Jump to content

Planning 3D enhancements


Recommended Posts

  • Vectorworks, Inc Employee

I am planning for the 3D enhancements for the next release. I keep on getting private mails about wishes and workflow improvements. Also we have discussed some of those in this forum.

Please post your top 5 wishes or annoyances on 3D Power pack functionality in reply to this message.

Thanks.

Link to comment

Hi Biplab,

here are mine:

1.....Reliability

2.....Reliability

3.....Reliability

4.....Reliability

5.....Reliability

Ok, ok, I'll be serious now....

There have been (more) recent discussions about the Dear-Old Fillet tool. Tom gave an informative precis on the complexities of 3D manipulation and why operations might fail.

My experience of this tool, though, (including in the VW11 demo) is that on duplicates of the same object, or on two objects made with an identical sequence of operations but in 2 seperate files, the Fillet tool will work on some objects, but not on others.

After repeated attempts to get it to work and having erratic, unpredictable success, the Red Mist starts to decend, and I start to feel that Ultra-violence on inanimate objects is a valid lifestyle choice. :-O

I'll leave off now and take some time to consider a more expansive response. ;-)

cheers,

N.

Link to comment

This applies to NURBS surface modeling....and boy are you getting close! I'm loving it!

1. Show points globally preference (interpolation points & control points)

When creating surfaces from curves, intersecting curves are important, and being able to snap to the interpolation points on another curve is imperative. Right now, you have to select a curve to see it's interpolation points (and only if you drew a nurbs curve by that method) and then use the cursor hints to snap to that point. Drawing curves using control points are more intuitive, but don't allow you to see interpolation points to snap to when creating curve networks.

At this point I usually use the polyline tool and then "convert to nurbs" so that I can see the interpolated and control points...BUT I do not maintain tangency on either side of the interpolated points so when I try to reshape the nurbs curve, I get "sharp" or "broken" points.

2. Lock points to first point (or maintain constant "Z" value into the page in orthographic views) when creating nurbs curves.

This pertains to creating planar sections for lofting surfaces (and creating curve networks). Planar sections for lofting are usually created in an orthographic view and in the course of drawing a nurbs curve section, you frequently have occasion to snap to other nurbs curves (if you can see the interpolation points) and right now when you draw in an orthographic view and snap to a curve, then finish creating your curve, the curve usually isn't planar because the points not intersecting the "snapped to" curve are on another plane. Now you could move the working plane to a interpolated point and draw, but if you had to snap to two points on a nurbs curve (for alignment reasons) then even that method wouldn't work.

3. Maintain same number of "points" when offsetting nurbs curves.

EXAMPLE: I create a planar nurbs curve with 5 verticies and then use the offset tool. I then get a curve with 22 verticies, making it tough to align "V" curves for creating surfaces.

4. Project curve to surface (and not trim or split)

This is more a convenience since it is can have a workaround somewhat. I would like to be able to project a curve onto a surface where the curve doesn't have to extend past the edge of a surface when in an orthographic view, and the cuve is an independent entity after the operation. I know you can project a curve onto a surface where the curve does not extend past the surface edge boundaries in ortho view, but you cannot extract this curve after the operation.

5. Better tesslation control over generic solids (or in general)

OK, there's probably not enough of us to justify this one, but here it is anyways. I'm using VW more and more for 3D modeling, but use other applications for animation and rendering....and to keep file sizes down and increase the chances for import success, I create generic solids when I'm ready to export. The 3D conversion res settings usually tesselate too high or not enough. It would be a convenience to have better control over this.

Link to comment

Top 5 wishes for 3D modelling:

1. Better 2D to 3D compatability: I'd like to be able to convert 3D nurbs curves into 2D splines and back again as I see fit. I'd like to have interactive weighting control by a slider or something like that. (I would wish to have weigting control of 2D splines as well.)

2. Some sort of 'hypernurbs' tool, that allows me to work with several nurbs surfaces at a time, keeping their edges stitched together.

3. I would like the extrude along path tool to become more tolerant of my design wishes. More than 50% of all attempts with this command fails. Also I would like a 'twist' function to be included, so I can twist the profile as it is extruded along the path.

4. The 'fillet edge tool' gives me error messsages all the time. In FormZ this tool is extremely tolerant, and will round almost anything. I'd like to see that in VW as well.

5. I would like to be able to move multible points on Nurbs curves with the '3D reshape tool'. Also I would like to be able to move the points relative to the screen instead og just the XYZ-coordinates.

6. Oops here is another one: Snap to the points and contours of a trimmed surface.

Keep up the good work:-)

Link to comment

I agree with all of the above, here are some more/ additional/extra

Vertex control.

