5 hours ago, Noel said:

I'm pretty sure of the 250 developers at least some are hyperaware of just how much most users want to see a substantial improvement in the quality of cloud types and weather depiction.  They get it from visiting forums it's impossible to to miss.  I think it's mostly a question of how much more they can get out of the current rendering engine w/o breaking the performance bank.  Is it impossible to improve 3D pixel density to help facilitate making more distinct cloud morphologies a reality for those w/ the hardware to handle it?  The only thing I can think of is that voxel density is locked such that there is no way to make it more dense for those w/ better hardware w/o putting that demand on their minimum standard hardware, including Xbox.  This may be why we see such little progress toward that end.  I had proposed something like DynamicLOD for clouds, such that one can dial in improved resolution at close range, diminishing the farther away clouds are as a means to improving the appearance of clouds at close range where they currently are very poor, low resolution, blurry to some degree.  Perhaps we'll see a serious jump in 2028 when next gen Xbox debuts which MS is touting apparently as the "largest technological leap ever seen in a hardware generation..."

 

 "The only thing I can think of is that voxel density is locked such that there is no way to make it more dense for those w/ better hardware w/o putting that demand on their minimum standard hardware, including Xbox.  This may be why we see such little progress toward that end"

You now know the whole truth. Everything must fit within the vertical rectangle.

On 6/12/2024 at 4:15 PM, HiFlyer said:

I'm convinced many people would even find real life weather unrealistic.....

You’re not wrong in my case...the other night I stepped out of work, glanced up at the sky, and was amazed at how absolutely fake everything looked.

The clouds were wrong, the colors were wrong, it all looked like a very garish, very fake combination of the worst of what I’ve seen in MSFS and the latest XP12 beta, with a bit of “The Simpsons” opening credits mixed with a Bob Ross PBS painting show.

I couldn’t believe that in the time since FS98 this was the improvement.

And (again) this was the REAL, actual sky!?!

It was the real time prototype...yet the entire combo of various clouds and colors looked VERY silly and fake.

Having just read a disturbing article about beauty pageants featuring very realistic (but not real) contestants, I’m convinced we will get there soon!

Both current sims often do a fantastic job of re-creating the current local sky...and (sometimes) the actual sky looks real too.

 

 

 

 

 

 

NVM

Edited June 14, 20242 yr by JBDB-MD80

1 hour ago, filou said:

 "The only thing I can think of is that voxel density is locked such that there is no way to make it more dense for those w/ better hardware w/o putting that demand on their minimum standard hardware, including Xbox.  This may be why we see such little progress toward that end"

You now know the whole truth. Everything must fit within the vertical rectangle.

I've always suspected this is the issue that there can only be one voxel grid. Do the other sims sporting volumetric clouds default to a higher resolution grid?  When looking at XP12 trailers it struck me that it was even worse, lower resolution.

 

ChatGPT3's take:

Creating volumetric clouds using voxels is a computationally intensive task due to the following performance demands:

1. Memory Usage

• Voxel Grid Storage: A high-resolution voxel grid requires a significant amount of memory. Each voxel represents a small volume of space, and high-resolution clouds necessitate many voxels.
• Density and Color Information: Each voxel often stores multiple attributes such as density, color, and potentially other physical properties (e.g., humidity, temperature), increasing memory requirements.
• 3D Texture Storage: When using 3D textures to represent clouds, the memory footprint can be substantial, especially if multiple layers or levels of detail are used.

2. Computational Load

• Voxel Processing: Generating, updating, and manipulating voxels requires substantial computational resources. This includes operations such as density calculations, noise function evaluations (e.g., Perlin noise), and procedural generation techniques.
• Rendering: Ray marching or ray tracing through voxel grids is computationally expensive. This involves calculating light interactions, shadows, and scattering effects at many points along each ray.
• Shading and Lighting: Realistic cloud rendering requires complex shading models that account for light scattering, absorption, and multiple scattering events. These calculations are intensive, especially for high-quality rendering.

