POST
|
We identified the problem of the broken drawing shown in the screenshot (the funky colorful triangles) when applying &disableFeatures=im-shading SceneLayers (which includes Integrated Mesh Layer) are cached in the browser's indexedDB for performance reasons a slightly different version of integrated mesh data gets cached when the url parameter &disableFeatures=im-shading is set loading a non-matching version from the indexedDB cache causes the broken rendering The fix/workaround: when switching between im-shading disabled and enabled using the url parameter, the browser's indexedDB cache must manually be deleted Clear SceneLayer indexedDB in Chrome (similar in other browsers): open developer tools navigate to Application tab under Storage / IndexedDB, select esri-scenelayer-cache click Delete database Apologies for the inconvenience on this, we hope this workaround is sufficient for you for the moment. And as stated earlier, we are revisiting this behavior for the upcoming 4.31 release. ----- I marked it solved by accident. FYI, I marked it as unsolved again for now.
... View more
08-14-2024
06:55 AM
|
1
|
0
|
540
|
POST
|
Update: we can now reproduce the broken drawing as seen in the screenshot, and are looking into it.
... View more
08-13-2024
09:05 AM
|
0
|
0
|
192
|
POST
|
@Shannon Is the screenshot what you get after applying &disableFeatures=im-shading? That should not happen of course. Is there a chance you could share the problematic scene with us (the url), with an instruction how to exactly get to the problem? It's hard to say more without seeing the actual problem. In case you don't want to share the url publicly you can also DM (direct message) me. Also, confused why you marked the case as solved. Was this done by accident?
... View more
08-13-2024
08:33 AM
|
0
|
0
|
413
|
POST
|
In ArcGIS Online SceneViewer, you can add this url parameter to disable the new integrated mesh shading: &disableFeatures=im-shading This should get you back to the pre-4.30 state. Does this work for you as a workaround for the moment? And FYI, we are reconsidering this behavior for 4.31, but no decisions yet.
... View more
08-12-2024
08:00 AM
|
0
|
0
|
485
|
POST
|
@ShannonApgar-Kurtz, @ArthurJensen: are you using SceneViewer on ArcGIS Online? Or a custom application built with ArcGIS Maps SDK for Javascript?
... View more
08-12-2024
05:22 AM
|
0
|
1
|
498
|
POST
|
With the 4.30 release, we "Added shading to IntegratedMesh3DTilesLayer and IntegratedMeshLayer" (https://developers.arcgis.com/javascript/latest/release-notes/#bug-fixes-and-enhancements). In case your imagery is an integrated mesh this likely explains the visual difference. In this context, surface normals (the local direction of the surface) on integrated mesh are used to shade (apply lighting to) the surface. This helps to improve realism under specific lighting conditions and integrate IntegratedMesh data better into the scene (since most scene elements are being shaded this way). Shading also helps to understand the shape of the integrated mesh better, but at the same time might expose issues in the data set. In the case at hand it seems that the data has incorrect normals at tile borders (normals not properly interpolated across borders), which get more visible now due to shading.
... View more
08-08-2024
04:29 AM
|
2
|
2
|
610
|
POST
|
Many thanks for looking at this Andreas and your useful advice. I'll have a go at implementing your suggestion. Great! Let us know if this works for you. Do you know when the use of the fov function with goTo will be available please ? Added issue to our roadmap, but can't give a concrete date when this will be tackled I'm afraid.
... View more
04-22-2024
06:52 AM
|
0
|
1
|
1218
|
POST
|
Suggested workaround: apply new fov before view.goTo() (ℹ Camera needs to be cloned to apply modifications: https://developers.arcgis.com/javascript/latest/api-reference/esri-views-SceneView.html#camera) document.getElementById("gotoVP2_53.5deg").addEventListener("click", () => {
const cameraCloned = view.camera.clone();
cameraCloned.fov = 53.3;
view.camera = cameraCloned;
view.goTo(
... this will instantly (without animation) change camera fov to the new desired setting then the camera will transition smoothly to the new viewpoint Unfortunately there is no way atm to smoothly transition between fovs. This is something we consider to improve in a future release. Is this workaround acceptable for you?
