Jump to content

Danielj1

Member
  • Posts

    106
  • Joined

  • Last visited

Posts posted by Danielj1

  1. I just tried this (and filed a bug report too). Here's what I found: a two-segment NURBS curve path (meaning L-shaped) won't display the problem, but a three-segment one (U-shaped) will. In addition, adding the fourth vertex (meaning the third segment) sometimes causes the extrusion's profile to be located at a different angle relative to the path.

    So in this particular case, I'd use more than one path in sequence, each with no more than two segments. Let's say you're creating a three-segment path for your extrusion, similar to the one shown in your illustration. I'd consider making two, two-segment paths (for the purpose of this first modeling step place them side-by-side without overlapping). The first one modeling segments A and B, and the second one modeling segments B and C.

    Once the models are done, ungroup everything and delete segment B from one of the models. Then move the remaining segment C over the first model, snapping to the appropriate corner, group it all, and you should now have a three-segment model with the corner joints represented properly.

    Dan J.

  2. Jeremy,

    Seems to me that the original 2D polygon used to create the extruded "gable end walls" may not be closed.

    Double-click on the extrusion to access the original polygon, then in the Object Info palette make sure the Closed checkbox is selected; see if that works.

    Dan Jansenson

  3. A technique that has worked for me is avoiding the use of rounded shapes for the tiles, and instead using faceted forms with a series of flat faces creating the roof tile. This reduces the rendering time and memory requirements dramatically with hardly any impact on the quality of the final rendering.

    In most cases the camera's viewpoint is sufficiently distant from the project so that the difference between faceted and rounded tiles can hardly be observed.

    Description and images here:

    http://tinyurl.com/yaol2fe

    Dan J.

  4. If it's a regular VW2010 with Renderworks (not educational; don't know about that) you should be able to install the program on two computers. However you will only be able to operate them one at a time, and not both simultaneously.

    You didn't indicate the nature of the problem you were having installing the software on the second computer. That computer needs to be connected to the internet, I believe, in order to authorize the full install.

    Dan J.

  5. I have to say this is a mystery to me as well: on my MacbookPro with OSX10.5.8 I can't replicate this problem at all (I used the replacement file). Tried with OpenGL (with Detail set to Low), CustomRW (with Ray Tracing on and off), and other combos all to no avail; I'm not getting the jaggies. I'm just wondering if it is a video driver issue...

    Dan J.

  6. Donald,

    The problem you're describing is, I believe, due to Vectorworks behavior that has changed since the book first appeared. Thank you for bringing this to my attention.

    The task described here uses the Clip Surface command to trim away unneeded portions of a 2D polygon containing an image-fill-based rendering of a building. To do this with the current version of Vectorworks, you must in essence first create a kind of masking doughnut sitting on top of the first 2D polygon. The doughnut hole will be in the shape of the building's outline.

    When you select both 2D polygons (meaning, the underlying one containing the building's image, and the overlying one functioning as a doughnut-shaped mask) and then invoke the Clip Surface command, the doughnut will cut away the underlying 2D rendering/polygon--except for the portion that can be seen within the doughnut hole itself.

    At first it will appear as if nothing has happened. But when you delete the top-most doughnut, what remains is the underlying rendering, trimmed to the shape of the doughnut hole; that is, the building's outline.

    Now, the challenge, a small one really, is how to draw the doughnut with the properly-shaped hole representing the outline shape of the building rendered beneath. One simple method involves first drawing a 2D polygon on top of the building rendering in the shape of the building's outline, and then drawing a second, large 2D polygon covering everything.

    Send the new large polygon to the back and select both the new large polygon AND the one representing the building's outline, and then invoke the Clip Surface command. The large 2D polygon will now have the shape of the desired doughnut.

    Move it back to the front, make sure the doughnut and the rendering polygon are both selected, and then invoke the Clip Surface command again to trim away the unneeded portions of the rendered polygon.

    Whew, I'm hungry now.

    As others have mentioned, all this can be done much more easily (if the outline shape is not complex, that is) by using the Clip tool, with the Removes Outside Mode selected, and the Clipping by a Polygon Mode selected too. But you only have one shot (except for the use of Undo), and you cannot adjust the shape of the marquee that the tool creates. You can, however, use this too incrementally, cutting away portions of the underlying image piece by piece for greater accuracy until arriving at the desired shape.