I'd like to be able to manipulate verticies In user defined groups, or "resolution" .That is, I'd like to be able to select one, two, or ten verticies or surfaces, and push or pull them -both "by eye" and/or by exact dimensions and have them behave like "one surface" while keeping the edges intact. See TouchCad for examples of this. See also Kaare's "hyperNURBS" entry along the same lines.

Maintaining the integrity of edges is particularly important where manipulating the "interior" verticies of a surface which abutts other surfaces or solids.

While on the subject of verticies; the current situation where you can drag a vertex with the 3D reshape is ok, but to move a vertex by a defined amount you have to use the select buttons in the OIP and cycle through them one by one till you get to the one you want, and then change it numerically in the OIP. It would be much better to be able to directly select a vertex to edit (by shift-clicking on it for instance).

Tom wrote "Drawing curves using control points are more intuitive"; well I'd disagree with that entirely.

I never use Beziers for this reason, always Cubic Splines, because I want to define the points through which a curve passes, not define points which are "some" distance away from the curve I want to draw, and which, when edited, dont usually change the curve by the same amount. Likewise I'd prefer more/better interpolated surfaces, so that if I move a surface vertex 10.00mm, I move the surface 10.00mm.

Surface/ solids modeling

I just want these to work more than anything new.

Tom in previous posts has outlined some of the mathematical complexities of NURBS surfaces, and the constraints involved in applying operations to them, but I'm pretty sick of seeing "whatever-it-is has failed" error messages 70% of the time when using the "PowerPack" tools. Having to find out by trial and error which particular order things have to be selected for the Combine/Connect tool to actually work (if it will) is tedious to say the least. I'd rather go back to using clay like we did in the Olden Days than bang my head against that wall for too long.

Tom's point; that there are rules which need to be addressed (that is to say learnt) for NURBS to work successfully, is valid enough, and certainly no-one expects (for instance) vehicle manufacturers to teach people how to drive, but for someone (me) who is learning NURBS modelling using VW, it's an uphill battle.

The manuals are close to useless in this situation. (NB: VW manuals are FAR from being the worst around, but they're also a long way from being in the "Great" class either). I'm actually one of that band of misfits who reads the manual before I try something new, and on NUMEROUS occasions I've followed the instructions step by step to have the operation fail, and to subsequently find out that there was some piece of vital information missing from the instructions in the manual. If it happens that NURBS are particularly difficult to deal with and require a particular order of operations or some other constraint, then it would be a good thing for everybody if something along those lines was explained early on. Think of the TechSupport time saved at least. I'm not suggesting you teach us how to drive here, but if the indicators on the car don't work unless the lights are turned on first in a particular order, I'd like to know that before I buy it and go for a drive.

Anyhow, getting back to the story; The NNA blurb for VW11 describes the 3D features as "faster and more reliable", and I've seen them described (here, I think) as "more Robust" . The logical implication is that not only they were previously less Reliable and Robust, but they are not yet Reliable and Robust, just more so than before.

So here's my wish for the 3D tools: I want them to do what the manual says they will do. I want them to be RELIABLE and "Robust". I want the combine/connect tool to work on 2,3, -or more surfaces regardless of the order they were selected in. I want the Fillet tool to work. If a loft produces a group or a solid addition instead of a single surface, I want to make it a single surface with one click.

In short: if you add no new "features" at all but make the existing ones both tolerant and "bullet proof", then I, for one, would be happy.

Well it's getting late, and I'm tired. There's probably a lot more to be said on the subject of Usability -SEE Sketch-up and Concept 3D for some good examples,- but not tonight.

bye for now,

N.

[ 07-05-2004, 09:00 AM: Message edited by: propstuff ]

Link to comment

Yes, but instead use 1 3D object to cut into more than 1 other 3D object at the same time...leaving any object that was totally cut to be left with 2 seperate objects (new objects will then have their own object info.)

Also a couple of year ago one of the other Software Engineers sent me an email asking me to tell him what I would like to see in Log Home tools....He thought it would be a great nitch. I am working on designs for my Log Home, and have had to learn some tricks that could possibly be done with special tools. I can not remember the Engineers name...I think it was maybe Mike...and I think the last name began with "G" Gerr? I do not know if you are even the one to work on these kinds of tools...but I figured I would pass it along anyway

Link to comment
  • Vectorworks, Inc Employee

If the multiple objects you want cut all have the same properties, and you don't care about making distinct objects, you can use Section Solids. If the sectioning object is a closed surface, it works like the Subtract Solids command, except that multiple objects have a single object subtracted from them instead of vice-versa.

