Cloud Hidden

If you are framing a wall with peaks, the wall framer will put a stud on each side of a wall peak. There seems to be no way to stop this. My walls are framed to a dome. I use many wall peaks to match the curve of the dome--as many as 1 per running foot. The framer is putting one stud on each side of each peak--ignoring the specified stud spacing--which means there are many, many extra studs. This is not how we'd frame such a wall, and vastly overestimates the lumber requirement. We'd follow the standard spacing and not let it be dictated by the wall peaks. The framer is especially useful for these walls, since it's hard to do a manual material takeoff on a curve-topped wall, but adding studs at each wall peak, rather than just following the specified stud spacing, makes it a whole lot less useful. Any advice on how to address this? Ideally there'd be a checkbox labeled "Add Full Stud at each Wall Peak" or "Use the specified spacing even at wall peaks" or similar. But until then, is there any way to get a 2D, 3D, and worksheet that reflects the right spacing? [ 09-12-2003, 01:53 PM: Message edited by: Cloud Hidden ]

I've become suspicious that GetWallPeak does not work. Could you, Alexandre, or anyone else, run this script and see if it does or does not show you the coordinates of the wall peaks in a selected wall: code: Procedure WallInfo; VAR pX, pY, zValue: REAL; hWall : HANDLE; i :INTEGER; str :STRING; BEGIN hWall:=FSActLayer; {Handle to the First Selected object at the Active Layer} i := 0; while i < GetNumWallPeaks(hWall) do begin GetWallPeak(hWall, i, pX, pY, zValue); str := concat('GetNumWallPeaks ', Num2Str(1, GetNumWallPeaks(hWall)), ', i ', Num2Str(1, i), ', px ', Num2StrF(px), ', py ', Num2StrF(py), ', zValue ', Num2StrF(zValue)); AlrtDialog(str); i := i + 1; end; END; Run (WallInfo); [/code] Edit: Upon further investigation, it always seems to return the coordinates of the last wall peak, regardless of the index parameter. Edit: And why is this a 0-based index? If you have 1 peak and you pass 1 as the index, you get 0's as the peak. Pass 0 as the index and you get the correct peak. I thought Pascal was more 1-based than 0-based. [ 09-12-2003, 09:59 AM: Message edited by: Cloud Hidden ]

Thanks for the examples. Did you "roll your own" trees, or use xfrog, or...?

Well, darn if that doesn't work! Thanks. I was looking for that parameter directly and even via other calculations, and never paid attention to 3DCntr. It's great they have VS, but frustrating at the same time be/c of missing functions like that one. Another I'm grappling with is that you can add a wall peak and get a wall peak, but you cannot (as far as I can tell) set a wall peak or delete an individual wall peak. Sometimes it's hard to follow their reasons... [ 09-12-2003, 01:20 AM: Message edited by: Cloud Hidden ]

Still stumped after trying a bunch of things...any idea how to get the "Delta Z" and "Bottom Z" values for a wall object?

Some of the conversations go back to December and relate directly to features being promoted in the new release as newly enhanced. As I said, I appreciate the efforts--I was a Mac developer, I can relate--but when the same bugs persist through multiple releases, it's frustrating and you start to question the time spent to create and submit the files.

Don't misunderstand...I'm really happy NNA keeps working hard and putting out new releases. But it's frustrating to report bugs, prepare sample files documenting them, email the files when and where requested, have the problem verified, have email conversations with people with recognizable names on the tech side, and then not see the problem fixed. And it's not all arcane stuff either. * I have two polylines that added together (Add Surface) produce a crazy result. * A simple clip surface from a polyline that takes 4 min + to run (if at all) and then is wrong. * Blend edge that won't. * Shell solid that won't. These bugs cost me hours to work around on each project, and I don't know what else to do differently to get them fixed.

quote: Originally posted by Cloud Hidden: look forward to trying my trickiest ones--saddle connections on intersected domes. NNA has had a test file that we know wouldn't work. Anxious to see if it now does. Well, bummer. Still doesn't work. Have to use the much longer procedure of subtracting away the saddle area and lofting the saddle at the radius I need. And the message it gives is really frustrating....says to try tangents when that's already checked, and to try a different radius, without a clue on what other radius might work. Bummer. And darn it, I get the same ol' non-result with shell solid, too. Just have 3 intersected hemispheres added with the base removed...kinda like a mickey mouse hat. They won't blend or shell. [ 09-10-2003, 11:16 AM: Message edited by: Cloud Hidden ]

Can't say I've ever figured out when I needed tangents or faces. I usually just try every combo till one works or I run out of combos. I've historically had big trouble with this tool, and look forward to trying my trickiest ones--saddle connections on intersected domes. NNA has had a test file that we know wouldn't work. Anxious to see if it now does. Also would love to have a way better than trial-and-error for setting those options. But I'll settle for predictability.

