I’m cautiously optimistic that MSFS 24 will have an improved architecture.
There were developer comments about a graphics project being very promising, so they decided to build MSFS 24 around the new work.
Time will tell.
The multi monitor option in MSFS is still beta and is never been further developed then that. Asobo has also never mentioned that they will do so in the future.
So we can hope for better times, but i think we have to get use to what we have now.
Or just not buy it.
You are right! 
I think that when nothing changes i also will stay with the 2020 version and not buy a new version.
I have no intention of throwing my money down the drain.
1 Like
Unfortunately I’ve had to go back to XP12 as the current setup is useless for me
XP12 is not much better then XP11, because the standard scenery is very bad.
The only thing that’s better is the multi monitor support.
I have XP11 with a lot of enhancements, scenery, airports, lighting, clouds and i won’t buy that again for XP12. Therefore i fly most of the time with MSFS 2020 (triple monitor Nvidia surround) and sometimes with XP11 (in a triple monitor XP configuration).
I beg to differ. The standard scenery in XP12 is greatly improved from XP11.
But of course if can not compete with FS2020 on that part. LR simply doesn’t have the financial resources for it.
I have to agree w @GodAtum when it comes to mult-monitor setups.
It is so sad that I had to go back to 1080 vids as 3840x1080 which I’m actually using in Sim is really destroying Immersion.
It simply doesn’t look real. Everyone easily recognizes that it must be fake/simulation.
So I need to cut of half the screen. 
Till now I’m not optimistic at all that visually flown patterns with head trackers will be somehow closer to real as we’re used to from yp3D an X-Plane.
1 Like
The extreme pincushion distortion on the outer thirds of wide screens is one of the main flaws of this sim. It’s ridiculous to look at runways out of the side window, or looking at planes in closer views from outside. Totally distorted. Everything appears twice as big and out of proportion when on the side, compared to being screen center.
I hope for a better projection like barrel projection. But nothing hints at that anyone at Asobo even cares. None of them developers there seems to be interested in flying planes strangely, since the sim is useless for real world visual approach simulation before turning final as it is.
The problem is augmented by msfs‘ lack of integrated multi monitor support. Which would make ultra wide screens a good idea. But they give the worst visualization, since most of the picture outside of screen center is heavily distorted.
A better projection math would be the most desired feature for me in the sim.
Here a good primer on the problem:
4 Likes
@PlatosCave4408
A good explanation of the maths involved. The only reason (IMO) that this has never been resolved in the 2020 version is because they are using old code and libraries. It feels to me that they have difficulty changing that code with a scalar bar. The only comment that I have heard Jorg commit to on this subject is that it would be “difficult”. If Asobo have not rewritten the code themselves or made serious efforts to adapt the old code, then we will not see this resolved in 2024.
My fear is that they have either gone down the route of adapting the old code to fit or done nothing at all. Asobo has a poor history regarding the adaption of the old code. 2020 has persistently developed its own bugs with every iteration of the game through patches.
I am hoping that we can find out the answers to this at the next event. This is the timeline Jorg has set where Asobo can answer queries in more detail on the sim.
The method described above is a post-process. In that respect it’s no different to the distortion which is already used for VR.
1 Like
Lockheed Martin also started with the FSX code and I didn’t notice any compareable distortion in P3D. And I ran it in the same home cockpit as MSFS.
So I assume it is a resolveable issue.
For me this is the main issue for a go/no-go decision for MSFS 2024.
3 Likes
No, there absolutely is distortion at the edges of the view in P3D, because it uses essentially the same projection that MSFS does, ie one intended for a 16:9 (originally 4:3) display. There are ways to get rid of it in P3D by re-projecting and ‘overscanning’ the viewport in the way described by @PlatosCave4408 above. You can’t do that kind of post-processing work with MSFS because the pipeline is not open to end-user extension in the same way that it is in P3D, so this is an option that would have to be added by Asobo (or they’d have to open the pipeline to extension, which IMHO would be better).
I don’t believe this is the case. AIUI the rendering engine from FSX was replaced with one derived from Asobo’s previous game titles. This will certainly be newer than the FSX-era code. It’s got lots of graphical capabilities that the FSX engine does not have, but it’s also more limited than the FSX engine in the sense that it doesn’t allow things that the older engine does with multi-view, multi-frustum etc.
The whole issue of ‘old’ code is a bit of a red herring anyway IMHO. Old code works exactly the same as it did when it was ‘young’. It’s the context in which the code runs that changes. We haven’t magically gotten 10x better at writing code since the 1980s; if you sat down to solve the same problems that FSX solved today, you would likely write code that looks very similar. The issue is that when a codebase gets large, it becomes highly interconnected, so when you poke it here something unexpected happens over there, because humans are not good at scale. This raises the complexity, and hence cost, of change. Good code is designed so that this is less of a problem by separating concerns and designing for maintainability, but I’ve seen the code for a lot of commercial products and I can say with confidence that it’s usually a patchwork quilt of hacks written by mediocre programmers (ie most of them, myself included). That’s as true today as it was 40 years ago.
The reason this issue has never been resolved is that Microsoft doesn’t consider it a big enough issue to dedicate the time and money to resolving it.
Have they addressed it in 2024? I’d be very surprised. Let’s see. I voted for the topic so I’d be pleased if they have done anything to improve the display system for users going beyond one 16:9 monitor.
