The thread for silly Alias questions...

I have a new question!

If you look at the image attached, you’ll see a number of curves I have ensured line up perfectly in the front view, grouped, centered the pivot, and then wish to offset. The screenshot shows the nightmare that happens.

  1. Reverse curve doesn’t seem to do anything in order to get the offset to offset in the same direction (when I use the tool, it does seem to change the direction of the tool arrow, but the curve itself just “flickers” and the U remains in the same position). What am I doing wrong?
    (EDIT: Managed to answer this one myself… a curve with construction history cannot be reversed… the “flickering” of the U swapping locations, and then snapping back because of continuity, should be considered a bug, imo.)

  2. Why does the offset in this view go left and right, even though the curves and the pivot are perfectly lined up? (They look quote different in the side view which is what matters, and sure, I can snap them back, but that’s not my point.)
    (EDIT: I think I managed to answer this one myself as well… like the project tool, the offset tool seems dependent on your camera view… but that cannot be right, can it? It should only be dependent on the pivot, surely? But when I offset the same curves, just by instead looking at the side view, and then went back to this view, the offset result was properly aligned… weird!!)

If you look in the offset settings there are 2 options for offset plane, geometry average and active view. Active view is usually the default. Just like project there are options where you normally want to select the direction based on your selected view axis.

Hey Alias users I have a couple questions about the rail surface tool.

Could someone explain the difference between using Implied Curvature vs. Curvature when using the rail surface?

Also what does Rebuild mean? I know that I am often prompted to turn it on or off to avoid some error, which I do blindly but I do not understand why.

Thank you!

Implied (tangency or curvature) is typically used for maintaining continuity across a centerline or mirror plane. For example if you’re building to a curve on the center line and just have it set to free or fixed, it will only try to get positional continuity to that point, which may or may not have tangency when you mirror it across the centerline. By setting implied tangency/curvature it tries to achieve tangency to that mirrored surface across the edge.

Rebuild means the tool will go in and rebuild the edge that you are specifying. You generally don’t want to rebuild if you are building to a clean edge (natural surface boundary or a curve) but do want to rebuild if you’re building to a complex surface or trimmed edge. This will reduce your # of spans and help improve the isoparm distribution while matinaining continuity.

Alternatively you can use explicit control to specify the exact surface parameters you want but this may not achieve continuity without further tweaks.

Thank you Cyberdemon, that makes a whole heck of a lot more sense now!

One new question, is there anyway once you have added isoparms to a surface to reselected specific isoparms and readjust their position on the surface? Or should I simply undo all the isoparm placements and just redo them?

A good habit is to always make a copy of your surface and put it on a backup layer before you insert any isoparms. Once you do, that’s pretty much the end of the road since you can not undo insertion. You can not reposition (or reparameterize may be a better word) the isoparm. If you think of your surface as a math equation, when you insert that isoparm it’s tied to a specific parameter on the surface so while you can move the CV’s around and the isoparms will move, their parameter (or relative position) will still be the same. At that point you either have to rebuild your surface (which may not yield the original result you started with) or recreate it from scratch.

If you want your isoparms to be nice and evenly spaced, consider upping the spans manually on the box in the control panel, or hold the “alt” key to automatically put the isoparm at the halfway point. (Ctrl+Alt will snap to an isoparm on an adjacent surface)

I forgot to thank you for this answer!

Here’s a new one I suspect has to do with the letters U and V, but I have no idea how to diagnose it properly:

(I especially like how the result of the union command is randomly an intersect and randomly a subtract…)

The thing is, I have other of these crossed bands which union fine, but most freak out like this.

Also, I have over a hundred of them… not just these five. Several hundred, in fact.

Swift help is greatly appreciated!

Here’s a wire:

http://www.eobet.com/temp/shellsgrr.wire

Found the answer myself: I have to blindly, through trial and error, reverse every U and V direction of one or both of the shells until the union operation stops behaving like intersect or subtract.

I still have no idea what’s wrong, or when things are right, I’m just switching back and forth, modifying every shell until it works.

Gotta love Alias!

Won’t lie to you - the shell boolean tools suck balls. It’s weird, I download your file and for X 1 3 and 5 Intersect actually joined them, and for 2 and 4 Subtract worked but only if I picked them in the correct order.

The only thing I can think of is something happened with the way those were created that was different or reversed from the others, or that since they were grouped, the group is causing some kind of weird heirarchy.

