Jump to content

mmyoung

Member
  • Posts

    155
  • Joined

  • Last visited

Posts posted by mmyoung

  1. Ha! brilliant.

    Thanks for that ? lovely thinking.

    I had been hoping they would've built in something for single-segment polys. Seems logical, because you can build wireframes in space with them that are handy for a bunch of things, shade-tree surveying, project envelopes, on and on.

    This function would realize the poly is a single segment, then tell you how long it is. They already know this (the information is there), I'd just like it exposed. Wouldn't that be cool?

    Probably a way to write a script for it...

  2. I cannot snap a control point to a surface, either a planar extruded surface (the side of an extruded rectangle) or to a NURBS surface. I CAN snap NURBSpoints to edges, vertices, and to contours I generate, but that's it.

    Is this what you needed?

    re/ your "rant," it does seem to me the tech support folks at NV are not deeply familiar with 3D, especially NURBS yet. I don't see many users pushing the app like you do, so I'm not sure they get much call for it, though that will need to change now that more of us are working that way.

    There ought to be an option to leave the controls visible, and snapping to a surface is actually the reason NURBS surfaces were invented, I believe. Will Rhino do that? I've not worked with an app that will do it yet.

  3. Hi Kaare,

    THANK YOU! for that setting (Nearest Point on Edge). Don't know how I missed it. I had some tedious workarounds.

    ..

    ? I cannot snap to the control points of a surface or NURB Spline.

    The control points are visible to me only when I'm in 3D Reshape.

    I can snap the control points to existing objects (help objects), but not the other way 'round (can't snap existing objects to them, because I can't see them).

    v2008 is no longer loaded on my machine. Were the control points visible in 2008 when you weren't in 3D Reshape??

  4. Kaare,

    A question for you (unfortunately not an answer to your problem above):

    If I want to snap one NURBS to another in 3D, it seems possible only at certain points along the curve. This is not the case in other apps (Shark, for example).

    Is this the same for you? This also seems like an oversight, but it seems important.

    --

    Also, I am not quite sure I understand your dilemma above; could you elaborate a little more. I know this kind of thing is sometimes hard to explain...

  5. Wouldn't it make sense to have an editable field in the OIP for length of a singe-segment 3D poly and NURBS (degree-1)?

    If we want to edit the line length, we could do it directly in the OIP.

    A single segment 3D poly or 1-degree NURBS has vector magnitude information in it that we should be able to edit. If you are drawing directly in 3D, this would be very, very handy.

  6. Hi Peter,

    Actually, I *am* kind of embarrassed by what I wrote. It ticked me off that there were these booby traps, and I had to take a few swings at it, but it's not as bad as I made it seem. It could be documented better I suppose.

    I duplicated the 2D poly layer and really heavied up my important index contours, and I was all right, so thanks for that.

    But truly, yeah, the TIN generator is one heck of a powerful tool. Especially if you have to shoot the site radially ? adding the smoothing points really helps a lot. My "adventures" notwithstanding, I'm saving days of work.

    Thanks for the understanding reply and the help ? very appreciated.

  7. As you are converting a diabolically complex contour map from 2D polygons to 3D contours...

    Why doesn't VW put the 3D contours on another layer or in another class as they are made?? That would make sense, I think.

    And why remove the color and line weights from the graphic you are converting? There are a lot of keys to the geometry you are converting in the colors and line weights which make it simple to orient yourself. Then VW takes them away and asks you where you are...

    There are work-arounds, but geez, they didn't make this easy.

    If you are interrupted in the process, supposedly you can resume by starting over on the red contour ha ha ha ha ha

    Very hard to find the red contour! And your original 2D contours are now all the same color and line thickness, they are overlaid by the new 3D geometry.

    So if you goof when you resume, and you don't have a couple virgin backups, you are cooked ? all your original geometry is doubled up, and you can't see where this has occurred.

    And if on completion you don't tell VW to delete your orignal 2D polys (because maybe you need them), they are left on the *same layer* as the new 3D contours ? very hard to separate well!

    -- If the newly generated 3D geometry could be put into a separate class, a redline class for instance (an option to do this?), or on a new layer, that would help.

  8. MichaelK ? yeah, I've got 3D Conversion res set Very High. Really, as Kevin McAllister mentions, there seem to be different rendering protocols for Design Layers and Sheet View. I think they got 'em backwards ? I'd rather see the "good stuff" in the Sheet Layer viewports.

    I did send this in to Tech. It just seems wrong, not a bug so much as a goof.

  9. The file I had a problem with (above) is buggered... looks like it didn't make it across the divide from 2009 to 2010.

    *However*, this thread made me drill more deeply into resolution issues.

    Something does appear to be wrong here ? or at least I am getting reproducible errors on this machine (see attached PDF and VWX):

    I generated the objects like this:

    - 26' circle

    - Offset & duplicate inside 6"

    - Trim above the centerline, as shown.