1 Like
AIUI, MSFS2020 uses code from the original flight simulator, which was originally released in 2006. And from the discussions on the dev streams, I think they (Asobo) inherited the project when Flight Simulator was released for steam in 2014.
As for coding being written the same today as it was in the past, I Disagree. You write code with the hardware that is used by your target audience today. The amount of CPUs we use today has increased massively since AMD introduced the Ryzen family. Asobo have often stated that they want to make more use of the cores we have available.
A 1080 Nvidia GPU had 2560 cores, whereas the 4080 has near four fold with 9728. RTX coding is written into the game now also. All of which is mostly ran by driver API, I know. But then drivers are being constantly updated to make games run when they are not.
Add in that Windows has changed a huge amount since Flight Simulator was released originally in 2006. We had DX11, and now DX12. JS used by devs for the MSFS API is being updated all the time. All of this is the reason why a lot of old games do not run on a modern system.
Note that reprojection from standard rectangular rendering to a cylindrical or other type of projection is just a post-processing effect that can be done on the GPU. The existing MSFS2020 codebase already implements an optional “lens distortion” filter, and they could implement others if they cared with little or no effect from “old code” in the flight simulation engine, which wouldn’t even be touched.
This is simply a niche issue that very few paying customers care about enough to affect a purchasing decision, so it stays at the bottom of the todo list with all the other general simulation issues. 
2 Likes
I don’t think it is as niche as you think. Asobo decided the issue was fixed and took it away from the forums. People just got bored with adding their opinions.
As for whether they can or can’t, Jorg has only ever said that it would be difficult to implement. From my knowledge, he never explained why that was the case.
All of which is moot at this point. Either Asobo have implemented a proper fix or they have not. Even if every person on this forum decided this was the feature that mattered, it will not change the wheels that are already in motion for a November launch.
2 Likes
Yeah, I’d probably bet money they haven’t done it even though they probably ought to. This is the nature of the beast; reasonable things that would be useful for real customers often don’t get done.
When companies have an obligation to shareholders, this is very true. People often forget that game developers are not the fanboys that we had of old. They are in it to make money.
You have to give some credit to Asobo though, they have provided a lot more than most game companies do. Yes, there have been many bugs and the game has looked like nothing more than a firefight, but the amount of content they have added from the original purchase is vast. It really is a case of never mind the quality feel the width.
People have accepted that because Asobo have consistently claimed it was down to old code and libraries. 2024 will most definitely receive a different response. They cannot claim it is old code anymore. All those restrictions are gone.
If Asobo has not added widescreen support, or multi-screen features that are better than the current iteration, then I think this failing will be widely commented on, or at the very least be something that will gain traction very soon after.
Mordor bears this out in their assessment of the monitor market overall.
What are the key trends influencing the Computer Monitor Market?
Key trends influencing the Computer Monitor Market are a) Rising adoption of ultra-wide and curved monitors b) Shift towards higher resolution displays c) Integration of touch screen technology
They also state that the monitor market is expect to grow by 25%+ by 2029. Screen tech is becoming a lot cheaper to produce. Larger faster screens are going to be areas that the world will want to buy into, and as speed is irrelevant for a flight sim there will be only one side of that argument that will matter to sim players.
1 Like
Yes, there is FSX code in the product, and yes, the codebase they started from was the FSX Steam Edition code. The Asobo team has said as much. We don’t know how much (probably quite a bit), but we do know that some of it was replaced / changed. For example, the new flight model will be all new code whereas the legacy flight model probably uses the FSX code.
The rendering engine, specifically, was replaced with one derived from Asobo’s previous work (I think originally from The Crew 2), which had been used for a project they were doing for Microsoft’s HoloLens that convinced Microsoft to green-light the Flight Simulator project. This has been talked about in previous streams / posts by the team. That should mean that none of the code responsible for drawing to the screen or rendering the scene to the GPU is original ‘old’ FSX code, though of course I can’t say that for certain since I’ve never seen either codebase.
I take your point, but that doesn’t mean that older software is automatically less performant on modern hardware than it could be. DirectX is a specific case where to take advantage of newer GPU features you do need to change the code, yes, but DX9 code still runs fine on modern GPUs and much faster than it would have gone on a GTX 780, for example.
On core use and threading, Intel (and AMD) CPUs have had hardware-scheduled threads since the 386 days. You’ve always been able to write software to be parallelised across multiple threads. Yes, in the pre-multicore days that meant time-sharing with everything else, but that was the case for all software. I recall back when FSX Service Pack 2 came out, the ACES team talking about moving texture loading onto separate threads which improved performance by reducing the wait time for the main thread. That these threads were all sharing the same core didn’t prevent there being a performance increase from that.
I’ve written a LOT about the multi-core thing on other posts, so I won’t do it again, but concurrent programming is hard to do, whether the code was written 20 years ago or yesterday.
MSFS2020, even with the old coding, works fine on a 16:9 display. The coding used in MSFS2020 will always work perfectly well on a 16:9 display. I think we both agree the coding used in MSFS2020 is not of a modern standard. We are both of the opinion that the coding used may be older than a decade. IMO, the lack of flexibility in the old code is the issue with compatibility of the ever changing ratios around today.