the PC experience worlds apart from the XBOX experience (VR, controllers being prominent examples) that your rational can't be true. As a dev I have an XBOX. I was not able to complete one single fligth because the handling and visual appearance is so subpar compared to my PCs. Also PC on best settings does look much better than XBOX and performs much better. Therefore they have not dumbed down anything for PC. At best the system architectures between PC and XBOX version had to be aligned causing temporary degradation. In SU10 latest we see again the clouds as before SU5, imho. I mean this thread was filled for pages with negativity because one user had wrong settings on his system. 

There are multiple hard facts, why clouds even are better than originally at release:

- The density setting is working nicely now and allows to create a wide range of visual appearances
- The couds are changing the outer shape all the times (dont know the prof term for that)
 

Edited August 3, 20223 yr by mrueedi

1 hour ago, abrams_tank said:

To get a meaningful improvement in the quality of the cloud graphics, a rewrite of code may be necessary. 

If MSFS is the 10yr project it's purported to be, by Asobo, it would be hard to rationalize basing its architecture on 3y/o moderate hardware and then ignore the pathway to upscale as the years march on.  We are now at not quite 2y out from this 10y service life and by the time years 4, 6, 8 and 10 arrive hardware will have continued its relentless march towards greater capability, including subsequent Xbox releases, and to leave folks with something that looks essentially identical in 2030, as it did in 2020, would do nothing to promote MSFS' continue growth and acceptance over time.

 

2 hours ago, March Hare said:

However, to get the kind of detail, resolution, whatever, gains in cloud structures along with everything else the sim is now doing (and everything still on the back burner to add in), it might take another major rewrite of the code that absolutely depends on, let’s say, 8 cores to get that true power.

Is this not the optimization of DirectX 12 we're starting to see already?  You've mentioned this a few times:  "and everything still on the back burner to add in..."   Is there something coming that is more critically relevant than clouds and weather depiction, that place we spend most of our time flying thru?  What is it you imagine Asobo is saving processing resources to be able to manage?  Already, in 2022, we have major resources untouched in top end hardware, and there are new versions of Xbox coming as early as 2023--which we will still be in Year 3 of a 10yr project. 

49 minutes ago, Noel said:

Is this not the optimization of DirectX 12 we're starting to see already?  You've mentioned this a few times:  "and everything still on the back burner to add in..."   Is there something coming that is more critically relevant than clouds and weather depiction, that place we spend most of our time flying thru?  What is it you imagine Asobo is saving processing resources to be able to manage?  Already, in 2022, we have major resources untouched in top end hardware, and there are new versions of Xbox coming as early as 2023--which we will still be in Year 3 of a 10yr project. 

I don't imagine anything at all. Seb said they created headroom for more features people want, or things they themselves want to do, and that was only a few months ago. They have already started on those features, such as water wakes, better physics... But I don't know what else they have planned; I can only paraphrase what Seb said, and see the plethora of feature requests still coming from the community on the forums. How relevant any new features are compared to what you want the sim to do isn't for me to decide (and also why a developer can never really win).

DX12 can bring potential benefits, sure. They will likely introduce GPU-based ray tracing for RTX cards and you might get better reflections and shadows then. But it's not the magic wand that will fix all, or utilise every bit of the latest hardware. It might not bring clouds much closer to realism. None of it is for me to say or predict.

But I presented a scenario based on logic and experience of the industry: That to achieve the kind of realism some want eventually may require a more full and different base code rewrite than DX12 implementation.

Say, to reach the level of realism you want (and we would all like) from clouds, and still do everything else the code still needs to do (and whatever Asobo and MS may have planned) might demand a minimum of 8 cores processing concurrently -- handling processes and data in a significantly different way from standard practices -- and ensuring the elimination of data dependencies between cores (hypothetically). Well, if that is the case then anyone with an i5 processor who bought the sim can kiss goodbye to it, because an i5 doesn't have 8 cores. Part of my point was, the publisher certainly won't leave the minimum spec customer to eat dust while wealthier people cruise on into the future with this version of Flight Simulator. I mean, at that point it's arguably a different version of FS anyway.

So when they are ready to start truly exploiting the hardware that has appeared later on the market, and have the tools and time to do so, it would make more sense (for the reasons already mentioned in my previous post, including commercially from a business standpoint) to start from scratch with a MSFS 2030.

That's usually how it works in the industry. Technology moves on and software catches up, but to catch up properly you start fresh and that's why you have another version of Flight Simulator or a Red Dead Redemption 2 instead of "simply" patching Red Dead Redemption, and so on...

 

 

Edited August 3, 20223 yr by March Hare

8 hours ago, rayharris108 said:

Well this tells us a bit about how great one of the planes are going to be in X-plane 12. Not sure what it tells us about X-Plane clouds or MSFS ones though.

I saw plenty visibility clouds in that video, not sure how you missed them.

2 hours ago, March Hare said:

Say, to reach the level of realism you want (and we would all like) from clouds

First of all, neither of us knows the extent their currrent rendering model will allow for increased cloud detail.  Second--the only hint I gave of the 'realism I want' is moving what we have today, about 20% closer to the image I posted of a RW cloud.

If we say that today, cloud depiction detail sits at an arbitrary value of 100, do you believe the current rendering model could move cloud rendering detail up to 10% or maybe 20%?  If not, why not?  Certainly the hardware resources are there even today, and they improve steadily including for Xbox.  10yr project--at 1.9y today.  

Capt Piett describes his CPU coasts along at 30% utilization and that is with everything maxed and displayed on a 4K screen.  Let me say that again, 30%.  And that 30% is do a huge amount of processing--all the flight dynamics, pre-rendering, weather/clouds, audio, ATC all the stuff that runs handily while taxing his CPU only 30%.  Bumping up cloud complexity/detail by 20% is going to suddenly shut down the rest of the sim?  I don't think so.  But moreover, what's in it for Microsoft to essentially halt development of everything above ground for the next 8.1 years?  Yes, you said Seb said there were a few more things on the table.   Sure, and there's ample headroom to accommodate this.

Until someone produces a cogent argument for why the current rendering model used can't bump up detail/complexity by 20%, I see zero reason to believe otherwise. 

June 21, 2021:

Work is being done with MeteoBlue to improve live weather. Discrepancies with METAR will be solved by the end of the year, but improvements will continue in 2022 with a roadmap that, according to Neumann, will “blow people away.”

Kind of a bold statement to make considering this was done after SU5 and after METAR integration.