terrain mesh, will it improve

[Brandon01110 407]

I hope they update map data more often as the image for the city i live in is about 5 years old as my house is not there yet.  Google earth image is only 1.5yr old.  So it seems for now i wont be able to fly over my house....

[Dominique_K 10,138]

5 hours ago, Shack95 said:

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

Thanks again for the good job. I agree with your conclusions and would add that the Himalaya pic is also difficult to interpret as it is probable that the resolution of the mesh decreases with the distance .

If their sleeves are empty, I hope that we will be able to add a better mesh without anomalies cropping up.

 

[Shack95 2,658]

1 hour ago, domkle said:

Thanks again for the good job. I agree with your conclusions and would add that the Himalaya pic is also difficult to interpret as it is probable that the resolution of the mesh decreases with the distance .

Anytime. Indeed, that would be an explanation for that Himalayan mountain. Let’s hope. I have to admit that I‘m a bit sceptical here as the similarities between Bing and the sim are striking.

1 hour ago, domkle said:

I hope that we will be able to add a better mesh without anomalies cropping up.

Same here.

[Chock 17,769]

 A difficulty here, or perhaps more correctly called a complication/expense, is that whilst pretty good terrain elevation data is generally available for most places, even on the shuttle terrain elevation radar mapping mission, some areas were covered better than others. Another complication with that data's resolution, and other data, is that whilst NASA's Shuttle mapping mission data is in the public domain, based on the notion that since Joe American Public fund NASA and the US Government, the data is freely available to Joe Public, other nations are not so philanthropically inclined.

For example, at the turn of the new millennium, the UK Government funded a project to aerially map the UK with high resolution photography. But did they make that data freely available to Joe Public in Britain? Not on your life, which I personally find outrageous since me and my fellow citizens in the UK, in a manner similar to those of the US, funded that operation. The UK is by no means alone in being mean with data availability either, lots of other governments have been similarly stingy with what they are in possession of. There are of course also military missions which map such data and spy satellites which do it too, and how good that data is might be subject to the requirement to deny its existence, so how much of that is likely to be released by any government for wider use, by nations generous to the world at large or otherwise, is questionable.

Since the BGM TLAM (aka the Tomahawk cruise missile) and some optically-guided versions of the Maverick missile (i.e. the AGM-65A/B variants), actually navigate by comparing an image database gained from reconnaissance missions, with the terrain their image seeker is detecting as they fly over it, decent terrain elevation mesh data and photographic maps of that terrain have to be pretty good and also have to be regularly updated too, which is often what things such as the SR-71 and U2 etc were up to. Not only this, the systems have to also be capable of detecting seasonal changes too, since a war can break out in winter just as easily as it can in summer.

So I'm willing to bet that there is particularly good mesh and terrain data for North Korea, Iran, Iraq and other fun holiday destinations such as these, for obvious reasons. Whereas since the US are fairly unlikely to be launching a cruise missile attack on for example, Cardiff in South Wales any time soon, the freely available data for there might not be so high resolution, but ironically, it might actually be more likely to be available. Which kind of explains a lot if you think about it.

Edited March 22, 2020 by Chock

[Claviateur 789]

The terrain components including the Mesh could have a variety of resolution layers depending on the importance of the location and its elevation as well as the connection mode.

But hopefully, the terrain level of details (LOD) won't change and morph with distance like FSX.

However, I am almost certain photogrammetry areas will morph depending on distance/LOD. It seems so in a couple of preview videos as well as in many other appliactions including GE... 

Edited March 23, 2020 by Claviateur

[Noodle 626]

OT: No variant of the AGM-65 navigates using terrain or image comparison. All non-laser guided versions are optically guided based on edge detection of target objects based on video contrast.

[Marcoh 5]

On 3/22/2020 at 10:08 AM, Shack95 said:

 

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.

 

Hello, the rock in St-Barth is quite probably hand-crafted, the rest of the mesh in St-Barth seem to be LOD 9.

Here a comparison, first from my scenery, with a mesh from IGN at 20m/pixel, corrected localy.

The second is a screenshot from a MS discovery episode

We can see that the skyline is more smooth in the second picture.

We can also see that the color of the water is too dark and greenfor this lovely beach.

Note : the red ligne was a test to verify that the runway is sloped but strait (not ondulated).

But for me, it's not a problem. It's mean that there will be plenty of room for developpers for enhance the virtual world.

[Claviateur 789]

31 minutes ago, Marcoh said:

 

Hello, the rock in St-Barth is quite probably hand-crafted, the rest of the mesh in St-Barth seem to be LOD 9.

Here a comparison, first from my scenery, with a mesh from IGN at 20m/pixel, corrected localy.

The second is a screenshot from a MS discovery episode

We can see that the skyline is more smooth in the second picture.

We can also see that the color of the water is too dark and greenfor this lovely beach.

