From the changelog of world update 3: “Several terrain spikes have been fixed”.
I am afraid that this implies that fixing spikes is a manual process? Oh dear: so it may take a while until all spikes are gone (assuming that no new ones will “pop up”).
I was hoping that this was merely a mesh generation parameter (or fix) that would solve all spikes all and forever. Just like they had appeared out of nowhere, caused by something.
But it seems that the spikes are caused by something more fundamental, possibly a new world mesh generation method, new (server-side) mesh data or whatever… something that Asobo cannot (do not want to) “turn back”.
So they were really serious when they said that we should report individual spikes (via Zendesk I guess), including GPS coordinates… sigh…
As I haven’t yet finished downloading the update I cannot really tell right now how many spikes are left, but I am definitively going to check out “the Dark Tower” near Thun (Switzerland)…
Yes there are gone i have checked it in my home region area near edrk(Koblenz-Winningen) Before the newest update was out there are a lot of spikes on the terrain after the update there are gone. I hope this helped you a bit.
The ones that I filed tickets for in the previous version are still present in this version. When they say “fixed several terrain spikes” it seems to mean they didn’t fix the actual issue and just used terraforming to band-aid over some major spikes…
They’re likely refining the noise filter they use to smooth out these terrain spikes. You can see these spikes in the raw LIDAR data government sources like the USGS provide as well. It’s a difficult balancing act, because if you go too aggressive with your noise filter, you end up with the problem we have in areas that haven’t been updated with better elevation data: the filter causes mountains and ridges to become overly rounded and ill-defined.
My guess is that all of the really egregious stuff that was obviously caused by a bug prior to the latest update was fixed, and there’ll be ongoing changes going forward as they’re able to find a happy medium between few/no spikes in a region without compromising the good data.
The weird thing is that the spikes only appear at certain distances, which means they’re either not in the original data sets, or the data sets at different resolutions are derived from different sources, some of which contain spikes and some of which don’t.
If they were in the original data, and the data at all resolutions were derived from the same source, then getting close to a terrain spike would not make it disappear as they do.
spikes are gone.