Yes, Rio has a few "visual" issues (have seen it a few weeks ago). For one, its SRTM lacking some features like the Sugarloaf ... its just not in the data (I checked it). And second, obviously some mismatch happened between the coastline as SRTM sees it, and the coastline in the OSM (OpenStreetmap data). This is important, because the OSM coastlines "induce" the water, and the scenery creation process tries (and as we see, sometimes fails) to deal with it correctly. it has two options if there is a mismatch, either let water run up the hills of the SRTM data (does look funny), or the water tries to remain flat (thats what the processing is going for), but then the "land" has to give way. The result is, that sometimes, where SRTM still thinks there are hills, the OSM coastline bites out some pieces ... resulting in those steep cliffs. Who is exactly at fault, I can't tell for sure (either SRTM or OSM ... or just a little shift in their spatial alignment) but its hard to take care of all of these issues with a mostly automatic process for the entire planet (neither can we check every square mile visually) ... the only thing we can do is, to try to fix issues like this .... If you posted it to Laminars Bug database, we at least have a "reminder" to look at it the next time new scenery tiles are being cut. But especially if the data is faulty, its not trivial to get it fixed ...And for the mesh resolution ... its not working as trivially as you know it form MSFS. its a bit more complex (for example, an irregular triangle mesh! If you want to understand the - little bit hardcore :( - basics, read the interview "xplane10s Blog" did with me in December: http://xplane10.wordpress.com/2011/12/10/developer-interview-andras-fabian-mr-x-terrain/

Thank you for the excellent explanation! No doubts after reading that.So, there's not much that can be done in my case, I won't even bother with the Bug report. From what I understood from the blog post, it's incredibly more complex to make Mesh for X-Plane than for FS, even if the results can be much better. That means it's probably going to take a long time until mesh developers can learn how to do it and start selling those. Not the kind of news I wanted to hear :(I guess that's the price of trying to improve something that's been around for so many years. It's a project for years, not months.Thank you again!

Mesh Looks fine at São Paulo, so I guess it's just Rio de Janeiro that got screwed up.

I've constantly asked numerous scenery developers to make Australia with accurate SRTM data and they all say the same thing. Updated SRTM data for Australia is too expensive.

Just to clear up what you are being told by other developers, the U.S. handled the SRTM process. It was flown to gather 30m material from 60N to 60S. Once the government got the data they deemed it classified, though released it at a lower 90m resolution. I could ponder MY government's thought process at deeming 30m source classified and 90m source as open to the public, but it boogles my mind TBH. Some of the 30m source has been shared with other country's government, subject to the same (twisted!) classified tag.And then a few years after the SRTM mission the U.S. and Japan release 30m ASTER source to the public. Go figure that one out!The Land Down Under may have other source material available, but it's not part of the SRTM mission. Many times when doing flights for higher resolution aerial imagery they will also include laser based terrain mapping as part of the process. That can be expensive, though for the most parts, in the U.S. it's generally considered public domain data.For right now, 76m is the best for your area in wide scale availability, in FSX. That may or may not change in the future?