And fwiw, this is a document preference in VW 10.x. Whether as a plug-in or a pref, if you do curved surfaces, it's the difference between the product being usable or barely usable for sections/elevations/wire-frames. I haven't figured out all it's tolerances, and usually just set it to an arbitrary number higher than 45 and less than 90.

I created a new file. ~4"D circle. Extruded 5". Blend w/ 1" R and Select Tangent Entities and not Select Faces. (Don't know if either of those two would make a difference.) It worked perfectly the first time, which is a nice change of pace for me. Dressed it up and here it is: So, at least there evidence that it does work under some conditions.

I can get some wall info through Get3DInfo, and some through WallHeight, but neither tell the bottom Z value of the wall, such as would show in the OIP. Is that a function I'm just overlooking? As a corollary, for a round wall, how do you find out the curvature...radius, center of curvature, etc. Thanks. [ 09-08-2003, 08:17 PM: Message edited by: Cloud Hidden ]

"GetParameterOnNurbsCurve (h:HANDLE;pointX:REAL;pointY:REAL;pointZ:REAL;VAR?parameter:REAL;VAR?index:LONGINT) :BOOLEAN ; Description: Given a NURBS curve handle and a point (in world space), this function returns the parameter of the point obtained by projecting the input point. The function also returns the index of the piece in the piecewise NURBS curve on which the projected point lies." Don't know if this works on surfaces, too, but a handle and a point is what I want to provide. But I don't understand "returns the parameter of the point obtained by projecting the input point" What is the parameter of a point? And in what direction is the input point projected? Seems like an input point could be projected in infinite directions.

What's the best definition of U and V on nurbs? I see references to them in the documentation, but don't know what determines which direction is which.

Used the wall framer for the first time, and even though I had to manually curve (segment) all the wall tops in my dome, the framing calculations are the cat's meow. Very cool. Only thing I can't figure out from the worksheets: what things go into the calculation of "sill stock"? Thanks.

Just looking for a point in the right direction... Have a point on the ground plane and want to find the z point on a nurbs surface at that x and y. Any ideas? I've read every function definition, but between nurbs nodes and knots and u's and v', I'm getting more than a little confused...

The skydome creates equal brightness from all the lights. What settings have worked best for y'all for the skydome versus the main directional light--the "sun"? 50 for sky, 100 for sun becomes too bright. 50/50? 75/75? I'm experimenting, but maybe there's a good rule of thumb running around out there...

Thanks Dave. I just saw that I created images with the bottom of the image not at the bottom of the file. Doh. I'm posting this progression so that the difference from step to step--VW textures and trees, then imaged grass, then image props--is clearer. Improvement each time. Gotta figure out the best way to clean up the tree at it's bottom left edge, but still, an improvement. Gotta create more interest along the horizon and try the skydome yet. But step-by-step, learning and improving. Had fun yesterday driving all over town getting pix of green stuff. Less fun trying to isolate the green to keep from the background. Getting there. As always, any ideas welcome. ### W/ skydome 50% sky 75% sun. Overlayed on a background and cleaned up as much as I had time for. Still gotta work on things like shadows going in the same direction and extending shadows to the background from the foreground objects. But certainly getting close enough that I can get by with the products I have for now...VW, Photoshop Elements, and GraphicConverter... Next challenge....QT Virtual Reality! We gonna see that sometime, Dave? [ 09-02-2003, 04:23 PM: Message edited by: Cloud Hidden ]

Is it true that Send to Surface will set the elevation based on a corner of the prop (bottom left) rather than the center? My prop trees are levitating over the dtm and I have to individually adjust them. It'd be nice if there was a way to mark the point of the prop to use for setting ele. Second, what's the way to create a new instance of a prop? Say I create a new image prop, which places a copy of the tree on the drawing. Later I erase it. Then I want to use it again, but don't have one to dupe. The prop is in the Textures area of the resource browser, but not in the PIO area. I checked the PIO option when creating it... Third, the "softness" sliders don't have a labeled scale. Is the left soft and right hard or vica versa? When creating the prop, I encounter 3 different softness sliders: once with the image, then image color (I think), then transparency mask. Do they all do the same or different things? Any redundancy? Fourth, seems that having anti-aliasing on in photoshop when creating images (trees) hurts rather than helps, be/c it creates a slight off-white halo around the image that survives the transparency mask. Does this sound right? Creating the image with hard edges is the way to go, or is there a better approach? Thanks... [ 09-02-2003, 10:12 AM: Message edited by: Cloud Hidden ]