    Dan J.

  7. The author of that book occasionally lurks hereabouts to see what's going on...and is actually currently in the process of working on an update to the book. So feel free to email him directly if you wish, although his day-job workload may preclude a quick (or any) answer...danielj1@earthlink.net.

    As for the answer to the poster's question, yes the Clip Surface command is still there and available: Modify>Clip Surface. Of course, as others have noted, we also have the clip tool as well, which is extremely useful in certain situations.

    Personally I have a bias toward the command in these situations because the Clip tool requires drawing a clipping shape using its marquee function, and in complicated clipping situations it is sometimes necessary to first draw a 2D shape outlining the clipping area, and then trace over it with the Clip tool marquee. Using the command allows you to use the 2D clipping shape directly, without needing to trace over it with the marquee.

    But it's really a matter of taste and habit, and I'm glad we have multiple options to suit our individual work styles.

    Dan Jansenson

  8. I often use layer links because they can provide some important (to me) capabilities not available in DLVPs.

    Suppose I'm in the midst of a quick design session for a three-story building, and I want to see what the building would look like without the middle story, and with the top level sitting directly on the first. With a layer link arrangement, I unlock everything, click on the middle layer and hit delete. Then grab the top layer and simply drag it down.

    With a DLVP, I have to go to the OI Palette, hit the Layers button, select the middle layer and make it invisible, and then hit Edit. Finally, with the layer selected I enter the Edit Viewport Design Layers dialog, and alter the Z height; click OK and then click OK again. Only now is the result visible. If I lowered it too much, I must start the whole thing all over again till it's aligned properly.

    Another example: suppose I'm working on the model, and would like to separate out the three different levels, and show them with 10 or 20 feet between one and the next (an exploded view). This could be useful when trying to understand vertical HVAC chase locations, for example. Again, with a layer link, I unlock everything, switch to front view, select the desired layers, hit command-M and enter the distance. That's it.

    One more example. We can crop individual layer links; this works even in 3D; even in_rendered_3D, which is something really phenomenal (go to Help for a very good description of cropped layer links). This means that in a multiple-part linked layer model, I can switch to an isometric view, select the middle layer, crop it to display only a portion of it, then render the entire wedding cake. If you have structural elements that span the three levels you can make a portion of the exterior building envelope disappear in order to reveal the inside structure this way (as long as the structural elements are not on the same layer, of course). As far as I know you can't do this cropping thing with DLVPs, since a DLVP crop affects the entire viewport and not only individual layers within it (although you could accomplish something similar with classes, possibly).

    I'm only saying that layer links have certain advantages not shared by DLVPs; each has a different function. Having both available expands our horizons considerably, and there's still a viable role for the layer-link capability beyond what's already been described in previous messages.

    Dan J.

    Dan J.

  9. You can't compare the Australian and US versions of VectorWorks, because the Australian version has tons more content, tools, and localized stuff that doesn't exist in the US version, as well as a local distributor who provides intense and high-quality tech support. If you're going to compare to Allplan (or any others) you'd need to make a direct comparison of the specific content and capabilities of the two--including your localized versions. Using the US version as a reference will lead you astray in your price calculations.

    Dan J.

  10. I'm not sure this will work, but it may be worth trying.

    Make the glass component invisible (by class selection, say.) Then prepare simple extruded shapes to represent the glass in the rendering (textured properly) and place them BEHIND the windows a small distance: say, 4-6 inches. Or perhaps just behind the muntins, rather.

    I wonder if this would provide the right amount of muntin definition you need.

    Dan J.

  11. I've been seeing similar corruption happen with my neighbor's setup, but not with mine. He's using a G4 powerbook with 512mb RAM, and I'm using a G5 iMac with 2.5 gigs RAM; that's one difference. Another is that as a landscape architect, he routinely inserts large amounts of DXF- or DWG-imported data into his drawings; often very large ACAD drawings, such as extended topographic surveys. I've been working with him on one project doing a small building to be inserted into his landscape designs, and all the files I gave him became corrupted after he touched them. It turns out that he copied-and-pasted objects from his larger design drawings into mine, somewhere along the process. I believe that those larger drawings of his included topo surveys originally prepared in ACAD.

