Improvements to vaPlanView, vaSectionView

Hello!
I feel the urge to list how vaPlanView (and vaSectionView) could and should evolve, since I need VA for an older project, and it’s still a hassle to create good 2D from the 3D scene.
Can’t find much related in your roadmap also.

What’s working:
curves → curves
surfaces/breps → outline curves
section attributes (Comes to Rhino8, too. However, clipping planes aren’t as flexible as vaSections).

What’s missing:

  1. hatch → hatch (also transparent ones)
  2. text → text (Make2D can do it. In VA, boundary curves are created. What to do with text crossing the frame? Could/should simply remain text.)
  3. breps with materials of constant color → hatches with this color
  4. breps with materials with textures → rendered bitmaps (similar to what texture baking does)
  5. blocks → blocks (Oh yeah! Together with DWG export from Layout space, this would rock!)
  6. Support placing plan and section views into a Layout (You mentioned this is planned. Is it on the roadmap?)
  7. annotation → annotation (You supported this once, then disabled it again. Should vaPlanView be placeable in Layouts in the future, annotating in Layout space is preferrable, so this feature might make less sense.)
  8. Automatic updating seems to be broken.
  9. Curves etc. that were part of a brep could optionally be grouped for convenience. I know that’s a problem because the vaPlanView is a block at first, and groups inside blocks are unsupported in Rhino as of now.
  10. vector drop shadows

Since I also work with Revit and ArchiCAD, and both have a strong 3D->2D vector engine, I wonder if you are ambitious enough to beef up what vaPlanView/sectionView can do, so Rhino+VA can become a serious competitor here, or maybe even surpass them in some ways?
These programs would not be taken serious, or be successful without such a vector engine. It’s central to any BIM workflow. Nobody in his right mind wants to drag along 2D plans after the 3D model manually without the necessity. However, it’s still unavoidable in Rhino/VA because of aforementioned shortcomings.

Much has been said about all this, and Rhino8 will bring some improvements o.o.t.b, but I see the best chance of Rhino getting better in this regard (in time) if VA took up the effort.
Thanks!

2 Likes

Hi @Eugen,

We have plans to remove vaPlanView and vaSectionView commands in VisualARQ 3 and convert them to an “Export to PDF” and “Export to DWG” commands, just like Revit does. I think it’s not possible in Revit to have a plan/section view object inserted in the document, it uses viewport and details, which is also available in VisualARQ. Moreover, most of your requests work fine in VisualARQ real-time plan/section viewport/details.

Can’t talk about ArchiCAD, as I don’t know how it works regarding 3D to 2D export.

Plan and section views are slow objects that use Make2D as 3D to 2D engine. It computes hidden lines using real surfaces, computing booleans, which is slow, and some objects are not supported, like solid hatches or texts. Real-time viewports (like Revit) are implemented using a similar approach used in display. Instead of using real surface, their display meshes are used, which have less precision (jagged arcs, etc.) but are much faster to compute, and solid hatches can be easily clipped.

Our target is to have a fast and good DWG export, as in-document plan and section view makes the document big and slow, as thousands of curves must be added to the document.

Enric

Enric

1 Like

Why did I bring up this whole topic again: since there’s no DWG export from layouts (yet), and I had to deliver them, I needed to fall back to these ‘oldschool’ tools, and ran into the usual problems. No hatches, ouch! (Copypasted them over in a second step).

This layout DWG export option was missing bitterly in the last years. Looking forward!

Trying to understand this display thing better: it’s clear that nurbs get coverted on the fly to a fast mesh version which then gets sent to the graphics driver.
Questions:

  1. What does the PDF print from a layout export? This mesh representation?
  2. What is the DWG version supposed to export? The’ meshed’ version will be unusable, since people will need the actual nurbs geometries’ outlines (polylines) in (stpd old) Autocad, and if so, that’s pretty close to what vaPlanView does (if it could sit in a layout).
  3. Will another special display mode à la ‘Hidden’ be necessary for DWG export?
1 Like

Hello,
Enric, I would like you to keep these vaPlanView and vaSectionView, both… all my workflows now rely on them, as dynamic, real time views are a bit to heavy to maintain… As you say, export to DWG is a must, but keep these two solutions, please :wink:
Cheers, Jaro

1 Like

indeed !

It’s doable, but not to forget the mess with layout/detail layer visibility. You know: model space layer visibility overrides layout’s/detail’s (being worked on now, as i gather).
Also, (custom) display modes in the details are not persistent across different rhino installations, which they should be. They should be saved with the scene.

So, the basic idea to rely on layouts for DWG/PDF export is preferable, but there are still a few quirks involved which need to be fixed on the McNeel side.

I also do see this as a problem!

However, there is one thing that does not work on real-time sections that works fine with vaSectionView, jagged sections. If the jagged section “parts” are not parallel, the vaSectionView does a great job of getting the viewing direction right for both parts of the section line. In the real-time section you can only have one viewing direction. This is a limitation that I don’t think can be overcome, the camera in a viewport can only look in one direction.

Although, having section/plan objects is slow, we have many use cases for them, beyond the above, for example adding hatches over the plan/section/elevation object is easier in 2d.

I don’t understand. Both, real-time sections and vaSectionView will produce identical results for a jagged section line. If you find a case that is different, then there is a bug.

Please, can you post here a sample so I can take a look?

Enric

Take into account that in vaSectionView, there is also a single viewing direction, and not parallel segments will be projected to the viewing direction.

I think you are thinking in unrolled sections, which is a planned feature for VisualARQ, that will support also non-linear segments (nurbs or arcs).

Enric

1 Like

Hello @enric

Here is an example of what I am saying. It is a project one of my colleagues is working and by coincidence he came into this issue last Wednesday.
Maybe jagged section is not the right term, the section line is made of two non-parallel segments. In the image bellow the section object is selected and you can see the dashed lines with the direction perpendicular to the section line on both ends of the two segments. The viewport on the lower left is the vaSectionView object, while the viewport on the lower right is the real time section.

As you can see, on the right, although the section is perpendicular to the stair and the sectioned walls, in the real time section we can see the projection of the stairs steps and the wall end. On the other hand, this does not happen in the vaSectionView.

I can send you the file if you want to have a look.

Best

Hi @Filipe_Brandao,

Did you set the view direction perpendicular to the first segment? (first segment is always the one of the left when viewed from the section front).

Again, both sections should be identical. They will be using exactly the same clipping planes, and if the viewport has the same viewing direction, the projection must be identical.

Maybe I’m missing something and you’re right. Can you please send me the file to visualarq@asuni.com?

Enric

I didn’t know that it defaulted to the left of the section. I assumed it got the direction priority from the direction of the section line (start to endpoint), which I find more intuitive, coupled with the direction of the dotted line. In fact, this is what is apparently happening with the vaSectionView object. If I don’t make the dotted line perpendicular to the right section line then the view direction is parallel to the left section line.

Best,

Hi @Filipe_Brandao, it seems like the issue that makes some geometry do not appear in the section view is caused by a boolean error in Rhino. Enric has reported it here: https://mcneel.myjetbrains.com/youtrack/issue/RH-72023

Hello @fsalla! Thanks!

@enric -
FWIW, when you manually trim the surfaces in the file that is provided in RH-72023, you’ll see that there’s a surface that is collapsed to a line. Rhino doesn’t flag it as a bad object, and, possibly, the Boolean operations can be made to deal with that, but the real bug might be an operation that gets an object into this state…


-wim