Dave, Kristen, thanks! Those are the kind of practical suggestions that I can't find in the manuals, but are a lot more useful. I tried some of them... Good eye on the grass size. Used the lawnmower tool to knock it down by 4 to realistic size. Was worried that I'd see too much of the mosaic effect, but it worked out ok. When I tried to soften it though, the repetiveness of the pattern became lots more obvious. Maybe I need to add a bump as I soften it? Any rules of thumb for what makes a good bump? Used the chainsaw tool to cut down the palm trees. Later the other trees will go, too. No time for that right now. The SUV presented two challenges, though in principle I like the changes. First, the glass on the passenger's side windshield is giving a reflection of the sky or something, but it makes it seem as though I used the wrong material. Don't know how to prevent this. And I had trouble with the Custom Selection menu command. The SUV is nicely designed with different pen colors for the different objects. But I couldn't get it to select all the blue objects (bumpers), for example, and that makes putting textures on all 386 objects much too difficult. Did some, but not the body. Any ideas on why they wouldn't select despite matching the choice from the color palette?

Kurt, the only way I know to create a seamless texture is to mirror the image. Is there another way? Here's the prelim results of yesterday's experiments with grass. Grass texture. Before and after: Also followed Dave's advice from another thread to turn off ambient light. I'm pretty darn happy. Still have work to do, and open to suggestions. Haven't done the skydome lighting he describes yet. Gotta work on trees and stuff. And need to experiment more with sizes. My "grass" image is currently 7 mg. Don't know how far I can reduce that before the results degrade. Also gotta work on those seamless edges. Also could get a better grass image that doesn't include the clippings that will make a repetitive pattern more obvious. And haven't figured out good settings for a bump--anything I tried made it look worse. Still, huge improvement, and much closer to where I need to be. Further suggestions?

After going through all the considerations on C4D and alternatives, and making the decision to not yet make a decision, I'm gonna spend a bit more time seeing what I can accomplish within VW. From what MikeB and others write, it seems to come down to improving the textures I use. In all the playing around with textures, I'm sure I'm missing some useful tricks. I can find all the dialog boxes, and push all the buttons, but there are so many combinations of options that it feels hit-or-miss as to whether I get a more realistic result or not. Anyone have any guidance for getting beyond the basics? For example, I am really hung up (obsessed?) with getting fuzzy looking grass for the lawns surrounding the houses, and I can't figure it out. Should I take a digital photo of the prettiest lawn in the neighborhood and use that as the basis of a texture? What are the options I should be looking at to yield the best results--trial-and-error has its limits. Thanks for any ideas for getting beyond Textures 101.

Even at 1'....heck, even at 2'.....I was getting a message that the interval was small compared to the range. But even at an inch, I'd still have a shadow line where the contour was 1" lower than expected. So I'll either offset the pads or use the triangles. Heck, maybe I'll even try to make a nurbs surface follow the curves of the ridgeline....that oughta be fun! Knowing that contour thing is an anomoly and knowing that cut-and-fill works better with rectilinear pads than curved ones, I can make the adjustments here for now. Hopefully, it's also good feedback to NNA for a future VW version.

"anomoly"....you once worked for Microsoft, right? --Just kidding! In my software company, things like this were called #&%*!$% -- I had never before used the 3D Triangle....guess I just liked the look of the 3D Contours for some strange reason....but I tried it here. With the single only change being a trip to AEC->DTM Processor->View Options and unchecking the one and checking the other, the model draws properly, with the top surface matching the pad site modifiers rather than sitting one contour level below them. Thanks for following up on this.

Don't believe so. When I created it, the actual highest elevation was 50' and lowest was -50', and both existing and proposed still show that. When I view the model in front view, I have a bottom contour that properly touches -50 on its bottom edge, but the top of the top contour only is 49. Send to surface places an object at 50' elevation, leaving a 1' gap. I can certainly set my pad's one foot higher than I want them to make it look right. Which raises another point--the gap is always the contour interval...1', 2', whatever interval I use for the contours. That speaks against having dropped the DTM by a fixed amount. Somewhere, it seems a contour layer is missing and I still don't grok DTM enough to guess where it went. On the cut-and-fill, I switched from curved pads and fences to rectilinear ones so I could more easily try your suggestions on tiered pads. Cut-and-fill no longer gave error message and seemed to work as advertised. The error only happens when I use the cubic vertex mode (didn't try the other curved modes) for creating pads.