Note : the red ligne was a test to verify that the runway is sloped but strait (not ondulated).

But for me, it's not a problem. It's mean that there will be plenty of room for developpers for enhance the virtual world.

Nice observation but I think the runway must have a slope otherwise the link to the sloped terrain all around would be very harsh at some sectors. It could be that the slopes were softened and refined but I think logically the runway can't be flat if it must link to the the global terrain without harsh differences.

[Dominique_K 10,138]

32 minutes ago, Claviateur said:

Nice observation but I think the runway must have a slope otherwise the link to the sloped terrain all around would be very harsh at some sectors. It could be that the slopes were softened and refined but I think logically the runway can't be flat if it must link to the the global terrain without harsh differences.

The question is, is in FS20 the runway an objet shoehorned into the terrain with a little trimming around or is it a portion of the terrain defined as a runway and textured as such.

Edited March 24, 2020 by domkle

[Claviateur 789]

10 minutes ago, domkle said:

The question is, is in FS20 the runway an objet shoehorned into the terrain with a little trimming around or is it a portion of the terrain defined as a runway and textured as such.

I believe it is not an object but rather a native engine feature. It's a sort of draped polygon that has, perhaps, a terrain softening parameters and auto refinement most probably. The Airport video shows the in-engine tools to create airports including runways and I believe these must have some parameters somehow for the terrain. Or, an airport as a whole has a terrain softening parameter.

Taxiways could also have this capability of softening the terrain varatiations underneath. 

[Dominique_K 10,138]

I intuitively agree with you. That is why I am not too worried  about slopes and undulations.

[Marcoh 5]

I'm not worried too.

MS sayed in the video that the runway slope is defined from data.

I have data (From ICAO i guess) for this airport. Only two point of altitude are defined (the extremety).

That could explain why the runway is sloped like this.

If we have more data, we'll place more altitude points along the runway, to match the mesh more precisely.

[Claviateur 789]

27 minutes ago, Marcoh said:

I'm not worried too.

MS sayed in the video that the runway slope is defined from data.

I have data (From ICAO i guess) for this airport. Only two point of altitude are defined (the extremety).

That could explain why the runway is sloped like this.

If we have more data, we'll place more altitude points along the runway, to match the mesh more precisely.

If the data means only the 2 points from ICAO as you mentioned, then the whole airfield will be based on the average of the runways tips. For small airfields with one runway, perhaps this could function but I can't see this being practical or done easily in large complex airport layouts with multi runways (some crossing each others). 

However, the runway tips altitude could still be the references and the in between is a terrain average + transitions + auto smoothing I believe.

With the airports video from Asobo, we see that the creation of an airport layout is the most practical airport layout design to date in an airport creation tool for any simulator I know... And as far as I can judge from the in-sim / in-engine tool usability and style, I believe as I mentioned above, an auto smoothing algorithm could be very well involved. 

Auto smoothing of the terrain could make sense since average terrain mesh resolution in some corners of the world won't allow a practical use of an airfield (i.e: it will be too rough).

But all in all, I don't think the 2 points of the threshold altitude are enough to create the slope of any runway because in between, it must follow the global airfield terrain (no matter how the smoothing process goes), and because runways are linked and part of a network of taxiways that could be very well auto refined using averages or other algorithms...

 

Edited March 24, 2020 by Claviateur

[GlideBy 324]

8 hours ago, Claviateur said:

Nice observation but I think the runway must have a slope otherwise the link to the sloped terrain all around would be very harsh at some sectors. It could be that the slopes were softened and refined but I think logically the runway can't be flat if it must link to the the global terrain without harsh differences.

The MSFS image the runway looks flat.  Where if the had better terrain resolution and mapped to the terrain, there would be a slight ramp at the end.  I think this real life picture shows it better (and the water color).

 

And, as you say, if they are just using two end points for the runway, that just isn't enough.  And didn't Courchevel show changing slope?

Edited March 24, 2020 by GlideBy
typo, churchlevel

[Claviateur 789]

3 hours ago, GlideBy said:

The MSFS image the runway looks flat.  Where if the had better terrain resolution and mapped to the terrain, there would be a slight ramp at the end.  I think this real life picture shows it better (and the water color).

 

And, as you say, if they are just using two end points for the runway, that just isn't enough.  And didn't Churchlevel show changing slope?

Yeah Courchevel is one example for sure. And more complex airports won't be possible with 2 points for each runway.

I presume and hope the user friendly airport creation as per the Asobo video takes care of the terrain smoothing and I suppose parameters could be involved.

Any auto-smoothing process with some parameters for ref points or any other in-engine process to tune the terrain of the airfield is way better than manually exporting and editing the airport mesh/terrain using a 3D app and super technical 3rd party programs, as I did for my project for the other simulator.

 

Edited March 24, 2020 by Claviateur