Hi @Gabby_Getz
The code you shared is nearly the same as the one in Case 1 of my very first email in this thread, except that:
- I had used
geoCoder: true (just to be able to move around while debugging); and
- I wasn’t calling
viewer.zoomTo(tileset)
Barring the above two, my Case 1 code is identical to yours. I reconfirm that this code works without any gross issues such as exceptions, blue-colored background, a missing globe, or imagery reduced merely to a few faint gray patches (possibly corresponding to full-color building tops in the live, hires map).
However, the problem with my Case 1 code was/is that it displays the map in lowres, enough to make it unusable in our application. We need at a Zoom Level 18 or better.
My (quite possibly incorrect) thinking so far has been: Because the assets that come packaged with the NPM cesium module (version 1.120.0) could no way be hires (owing to the bulkiness of any hires imagery data, in general), the “clip and ship” feature of Ion would possibly afford me data that is at least a tad higher res than that, if not as hires as Bing Maps or Google Photorealistic Tiles data (which I am well aware isn’t allowed to be downloaded as part of the clip). Further: If it’s not a tad higher res, it shouldn’t at least so low res as to make it unusable (in our specific application).
So, in line with the above thinking, I’ve been trying to use .../imagery-and-terrain/tileset.json from my local tile-server, and that is where I’ve been running into issues.
I’m not sure what the content of imagery-and-terrain is in your example. Could you explain more? Does it contain a global imagery data set?
The clip is for a small region over Lhasa, China.
I tried both Sentinel-2 and Natural Earth II sources for imagery and terrain, in case it made any difference. It didn’t, except that one of these two sources for imagery displayed far worse than the other.
Here are the details of the clip.
(Just to reiterate: I tried the 3D Tiles data with both versions of the imagery-and-terrain data with details below.)
Select clip type: Clip 3D Tiles
Select 3D Tiles: Cesium OSM Buildings
Select Region:
North: 29.681881004
East: 91.184111274
South: 29.646959451
West: 91.139132862
Clip options:
Name: lhasa-a-3d-tiles-only
Format: 3D Tiles (not glTF)
Select clip type: Clip imagery and terrain
Select imagery and terrain:
Imagery: Sentinel-2 <------------
Terrain: Cesium World Terrain
Select Region:
North: 29.681881004
East: 91.184111274
South: 29.646959451
West: 91.139132862
Clip options:
Name: lhasa-a-imagery-and-terrrain
Format: 3D Tiles (not glTF)
Select clip type: Clip imagery and terrain
Select imagery and terrain:
Imagery: Natural Earth II with Shaded Relief, Water, and Drainages <------------
Terrain: Cesium World Terrain
Select Region:
North: 29.681881004
East: 91.184111274
South: 29.646959451
West: 91.139132862
Clip options:
Name: lhasa-a-imagery-and-terrrain-natEarthII
Format: 3D Tiles (not glTF)
You also mentioned that only the tops of buildings are showing. I’m not totally sure why without looking at the data, but I suspect this is due to the height of the buildings being below the WGS84 ellipsoid. You can have the buildings show up overtop of the base globe by adding the following line: viewer.scene.globe.depthTestAgainstTerrain = false;
My building-tops related observation could possibly be a wrong interpretation – all I can (re)confirm seeing is a few, very faint, gray patches over a black background. But the more concrete part of the observation was, the imagery-and-terrain URL wasn’t even getting hit on my local tile-server; only the .../3d-tiles-only/tileset.json was getting hit(!), telling me that my API usage for loading the .../imagery-and-terrain/tileset.json from the local tile-server was wrong.
I’d tried viewer.scene.globe.depthTestAgainstTerrain = true earlier because I did want to be able to depth-test (in our application) but it made no visual difference to the display of the imagery and the globe. Also, I tried setting this property to false (per your suggestion) and, again, it made no visual difference.
I could share the clipped data if needed. I’m including a screenshot of what I saw in Attempts 3 and 4 earlier. (Both attempts produced the same results, as I noted earlier. The entities in red in the screenshot are my own. The faint gray patches are what I’ve been positing as “building-tops” but they could just as well be some other pixel artifacts.)
Thanks.