Link to comment

I think the subtract solids option Tim Firkowski is requesting is one which I have wished for as well.

To explain:

Problem-

I have a stack of objects and I want to subtract a hole straight through all of them while retaining the individual editability of each object.

Presently I must create a duplicate 'hole' object for each object subtracted from. If the stacked objects are added, the operation can be done with one 'hole' object, but the stacked objects lack individuality.

Solution-

Provide a radio button in the Subtract Solids dialog box that allows the highlighted object to be subtracted from all others, ie. the 'hole' object would be duplicated and grouped automatically for all subtracted objects.

Problem'-

This brings up the issue of editing the hole profile. Are all instances linked to one editable solid, or are they individually editable.

Solution'-

Editing the hole profile should elicit a dialog box that prompts one to choose between; 'all instances' and 'this instance.'

[ 07-07-2004, 04:09 PM: Message edited by: EvB ]

Link to comment

EvB,

I think you hit the nail on the head (glad you did not get your thumb [smile] )

My intial use for this tool is designing with Logs. I "stack" logs, but then need to cut openings for windows. (My Logs are not recognized as walls) Some logs are not cut through completely, but those that are cut "in half" would want to be recognized as 2 objects so as to facilitate the size and quantity of logs needed.

I have found many other places that this tool could be used, but the logs/windows started it all for me.

OH...I did find the name of the Engineer that had sent me an email about tools for Log Home design....John Kerr..that was back in 1999.

Link to comment
  • Vectorworks, Inc Employee

It might be possible to rig up a script that does a similar operation. If the subtracting object is turned into a symbol, editing that symbol would then affect all the solids. Ungroup the symbol and edit if you want to change only the active instance.

The fast way to do this currently would be to convert the hole profile to a symbol with the leave instance in place checkbox checked, cut it, and then paste it in place repeatedly as you select and subtract or section the various objects you want the hole in. Certainly it's less efficient than having it built in, but it's not too bad if you're not doing this very often.

Link to comment

1. On nurbs surfaces when using the create contours tool i want to obtain more clean nurbs curves.

2. On extrude along path an option to define the object resolution. Maybe in that way we can get more accurate conversions.

3. More exactitude on cut 2d sections or convert to lines. Maybe we can get real polylines to define a curve. All this reinfoced by a better tessellation.

4. Faster nurbs calculations.

5. And OVER ALL: MORE EXACTITUDE, VISULA AND INTERNAL.

Link to comment

Biplab,

I've been thinking some more about 3-D enhancements.

It recently occurred more forcefully to me, regarding both 2-D and 3-D applications, that a crucial issue for VW are the 2 quite different modes of Designing (whether designing with a pencil, or with the computer).

The first is the "conceptual" development phase of the design process, and the second is the design documentation phase.

Before I start, I'd like to point out that I'm coming from a "Product Design" point of view here, and worse still; product design from a commercial sculpture background. Despite the pleasing growth of curved Architectural forms, architecture is still, predominantly, a "planar" activity and hence the following discussion probably doesn't have quite the same relevance.

Many of the problems that I've been having using the PowerPack (PP) functions can probably be put down to me trying to use those PP functions for design development.

The criteria for successful use here would include;

1; Ease of use, including an intuitive environment, and tolerant and robust tools.

2; Editability, including flexible editing options and the ability to push and pull forms (both interactively and numerically).

3; Implementation, including the ability to effectively translate a design "form" into the practical Documentation-for-manufacture phase of the Design process.

The traditional media for design development has been modelling clay (in all it's forms). Automotive designers, with all the CAD-CAM options available to them still use clay for modelling (some of) their designs. For experimenting with curved (as against planar) forms nothing beats traditional, soft, water-based clay. When the form is more resolved and proceeding to detailing, moving to wax or oil based clays that are stiffer and non-drying is usually better.

The point here is that clay offers the designer a fluid, flexible medium with (basically) no technical modelling issues to resolve. (there are, of course some technical limitations to what can be modelled, but not much that some wire armatures and sticks can't deal with) That is; if you want to add or subtract or manipulate a surface, there is no " geometry can't be computed" or "operation failed, have another guess" messages. If a bit falls of your model, it's obvious that you need to stick a bit of armature through it. With an understanding of Casting, Machining and Fabrication processes and their applications, it's really relatively easy to produce any form in traditional modelling materials that can be manufactured by contemporary processes.

Consider now, the current 3D PP. The tools available have come a long way from their first iteration in 9.5 in terms of power, interactivity, ease of use, and flexibility; and I'd like to stop here to thank you Biplab, and the team, for their introduction and continued development. Congratulations, and well done!