3. Real-time Constraints

• Frame Rate: For real-time applications, such as video games or interactive simulations, maintaining a high frame rate is critical. The need to update and render clouds in real-time adds pressure on both the CPU and GPU.
• Level of Detail (LOD): Implementing LOD techniques to dynamically adjust the resolution of the voxel grid based on the camera's distance and angle can mitigate performance issues but requires additional computational overhead for LOD management.

4. Parallel Processing

• GPU Utilization: Utilizing the GPU for voxel cloud generation and rendering can alleviate some of the computational burden on the CPU. However, effective parallelization of tasks such as noise generation, voxel updates, and ray marching is crucial for performance.
• Multi-threading: On the CPU side, multi-threading can help distribute the load of voxel manipulation and other computations, but this requires careful management to avoid bottlenecks and ensure efficient use of resources.

5. Optimization Techniques

• Sparse Voxel Grids: Using sparse voxel representations, where only non-empty voxels are stored, can significantly reduce memory usage and processing time.
• Caching and Reuse: Caching intermediate results (e.g., noise function evaluations) and reusing them can reduce redundant calculations.
• Temporal Coherence: Exploiting temporal coherence by reusing previous frames' calculations and making incremental updates can enhance performance in dynamic scenes.

6. Advanced Effects

• Multiple Scattering: Simulating multiple scattering events for more realistic cloud appearance is particularly demanding, often requiring approximate methods or precomputed scattering data.
• Dynamic Weather Systems: Simulating evolving weather systems with changing cloud formations adds another layer of complexity and computational demand.

Conclusion

Creating volumetric clouds using voxels is demanding due to the high memory requirements, significant computational load for voxel processing and rendering, and the need for real-time performance in interactive applications. Efficient use of GPU resources, optimization techniques, and parallel processing are essential to meet these performance demands.

Edited June 14, 20242 yr by Noel

2 hours ago, lwt1971 said:

Lack of focus on default ATC (so far) doesn't mean anything, as default ATC in every sim is cr*p and lower priority to other stuff that's been addressed so far .. They have already put lots of focus on weather and will continue to, similar to how they've put focus on avionics, systems, flight dynamics, and other areas of core simulation in MSFS 2020.

I could care less about ATC and a lot of those things, why? Because MSFS allows third parties to actually work on tackling those developmental challenges. 
 

If Asobo wasn’t so hell bent on not letting any 3rd party entity touch the weather engine then a lot of us wouldn’t care about Asobo’s solution either. 

Well all I can say is we'll see.

5 hours ago, UAL4life said:

I couldn't care less about ATC and a lot of those things, why? Because MSFS allows third parties to actually work on tackling those developmental challenges.....

I corrected that for you.

9 hours ago, F737MAX said:

Note the text towards the bottom of this screenshot instead:

https://forums.forza.net/t/forza-motorsport-graphics-downgrade-research-and-comparisons/661530

2024 is stretching the lead in graphics and some people know that if it's flight model ends up exceeding their favorite sim, then it's relevance will be further reduced so the desperation is starting to manifest itself...

8 hours ago, UAL4life said:

If Asobo wasn’t so hell bent on not letting any 3rd party entity touch the weather engine then a lot of us wouldn’t care about Asobo’s solution either. 

Nope, they are not "hell bent" against letting 3rd parties get more access to weather, as they once again answered on the most recent Q&A (the bolded parts below), where Jorg also says they "are spiritually not against it". 

https://forums.flightsimulator.com/t/june-2024-developer-stream-transcript/644684

There’s something else. I think we’re sort of tapping on the subject, but not fully, right? So there’s, I think, three things. There’s the live weather. So that’s on us and MeteoBlue. You heard plenty about this. I just looked at the comments. Yep, yep. We talked about this plenty. We’ll have a call with MeteoBlue again. I think Seb did fix a bunch of stuff just recently, right? When you look at the, I think, where was it, Seb? Like South America or whatnot. But there are some issues for sure. Then there is the, can we share the data that we don’t own with y’all? That’s a business discussion we need to have with MeteoBlue. It hasn’t happened yet. I’m not very optimistic that they say, “oh, yeah, give our data away forever.” I don’t know. It’s not our stuff. But we will certainly talk to them about it. And then the last one is plenty of people seem to like ActiveSky. Do we let this, if there’s some sort of thing that we need to do in order to let this to be more open? That is a discussion we need to have. We haven’t had it yet. I’m spiritually not against it. But we need to talk about it a little bit more and see what the real impact is.

