1

After converting a Revit model to a WebScene I see some edges are jagged:

enter image description here

Where the edges in the original are sharp:

enter image description here

When this jagged part of the model is processed separately the results are as expected:

enter image description here

The steps I followed are:

  • In Revit, using the FME Exporter for Revit Add-in to create a .rvz (essentially an IFC export).
  • In FME Desktop (2018), read the .rvz, scaled the model (XYZ 0.001, mm to m), added a (projected) coordinate system and wrote it as MultiPatch to a File Geodatabase.
  • In ArcGIS Pro, created a Project, created a Local Scene, added the FeatureClass from the File Geodatabase, used the "Create 3D Object Scene Layer Package" tool to export a Scene Layer Package.
  • In ArcGIS Online, tab My Content, added the package via the Add Item button.

Initially I suspected scaling the model from millimeter to meter caused this. But when the only the jagged item is processed it works fine, so I think I can rule this out.

Would be very happy if someone can point me in the right direction.

Disclaimer: I'm just getting started with 3D data like Revit models and WebScenes, so I probably ask something obvious. Do work with 2D / 2.5D vector data and ArcGIS for Server for a while so not a complete novice.

Update 201903050902:

  • Processed via Shape instead of File Geodatabase, no difference.
  • Found and used the "Validate Scene Layer Package" tool in ArcGIS Pro. Some interesting results:

In the complete package generated by ArcGIS Pro I get a lot of errors like:

-------------------------------------------------------------------------------
Node validation:
...
    Warning: Invalid Value:
        The mbs of node 0 extends outside the mbs of its parent, node root, by 5949660% of the radius of the mbs of its parent (10543395,33 meters).
        The mbs of node 1 extends outside the mbs of its parent, node root, by 5237936% of the radius of the mbs of its parent (9282148,94 meters).
        The mbs of node 2 extends outside the mbs of its parent, node root, by 1348704% of the radius of the mbs of its parent (2390038,83 meters).
        The mbs of node 3 extends outside the mbs of its parent, node root, by 2754633% of the radius of the mbs of its parent (4881487,30 meters).
        The mbs of node 4 extends outside the mbs of its parent, node root, by 1475275% of the radius of the mbs of its parent (2614336,32 meters).
        The mbs of node 5 extends outside the mbs of its parent, node root, by 2329187% of the radius of the mbs of its parent (4127553,01 meters).
        The mbs of node 6 extends outside the mbs of its parent, node root, by 2344830% of the radius of the mbs of its parent (4155273,86 meters).
        ...

The complete package generated by FME has also issues, but the are different:

-------------------------------------------------------------------------------
Node validation:
...
    Warning: Missing Value:
        Node 0 has no content and is empty.
        Could not get attributeData array for node 0.
        Node 0-0 has no content and is empty.
        Could not get attributeData array for node 0-0.
        Node 0-0-1 has no content and is empty.
        Could not get attributeData array for node 0-0-1.
        Node 0-0-1-1 has no content and is empty.
...

As we have other models without this kind of issues I suspect the cause of the problem is in the source (Revit) model.

1

a couple things to try. If you write to a multipatch shape file, do you see any difference? I don't know if its still a thing or not, but longer ago when you wrote to some geodatabase formats, coordinates were snapped to a grid. And so the reason you may see different results if you write only 1 feature is that the bounds of the dataset may be much smaller and so the grid may be much more resolute. So I'd try going to Shapefile because there I know it is double precision.

I'd also be curious if you viewed teh resulting dataset with the FME Data Inspector if the jaggy is there or not. Both when you wrote to Geodb as well as when you wrote to Shape.

Lastly, I'd invite you to try reading the Revit .rvt file directly with FME 2019 beta just to give us feedback on that...though I'm 100% sure your issue isn't actually the IFC/RVZ input..but hey, I thought I'd throw that in.

Will be very curious to know where this is going off the rails.

Dale

  • Hej Dale, didn't post this on the Safe forum as I suspected this to be a problem on the ESRI side of things. Processing it via shape did not change anything. Both the GDB as the SHP show correct in the Inspector. Discovered the "Validate Scene Layer Package" tool in ArcGIS Pro and checked the exports. The complete package results in a lot of warnings like "The mbs of node 0 extends outside the mbs of its parent, node root, by 2015614% of the radius of the mbs of its parent (3574644,98 meters)." where the deck only has none of these. – nielsgerrits Mar 5 at 7:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.