    My suspicion is that there's an odd, but specific combination of things that come together to corrupt a VW file: the insertion of large amounts of DXF/DWG data, as I've indicated, along with a very small amount of memory. Also there's something odd going on with his in-office network that I suspect may be playing a role, perhaps the backup issue mentioned earlier in this thread (he hasn't upgraded to 12.5.1 yet).

    I've been playing with 12.5.1 and its predecessors for a while now, both on the office iMac, and on a Macbook at home (with 2gigs RAM) and haven't experienced the kinds of file corruption I've observed with my neighbor, so my theory is that the items I've mentioned above are directly responsible.

    Dan Jansenson

  12. The problem traditionally has been that any white 3-D object

    often renders as grey, and appears excessively dark (I think it

    actually renders accurately, it's only a perception issue I believe).

    Adding and adjusting lights is too time-consuming and often yields

    unexpected results (and increased rendering times). There's no texture

    combination (meaning with reflectivity, etc.), to my knowledge, that

    can lighten the appearance of such a 3-D object in a controlled fashion

    other than Constant, as Biplap suggests, and that one is uniformly white with little ability to accept subtle shadows. After some experimenting, I think

    there's a way to do this, and thought I'd post it here for those who

    are working with radiosity.

    For a ceiling: make an extruded 3-D object with a white fill. Duplicate

    it in place, and nudge the duplicate down by a couple of pixels, just

    enough to see in close-up wireframe. Create a Constant texture, and

    give it about 80% transparency. Apply this texture to the nudged

    duplicate 3-D object. The transparency of the Constant texture will

    allow the underlying object to be seen as a solid, but the 20% texture

    visibility will create a lighter appearance, while allowing subtle

    shade and shadow effects upon its surface.

    An example of such a texture can be seen here:

    http://www.danjansenson.com/whiteceiling/

    Note that the ceiling above and the wall to the right both have white

    fills. It's only the ceiling duplicate that received the 80%

    transparent Constant shader. The white-ness of the ceiling can be

    controlled by altering the amount of transparency in the texture.

    Dan Jansenson

  13. You could also try the crude/lewd method of using the Render Bitmap tool. View your scene in your desired perspectives, and use the RB tool to create a bitmap rendering that stays on the sheet. You can copy/paste the bitmap anywhere you like and resize it appropriately. Any sheet containing rendered bitmaps can be exported as an image using the export image file method.

    Dan J.

    ___________

    http://www.imageprops.com

  14. "I adamantly contend, to the objections of all my friends who do it their own way, that the fast way to work in 2D in VW is to, with only a couple of exceptions, use layers only for scale, use only as many as you need to account for the scales you need on each drawing, keep your class list simple, and resist the temptation to think that when you categorize the world around you it is somehow more intelligible or easier to deal with."

    Donald describes, almost verbatim, the way I work too, with the only exception of adding layers to account for different floors in a multi-floor project. Adding layers and classes only as you need them, and then keeping things as absolutely simple as possible (including layer and class names) not only makes life a lot more easygoing, but also makes it possible to open a file years later and understand what the heck you've done.

    Dan Jansenson

    ________________

    www.imageprops.com

    the Renderworks Recipe Book

  15. One more thing. Note that a moir? pattern that is visible on a monitor may not be visible when printed (and vice versa). So if you've been working hard to eliminate this interference in your images, it might be worth making a test print to see how it looks on paper before spending any more time on the problem.

    Dan Jansenson

  16. This sounds, indeed, as if it were the well-known moir? problem with brick textures, common to all rendering programs, as well as digital photography and imaging of various kinds. The problem is the result of the interaction between the "frequency" or resolution of the brick image, and the resolution of the monitor or output device, and occurs not only with brick patterns but with other repetitive high-frequency patterns as well. There's a great deal of discussion on this topic to be found on the web, including:

    http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=003zye

    and

    http://www.wfu.edu/~matthews/misc/DigPhotog/alias/

    and

    http://www.formz.com/forum/forum_archive/forumB/141.html#872166382-22815

    The only solution is to alter the resolution of the image pattern (sometimes to a very great extent) or of the monitor or output device.

    In my work I tend to create both high- and low-resolution textures in part because of this very problem.

    Dan Jansenson

×
×
  • Create New...