If you’re still doing it you can try using the other 2 boolean commands, but yeah it’s funky and terrible and I’m not going to pretend that Alias ever plays nice with solids. I have issues even when just using Alias primatives (like punching a hole in a block)

Thanks. I’m not going to jinx anything, because I’m not finished, but so far I’ve noticed:

When you use the Reverse UV command, the little crosshair can be either yellow or cyan. However, unlike the surface orientation command, this is not an indication of which direction the UVs are (in other words, UI inconsistency, and I’ve upgraded to Alias 2012 now).

Call me superstitious, but most of the original Alias development team is gone I believe and now the desk is left scrambling to figure out what all the source code really means and how to handle it. Nevertheless it’s still a great tool… for now. But they shouldn’t wait to much longer or it’ll fall behind the curve.

What curve? Please, please, please tell me that there actually exists a viable alternative? (And please don’t say Rhino, because that’s just nowhere near anything like Alias.)

I can’t believe I’m saying this, but I know of no Class A modeling software that is as flexible with degrees on curves and hulls, that has as many snapping and transformation options and modes, that allows you to switch construction options on the fly and which actually retains history, almost like a parametric modeler…

I don’t think I can go back to a software that doesn’t have query edit, extend, stretch, proportional modification, hide unselected, non proportional view scale, skins (with their implicit continuity guarantees) etc.

But then again, you discover things almost every day that makes you want to tear your hair out, like for example, just now, I noticed the fact that you cannot set a keyboard shortcut for “pick curve” even in Alias 2012. You have to use the bloody marking menu for it!

And of course, at the same time, I oh, so could live without “pick nothing”, the fucking triple key shortcuts, “zero transforms”, the lack of undo, and well, everything that I’ve been bitching about for eight pages in this very thread. :slight_smile:

(EDIT: I realize not liking “zero transforms” goes against me liking the history, because it enables the absolute tranformations, but the thing I dislike is that there is no proper interface to view and edit the local coordinate systems that this creates… it’s all spread out under the pivot, orientation and reverse UV tools and probably more that I haven’t found, because very often when I ungroup something, scale factors reset and duplications behave oddly.)

Hi
I’m new to alias and I wonder, what’s the difference between a cv curve and a blend curve?

I’m not an authority, but I suspect they’re the same thing, only degrees are handled implicitly with blend curves and there’s some faster UI options for aligning them (a normal curve you have to separately snap, use the align tool and make sure the degrees are correct).


Ok, so I’ve been (reluctantly) getting back into Alias (2012 this time), and maybe I forgot something very basic, but shouldn’t this work? Simple rail with implied curvature on a symmetry layer… yields tangency continuity… what?

Oh, and unless I’m mistaken, in 20112, they still haven’t fixed instances. I just did a duplication on a surface with history and it lost its history once the action was complete…

Look at your inputs.

You’re telling it you want implied curvature along your rail (which is what you want based on the picture) AND implied curvature along your generation curve.

This means it’s screwing everything up trying to generate curvature in those 2 directions.

Set your gen curve to free and it should let it flow properly. If you are actually trying to achieve implied G2 on your gen curve, then that’s a whole separate discussion and you shouldn’t be trying to use a monorail 1.

Right, so now I have tangency along the gen curve instead. I’m trying to get away with modeling 1/4 of the object in question, so let’s go ahead and have that separate discussion, if you don’t mind. :slight_smile:

I’m guessing I could build support surfs and do a square off them, but I’d like to hear your ideas on the subject.

Modelling 1/4 is fine - but if it’s giving you that result it more than likely means that where your gen curve is not G2 across the curve you are trying to create G2 across.

If you can post a screen shot of your 2 curves with the orthos and the CV’s on, you can probably see one CV isn’t lining up across your second symmetry plane.

Curves are made sure that the 2nd CV is aligned to the 1st. I’ve repeatedly snapped them and they seem fine.

Still, no go.

So, I remember that if the 2nd CV is normal to the mirroring plane, I only need tangency to get curvature across the surfaces.

Still no go.

As a last ditch attempt, I duplicate the curves off the failed surface, set up support surfaces and create a new square.

Again, no go.

Anyway, this was an exercise and not really needed for my final model. I’m going back to modelling 1/2 of it instead. Still weird, though.


EDIT: Going really overkill and making sure the 2nd and 3rd CVs were aligned in both corner curves, then duplicating these to form the other corner and building a square with implied tangency(!) from those did produce a proper curvature aligned mirrored surface… but man, why couldn’t it just work from the start? :stuck_out_tongue:

Right but remember when you are sweeping a curve, the rail it moves along wants to rotate that curve in such a way that may cause the CV’s in the opposite direction to no longer be where you think they will be. When you force them to, it can break the continuity along the opposite plane because of that transformation.

I model stuff in quarters all the time, you just need to make sure the curves you start with want to allow that.