@MarStrMind I've checked out your Orthographic tile around EDDB. The polys and footprints match well with X-World and it's overall a good demonstration of your information extraction tech from OSM and sat data. However, I think that the delivery format as what essentially boils down to a ZL16 ortho tile with generated textures instead of sat imagery is very space inefficient and frankly offers no advantage over simply generating a straight ortho tile.

Since you already have all the terrain feature polygon data extracted, you can, however, try to bake these polygons straight into the DSF file and have a pool of generic textures (and links to default textures) that are assigned via library.txt (which also allows users to easily swap out or edit textures to thewir taste). X-World is doing this in parts for meadows and some other terrain features. This has the massive advantage that you only need to ship your DSFs and a library, decreasing the required storage space to a fraction of an ortho tile.

For further experiments in that regard, I suggest getting in touch with SimHeaven, as they've figured out how to set up XP's DSF generation pipeline to bake their polys into the DSFs. Maybe the both of you can even team up to generate a base terrain for X-World that incorporates your poly extraction tech and essentially replaces XP12's global scenery.

20 hours ago, Bjoern said:

@MarStrMind I've checked out your Orthographic tile around EDDB. The polys and footprints match well with X-World and it's overall a good demonstration of your information extraction tech from OSM and sat data. However, I think that the delivery format as what essentially boils down to a ZL16 ortho tile with generated textures instead of sat imagery is very space inefficient and frankly offers no advantage over simply generating a straight ortho tile.

Since you already have all the terrain feature polygon data extracted, you can, however, try to bake these polygons straight into the DSF file and have a pool of generic textures (and links to default textures) that are assigned via library.txt (which also allows users to easily swap out or edit textures to thewir taste). X-World is doing this in parts for meadows and some other terrain features. This has the massive advantage that you only need to ship your DSFs and a library, decreasing the required storage space to a fraction of an ortho tile.

For further experiments in that regard, I suggest getting in touch with SimHeaven, as they've figured out how to set up XP's DSF generation pipeline to bake their polys into the DSFs. Maybe the both of you can even team up to generate a base terrain for X-World that incorporates your poly extraction tech and essentially replaces XP12's global scenery.

I tend to disagree... my ortho tiles come in at around 5 GB at ZL16 (with ZL18 at airports) - Orthographic tile sits at around 2.74GB for +52+013.

In the meantime I have noticed there is practically zero interest in the project, so I am continuing it on a private basis.

45 minutes ago, MarStrMind said:

I tend to disagree... my ortho tiles come in at around 5 GB at ZL16 (with ZL18 at airports) - Orthographic tile sits at around 2.74GB for +52+013.

For comparison, I am a big fan of ZL14 orthos because EOX as a source has them color matched and cloud free out of the box and a pure land tile is around 500 MB. But generating tiles is a huge pain because both OSM and EOX are throttling their connection requests to not be ground down by some external LLM constantly trying to leech their data.

45 minutes ago, MarStrMind said:

In the meantime I have noticed there is practically zero interest in the project, so I am continuing it on a private basis.

As hard as it is to say, but in its current form of delivery, Orthographic offers no advantage because each tile still takes up space in the region of ZL15+ orthos. But I'd argue that the gap for better terrain is still there - regardless of when we will get a new default scenery from Laminar. The current default terrain doesn't cut it ("default is fine" fans, no neeed to reply to this), orthos take up too much space, AutoOrtho or MapEnhancement have their "always" online drawbacks and it's essentially the same imagery as in O4XP, which means color matching and cloud issues (depending on Zl and provider). So if you deliver your tech as pure DSFs with a texture library, you'd sit in the gap for a space-efficient, more detailed alternative to default terrain. The O4XP DSF for +52+013 is 42 MB with Sonny's mesh. With the required ground texture polygon definitions, you'd end up at around 50 MB per tile, which is a far cry from the 500 MB ZL14 and 2.7 GB for Orthographic's ZL16 (not counting the texture library). For the entire globe, the tally will likely be around 300 GB (at worst). That's less than what ZL14 orthos for Europe take up.

Think about it.

No need for me to think about it as I am most likely canceling the project and remove it from public view.

I'm a bit late to the party, I just had the chance to see the screenshots you posted, but the patreon was already down. Too bad, because I'm interested. I've been developing something along these lines myself, for years, so it would have been interesting to be able to share some of our research. Well...

From what I remember, I've noticed one thing that others might not have caught: the fields division accurately taking the roads into account. I don't know if it was imported as is from OSM (not many places have that), or if you have an algorithm for that.

The main issue I noticed is colorimetry: too faded colors (from memory). That's what might have put the other people here off. But I recon that's a though thing to get right.

Well, in the end, I would have liked very much to see what was on your patreon. And let me tell you this: you should not rage quit a project based on a few people's opinions, especially if you believe in what you do. I've had a few hardships myself during the years, most notably a bug in X-Plane that totally blocked me during several years, and that totally disappeared from Laminar's radar until I realized it was not referenced anymore. I had a temporary project burnout because of that, but in the end I work on it again. Now I don't even know if it will be compatible with the new scenery format. Well, many things came out of it anyway, and I don't regret having invested years of my life into it.