... View more
04-19-2024
07:31 AM
|
1
|
3
|
1252
|
POST
|
Thanks for reporting, this seems to be a bug. Sent to dev team for further investigation. I did a few tests but at the moment I can't provide a good workaround. Use "on-the-ground" of course if that's an option. Or clone/recreate the whole layer (with new callout config in constructor), but clearly not great either depending on your use case. Will keep you posted if more intel comes up.
... View more
03-21-2024
11:41 AM
|
1
|
0
|
280
|
POST
|
Hmm, really hard to say what causes this without an actual reproduction case. Could you still share the codepen (even if it doesn't reproduce the problem), so we have a starter what to look? (We specifically need to know what data and renderer / symbology is used, and how featureFilter is applied) Also, can you check the following: Can you check if it also happens with 4.29? Can you confirm the flicker does not happen when FeatureFilter is not applied? As said, hard to say but from the current recording I would say: it's not z-fighting (happens only with coplanar faces) FeatureLayer, GeoJSON, OGCFeature should not make a difference (the internal rendering is the same on those)
... View more
03-21-2024
05:22 AM
|
0
|
0
|
234
|
POST
|
Identified as bug / regression in 3D SceneView (missing antialiasing). Fixed, lines are smooth again. Will be shipped with next ArcGIS Online release, Feb 2024.
... View more
11-23-2023
11:22 AM
|
0
|
0
|
2192
|
POST
|
Draco compression is currently not supported for 3D Symbol resources. This is in our backlog but not yet planned for a release. Basis Universal compressed textures are supported though. If the models contain large textures, maybe this could be worthwhile to reduce size?
... View more
11-17-2023
03:46 AM
|
0
|
0
|
597
|
POST
|
First answer from the linked post also applies here: Use Color VisualVariables. This concept works for both CIMSymbols and non-CIM 3D symbols. See samples: * https://developers.arcgis.com/javascript/latest/sample-code/visualization-multivariate-3d/ * https://developers.arcgis.com/javascript/latest/sample-code/layers-scenelayer-edges/ Hope this helps.
... View more
11-06-2023
06:17 AM
|
0
|
0
|
196
|
POST
|
> We've seen performance benefits from using 512x512 tile sizes over 256x256 on high-res or large screens. Oh that's interesting! Would it be possible to share some more specific details (either here or DM if you prefer)? What particular improvements? Improved FPS? Faster loading of data? Definitely something we'd like to hear more about. I assume a 512 tile cover the same area as 4 256 tiles (and not the same area at 4x resolution)? AFAICT this reduces the number of network requests to 1/4, which could indeed help with certain aspects of performance. The whole terrain is then split less granular (again 1/4 tiles), I wonder if this comes with any downsides. > If supporting mixed tile sizes (even limited to 256 and 512) in 3D is complex Mixing arbitrary tiling schemes is tough: Tiling is at the core of the terrain logic and controls a lot of both internal and user-facing behavior. Mixing 256 and 512 would definitely be less tricky but still not a freebie.
... View more
11-02-2023
11:18 AM
|
0
|
1
|
1321
|
POST
|
> Raster API tiles only work when all other layers have same tiling size. Correct. In SceneView tiling scheme of all (tiled) layers must match. First layer that is loaded determines the scheme (typically the first basemap layer). FYI: This limitation in SceneView stems from the fact that in contrast to top-down-only MapView, oblique views in SceneView requires tiles to be displayed in different scales (different distance to camera). Mixing different tiling schemes would require the current single-tiling-pattern of the ground surface to be extended to support multiple overlayed tiling patterns. > I think ESRI ArcGIS Maps SDK for JavaScript maps require a basemap layer right? No, a basemap layer is not required (was the case in older versions but not anymore). If you do not need the basemap, maybe removing it is the quickest "fix" for your case?
... View more
10-31-2023
07:49 AM
|
1
|
0
|
1345
|
Title | Kudos | Posted |
---|---|---|
1 | 08-14-2024 06:55 AM | |
2 | 08-08-2024 04:29 AM | |
1 | 04-19-2024 07:31 AM | |
1 | 03-21-2024 11:41 AM | |
1 | 10-31-2023 07:49 AM |
Online Status |
Offline
|
Date Last Visited |
08-27-2024
09:46 AM
|