Despite a lot of ex-P3Ders yearning for 3rd party weather to give them proper weather as that's the mindset they're used to (given that with legacy sims 3rd party weather was absolutely needed), the work MS/ASobo need to do (in addition to clearing contractual hurdles) just has not been high enough a priority so far compared to other higher priority items. It is what it is, and the overall userbase agrees given that wishlist item is not as highly backed as other ones. I along with others would rather they spent the resources to improve default live weather first since it's already proven quite capable especially pre SU7 and still continues to be unprecedented for default weather, and *then* look into opening up more SDK access.
 

Edited June 14, 20242 yr by lwt1971

1 hour ago, Krakin said:

2024 is stretching the lead in graphics and some people know that if it's flight model ends up exceeding their favorite sim, then it's relevance will be further reduced so the desperation is starting to manifest itself...

Especially when even with MSFS 2020 as it currently stands paired with properly developed aircraft and FMs, there is no major differentiation apart from ground handling which also has been revamped in v2024 and partially backported to v2020. But I'm sure the stale talking points and tropes will be desperately attempted to be propagated 🙂 
 

2 hours ago, Krakin said:

2024 is stretching the lead in graphics and some people know that if it's flight model ends up exceeding their favorite sim, then it's relevance will be further reduced so the desperation is starting to manifest itself...

 

1 hour ago, lwt1971 said:

Especially when even with MSFS 2020 as it currently stands paired with properly developed aircraft and FMs, there is no major differentiation apart from ground handling which also has been revamped in v2024 and partially backported to v2020. But I'm sure the stale talking points and tropes will be desperately attempted to be propagated 🙂 
 

Are comments like these really necessary? Let's keep it about MSFS please. Besides, given all the flaws in MSFS2020 there is room for competition. That will not be any different when MSFS2024 is released. Or do you think the competition is standing still? 

40 minutes ago, Rimshot said:

 

Are comments like these really necessary? Let's keep it about MSFS please. Besides, given all the flaws in MSFS2020 there is room for competition. That will not be any different when MSFS2024 is released. Or do you think the competition is standing still? 

I have absolutely no problem with what I said considering the types of posts I was referencing. The competition certainly isn't standing still but it is pretty obvious where the really long strides are being made. @lwt1971 and I just stated the obvious and there's no problem with that in my estimation. I also said the relevance would be reduced, not that any other sims would completely die out and this was not based on 2020 but a hypothetical scenario in 2024.

Edited June 14, 20242 yr by Krakin

1 hour ago, Rimshot said:

Are comments like these really necessary? Let's keep it about MSFS please. Besides, given all the flaws in MSFS2020 there is room for competition. That will not be any different when MSFS2024 is released. Or do you think the competition is standing still? 

As far as my comment, it was specifically addressing the common myths that continue to be pushed out about MSFS (including in this thread), that's all 🙂 .. Yes of course flaws exist in MSFS as with other sims, and all sims eventually progress in their own ways. It is a discussion forum, so such discussion is bound to happen 🤷‍♂️, I'll let the moderators determine what comments are necessary or not.
 

On 6/13/2024 at 10:12 AM, eslader said:

Same with lens flares. You only see lens flares when a camera lens is involved. Our eyes don't cause lens flares. Absolutely stupid to put it in a game, but almost every game out there has them. If you're lucky, you can turn them off.

 

Well, in MSFS at least, lens flares only appear when you're using the external camera. You don't see them when in the cockpit.

The external/showcase camera is meant to well, represent a camera.

Edited June 14, 20241 yr by Tuskin38