With more screenshots it becomes more evident that terrain mesh quality is quite compromised, maybe not everywhere, but even in key places like the grand canyon, the mesh looks blurry at the closest distance, suggesting that what we see in bingmaps or earth view, will be quite similar to what we see in the sim, at least in this alpha stage.

Still a monumental improvement over what we had previously of course, but I wish Microsoft or Asobo could say something in this regard; what the update cycles will be and whether mesh resolution will be on par with what google has to offer, dito for photogrammetry,

Or maybe it will be done procedurally. Autogen trees look great though and making it look more crisp.

Also wonder whether you will be able to land anywhere where landable, exit the aircraft and walk around.

 

 

I checked the grand canyon and see nothing wrong or low res about mesh. Maybe I am not seeing the specific image you're talking about.

So far everything has been superior to what we have. 

He's confusing terms; mesh can't be "blurry".

He either means the elevation mesh (DEM/DTED/SRTM) isn't detailed enough for his taste but is using the wrong adjective, or he means the imagery resolution isn't detailed enough for his taste and is using the correct adjective.

9 minutes ago, Noodle said:

He either means the elevation mesh (DEM/DTED/SRTM) isn't detailed enough for his taste

Does he fly in real life?  Other than ground skimming?

 

21 minutes ago, fppilot said:

Does he fly in real life?  Other than ground skimming?

 

I...don't know?

The rounded off mountain peaks bother me.  There are no sharp craggies, and the peaks look sanded smooth.

1 hour ago, BijanStudio said:

I checked the grand canyon and see nothing wrong or low res about mesh.

He is talking about the ground texture, not the elevation mesh.  Look at the full size image by user tomiaut1980.  At about the same height as the planes, at the top of where it starts to go vertical.  To me it looks blurry or stretched.  Also the close river bank texture just seems to be the default low res satellite image.  So hopefully they can add more details like bushes and ground textures to make it look better.

And why are you taking this off topic by talking about if he flies.  No wonder threads go off topic...

1 hour ago, GlideBy said:

And why are you taking this off topic by talking about if he flies.  No wonder threads go off topic...

It was valid point regarding what basis of comparison he is making.  Many here fully understand.  

He is right and wrong. He has confused terrain mesh with ortho texture. He is right to say the terrain mesh is too low. Many pics shows this, and I am very concerned, as this should not be a problem for such a nextgen sim. If this has not been improved until now, there might be something that's hard to overcome.

3 hours ago, dongdongliushui said:

as this should not be a problem for such a nextgen sim.

That's very difficult to say. Such detailed mesh data almost certainly doesn't exist or would be unfeasible due to storage requirements.

The only other way such detail could be implemented is by local interpolation and extremely smart tessellation. Which would be tricky to say the least...

4 hours ago, suncoastflyer said:

That's very difficult to say. Such detailed mesh data almost certainly doesn't exist or would be unfeasible due to storage requirements.

The only other way such detail could be implemented is by local interpolation and extremely smart tessellation. Which would be tricky to say the least...

The mesh issue is indeed complex. It has been raised by some us repeatedly, we have some worries about the sharpness of the peaks/ridges in some pics.

The recurring question is what kind of mesh do they use. The world is now served by the 1 arc-second/30m sampling free SRTM data. Many places in NA and western Europe have better meshes available, 10  or 5 m. 

You rightly underline the insertion/compatibilty problem. The plateaus and troughs we see in P3D are a direct consequence of a better mesh laid on the terrain. Will the sloped runways be flexible enough to adapt to a better mesh than default is a pending question.

Edited March 22, 20206 yr by domkle

Looking at it positively - it will be tough for it to be worse than what we've got. I mean, what we have these days is pretty decent anyway as far as mesh goes, at least until we take a 3D scan of the planet lmao.

2 hours ago, domkle said:

The recurring question is what kind of mesh do they use. The world is now served by the 1 arc-second/30m sampling free SRTM data. Many places in NA and western Europe have better meshes available, 10  or 5 m. 

Have we heard about the mesh?  I know they have talked about the three sources of images 50cm-3cm, Satellite, Aerial, and Photogrammetry.  But I don't recall hearing about height data or resolution.  Although I assume Photogrammetry would cover vertically as well.  

41 minutes ago, suncoastflyer said:

Looking at it positively - it will be tough for it to be worse than what we've got. I mean, what we have these days is pretty decent anyway as far as mesh goes, at least until we take a 3D scan of the planet lmao.

What we have now as third party meshes is not so very bad (eg Pilot's mesh) ! And Returning to the 38 m FSX mesh would be worse.

The SRTM is a 3d scan of the planet, Daniel, enhanced in some places by even better aerial 3d scan of the same.

https://www2.jpl.nasa.gov/srtm/

Edited March 22, 20206 yr by domkle

Judging from the screenshot comparisons with Bing Maps it all in all looks as if the mesh they’re using is very similar to the one in Bing Maps. Maybe it’s even the same.

While it looks very good in many places (in some even better than Google Earth) there are other areas where deficiencies are striking. Here are some examples:

Probably the most extreme case is this one:

They’ve shown two places so far where the terrain in Bing Maps and MSFS doesn’t match.

Now this could be an indication of better mesh or these two rocks could be handcrafted. Since we can assume that St. Barts airport got some special attention and Sugarloaf Mountain is a world-famous landmark, I tend to believe the latter is the case. That cliff in St. Barts in particular looks hand-modelled.

I really hope this is just an alpha issue and that Asobo have a solution up their sleeves.

Edited March 22, 20206 yr by Shack95

I think we know now from all the Alpha previews that the terrain resolution, the aerial image data as well as the footprints that give birth to procedural urban areas are not equal or consistent (yet) across the continents.

And this is simply related to data availability (generated in-house or acquired) or the choice of data resolution for performance purposes.

The Beta will tell us.