The problem, from my point of view, is that their "power, interactivity, ease of use, and flexibility" are not nearly enough for the kind of experimentation with Form that the traditional clay process gives us.

The second half of the story; that of turning a design into a product, is where the PP functions are possibly more effective. Given that one has a design already conceived, (and can find out from somewhere the various functional limitations inherent in VWs implementation of NURBS), it ought to be possible to devise a strategy for developing the required form. Note that for this to be successful requires 2 prerequisites: a practical understanding of the ins and outs of NURB behaviour, and an understanding of VWs implementationof NURBS behaviour and relies on both the VW documentation, and due diligence on the part of users.

That is to say; where you know in advance the shape of the thing you want to make, it ought to be possible to devise a strategy to build it, as against experimenting with shapes to develop form.

I've been recently working with the demos of Sketch-up and Concepts3D due to my frustration with the PP functionality, and my desire to concentrate on the problems of design, rather than the problems of software limitations. Both these products have a stated aim of "ease of use", and while sketch-up makes no claim to be a "CAD" tool in the traditional sense, Concepts is making real inroads in to the realm of combining a fluid and powerful working environment with the precision and exactitude needed for "production grade" results. Given the competition in the CAD market, the question remains: where now for VectorWorks in terms of 3D?

In an Email conversation with Sean Flaherty, some time ago he mentioned that NNA saw the

3D conceptualisation and visualisation field as an appropriate market for expansion for VW. That suits me; I rarely do Commercial Sculpture any more; mostly Furniture Design, and teaching Design. As I only do Architectural freelance drafting for occasional jobs when other work is slow, the Architectural side is not of strong interest to me. Clearly, though, VW has a core of users in the Architectural and allied fields, and developments in the 3D functionality need to take that into consideration. At the same time, the ease of use of sketch-up must have made many CAD users of many programs wonder "why isn't mine that easy to use?".

So, I've been rambling on again tonight, and it's time to come to the point. The following list of 5 is based upon; the issues that are close to my work, my looks at various competing 3D software, and the interests of the user base of VW.

1/ It has to be at the top: reliable, robust and tolerant tools that are easy to use.

Included in this is more specific and useful reference material and more specific and useful feedback than the "Operation has failed" type.

2/ Dynamic cross-sections. These have been at the top of many peoples lists for a long, long time.

3/ Non-orthogonal dimensions. These are also long overdue in VW.

4/ Branched NURBS objects. Concepts3D offers up to 5 branches; getting VW to do 1 is a battle.

5/ A history palette. Again see Concepts3D; Structure/history File tree. (I've forgotten what it's called). Not specifically 3D so much, but a very powerful way to edit 3D objects, and much, much better than double clicking to just go one level down each time.

That's enough from me now,

cheers,

N.

Link to comment

My 2 cts...

We need better navigation tools, both in 2D and 3D. Most 3D apps can make use of the 3 buttons of a 3 button mouse to navigate through space: Left for rotating, middle for panning, right for dollying/zooming - all in combination with a modifier key. This is how most other 3Dapps handle their navigation: 3 keys or tools, nothing more.

In the current implementation of VectorWorks, you have the regular shortcuts and tools for 2D navigation (ctrl-1. ctrl-2, v, c, ctrl-4, ctrl-6, space bar, arrows, scroll bars plus the tool icons) that also work in 3D. For 3D navigation, you have 4 tools (represented by icons) that can be combined with modifier keys. There are simply way too many keys and options, which results in a cluttered interface experience.

Another thing that needs to be solved is the perspective view. Right now you can modify the view area by draggin the perspective view handles. Please replace this by

(a) a four-view system (perspective, plan, front, right and so on)

(b) a correct perspective window. Right now you have no idea of the camera settings you're using. You can change the viewing angle by modifying the perspective viewport, but have no feedback on how much it has been changed.

Anyway, the whole 2D- and 3Dnavigation system needs a rehaul. The points above are just the highlights of the current situation.

Cheers,

BaRA

Link to comment
  • 3 weeks later...
  • 2 weeks later...

I would like to have a "noise" function. I'll try to explain what I'm talking about.

If I have a curved surface that represents an amorphous form and I would like to mess it up, not step by step, but generally in one click. It would be nice to add some random placed vertices, in random heights (relative to the surface). Of course we need few sliders that control the parameters of the "noise generator". I think it should work like the stipple dialogue box works.

Such "noise" generator would be very useful. For example, we can create a surface that represents land in a 3d model. We can sketch quickly the base surface, and then add the noise that will gives it the "real life" fill. Or another one: we can quickly create a 3d object that represent the treetop, use the "noise generator" to change it into something more complex, apply transparent leaves texture, and we have a great tree.

Yovav

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...