    - Select everything and Compose.

    - Extrude 24"

    - Modify / Rotate / Rotate Left 90? to erect it upright.

    - Set to perspective view and adjust with Translate View tool.

    - Render in Custom Renderworks (Sht-4 in attached PDF); options for curve detail Very High and in Hidden Line with Line Render Options / Smoothing Angle set to 0?.

    The rings behind the arch were made the same way, duplicated upward using Move by Points.

    The model renders nicely in the Design Layer.

    ** BUT when I make a viewport (Sht-1, Sht-2, and Sht-3), Custom RW and Final Quality RW both render an obviously faceted profile!

    You can treat this somewhat by making a Foreground Render with Hidden Line (see Sht-2), but it doesn't actually solve the problem.

    ** Why ought it render well in the Design Layer and badly in the viewport?

  10. When I render in wireframe, arcs are fairly smooth. In hidden line, they are faceted (see attached).

    If I use hidden line as foreground rendering in a viewport, it looks terrible.

    It seems like I figured this out once before,

    but I'm stumped now.

    How do you get the facets out of a hidden line render?

    I have Custom Renderworks Options/Geometry set to Very High curve detail.

  11. Thanks Miguel.

    Are there directions somewhere for how to structure the metadata for shapefiles? How to set up .dbf and .prj records? So far I haven't been able to find information I could easily make sense of.

  12. I'm new to DTMs (which ought to be obvious from my question):

    When I import a stack of shapefiles, they align perfectly with respect to one another, but they can be offset from the "paper space" origin (0,0) of Vectorworks by miles.

    It's all right; I move my page over to them.

    But how is the origin of a shapefile established?

    Is there a mapping protocol that determines where (0,0) is? Is there some sort of master grid it references?

  13. Thanks Ozzie,

    Yeah I forgot you had Landmark. Great comments, helps me understand what you are seeing and what sort of things arise in this kind of work.

    I'd like to get Landmark. Seems like a lot of application for TINs, as well.

    all the best?

  14. Ozzie,

    If I had Landmark (as Alfresco does), I would make a TIN (a "Triangulated Irregular Network") with my surveying points, from which it is much easier to make a plausible surface from a point cloud than it is with NURBS surfaces.

    Without LandMark, as you suggest, you'd like to make the model using NURBS surfaces. Bunch of ways to do that. I'm not precisely sure what you are seeing in your imagination, so here's a guess.

    I began by tightening up the existing model to eliminate overlaps and address the low corner of the intended driveway. Executive decision... ha ha. It's easier to edit NURBS curves that surfaces, so do as much editing as you can before surfacing your curves.

    You are aiming at making a set of NURBS surfaces which correspond to your surveyed data points. To do this with NURBS, I would lay in my data points using 3D loci: Pop in a locus, then edit it coordinates in the OIP. Draw NURBS curves through the loci and make surfaces.

    You could also make interpolated surfaces, which are guaranteed to intersect all the data points, but getting the UV lines oriented well to your data points would be difficult.

  15. Ozzie,

    If I had Landmark (as Alfresco does), I would make a TIN (a "Triangulated Irregular Network") with my surveying points, from which it is much easier to make a plausible surface from a point cloud than it is with NURBS surfaces.

    Without LandMark, as you suggest, you'd like to make the model using NURBS surfaces. Bunch of ways to do that. I'm not precisely sure what you are seeing in your imagination, so here's a guess.

    I began by tightening up the existing model to eliminate overlaps and address the low corner of the intended driveway. Executive decision... ha ha. It's easier to edit NURBS curves that surfaces, so do as much editing as you can before surfacing your curves.

    You are aiming at making a set of NURBS surfaces which correspond to your surveyed data points. To do this with NURBS, I would lay in my data points using 3D loci: Pop in a locus, then edit it coordinates in the OIP. Draw NURBS curves through the loci and make surfaces.

    You could also make interpolated surfaces, which are guaranteed to intersect all the data points, but getting the UV lines oriented well to your data points would be difficult.

  16. Something else along the lines of the Zaha shape thing, so folks who want to fool around more in 3D can get an idea of a couple of the tools and ideas. Much more in there, but this is a start.

    >> Step through by clicking on Saved Views and going through the numbered sequence.

    The primer was a little large; I divided it half so it would post.

    The first section, frames 00 through 07, concern unfolding developable surfaces. The unfold tool is usually very robust and accurate (we've built things with it using a CNC plasma cutter). On this model, one of the surfaces wouldn't unfold, but I left it the way it is...

    The second section, frames 08 through 12, takes the same model and elaborates on it a bit, filleting it then extracting contours and isoparms and using them as extrusion paths.

×
×
  • Create New...