Grid deform by curve and interpolate between outer curves

[sant0s]

Hello everyone

I attached an image of what I am trying to receive.
In ICE with particles its quite simple by using the curves tangent, distance between etc.

But how can I do that with a normal grid?
At the moment I am doing that with a Lattice - that works, but it would be way more easy if I could use outer curves to interpolate since it is more accurate.
The Loft-Op does not work of course, since I can not animate the grid along a curve anymore.

Thx alot and god xmas to you all,
sant0s

Ps: the example is made with the Loft-Op to show what I mean.
And btw, are the videos on Paul Smith Vimeo site somewhere else? I cannot find all the great stuff anymore.

deform_by_curve_and_interpolate_between_outer_curves.jpg

[rray]

One approach that could work is if you extract 2 subcurves and loft between them, then change the extract parameters so the grid travels along the curve.

that might change the topology though if the extraction changes the number of points. Maybe it's better not to use extract but 2 shorter curves that are curve deformed to travel along the 2 main curves.

And btw, are the videos on Paul Smith Vimeo site somewhere else? I cannot find all the great stuff anymore.

Can't find them either, I really hope they're not gone for good They're one of the few valuable things left online concerning XSI

[sant0s]

Thx rray,

the probme is, that the grid must have an UV since I add displacement and lots of different shader stuff like gradient, curvator, etc.
When the topology changes, it probably gives me alot of problems.

I attached a video, on what I am workin, totaly WIP so looks like crap and has lots of errors, overlappings, etc - but gives an idea, why it needs to be a grid with UV.
I used Lattice there.

[rray]

For something like this the other method with 2 shorter (straight) curves that are curve deformed along the main curves might work. That shouldn't affect the uvs.

[sant0s]

For something like this the other method with 2 shorter (straight) curves that are curve deformed along the main curves might work. That shouldn't affect the uvs.

I think, I dont really understand that technique.
You mean something like attached?

How can I let the curves run between two curves?

curves.jpg

[rray]

The 2 curves would run along the main curves, after the deformation - wait I'll set up a quick scene
__

here's a scene, the null controls the path position - - turned out though might get tricky to control the position of the opposite side

[sant0s]

wow, thats crazy, never have thought about something like that!
will build that into my scene.

thx rray,
sant0s

[rray]

you're welcome- a problem with defom by curve is that the tranlation always measures curve length from the first point, so you can't really align the 2 sides well along the path by tweaking the curves.
good xmas to you too btw!

[sant0s]

I, with my absolut little math knowledge, though maybe it could work by doing it that way:

I have the curve to deform the grid with the normal "Deform by Curve" (I also found a ICE Compound that is doing the same, so even more procedural).
Than I have the two outer curves.
Now I take the distance between the outer curves to the deformer curve.
The grids PointPosition should than, based on the distance to the outer curves, move interpolated.

Ill just make an example what I mean, because I think, I was close.

And thx, wish you also good xmas!

[sant0s]

here, that is what I mean...
I cannot explain actually what I mean, but thats how I understand that.
One of Puals Tutorials was doing something like that, but with particles.
They where linear interpolated I think, but some time ago I did that.

edit: the numbers are just fantasy of course.

math_lol.jpg

[rray]

I think that's possible with meshes too, it's after all just PointPositions. But not sure what's the missing element atm in your setup. Too bad those videos are gone (I hope they'll be back).

You could look up closest positions on the curves, but with closest position there's alsways the problem that sometimes several points can have the same closest position which could lead to strange deformation artefacts.

[sant0s]

I just found a scene thats based on one of his tutorials.
Its different since its tangent related...
But maybe there is something that helps?
It works with the nearby points compound.
I was also just looking for all kinda words in your softimage ressources database - seems like a very uniqe problem...

tan_test.zip

[FXDude]

Hi Santos,

As far as I understand, the scene RRay posted seems to address what you wanted no?

Could you be more specific as to what you'd like your end result to be?

From what I can tell, just a third center curve (offset in Z) in the loft for that setup would do the trick, no?

[FXDude]

And btw, are the videos on Paul Smith Vimeo site somewhere else? I cannot find all the great stuff anymore.

Hey! Just noticed that...

I really hope they're not gone for good They're one of the few valuable things left online concerning XSI

and that statement is an increadible understatement. (Pooby!?)


I personally quite often end-up finding solutions to specific things in one of those videos,
(not neccessarily frequently, but whenever making new setups, everytime being like godsent)
and I'm pretty sure that also goes for anyone still using ICE ...


That apart that they are excellent references to what is possible with ICE (for ICE users),
and is a go-to, one stop shop for what ICE is all about (for ICE, and non-ICE users)
( for future references, reminders, and possible eventual (non-Houdini) replacements.)

All with very down to earth, intelligible, to the point, and sometimes humorous narration.


Does anyone happen to have downloaded them?

Or Pooby, could you provide a zip?

I hope they just have been made private, and could therefore easily be remade public, or be made accessible with passwords..


This is equally as big (but bad) news, as Mootzoid tools being made widely accessible was big (but good) news,
and very-much hope it's just a temporary situation.

pretty baffled here .....

[sant0s]

Hi Santos,

As far as I understand, the scene RRay posted seems to address what you wanted no?

Could you be more specific as to what you'd like your end result to be?

From what I can tell, just a third center curve (offset in Z) in the loft for that setup would do the trick, no?

Hey FXDude,
thx for joining the discussion!

Yea - actually its perfect, very fast and for what I wanna do its great.
But I just would like to know, if there is a way to archive something like I was planning for learning purpose.

It also would be interesting, since you could for example animate a NULL/geometry on the middle curve and plot two outer curves.
With a little bit of randomize or based on slope etc, it would be cool to do some fake rivers etc. just procedural without doing 1 or 2 more curves by hand.

As an example: If I have a massive map with alot of river, only creating the middle curve (without the shore line) already takes forever.
So I could just animate the NULL, generate two more curves with fake realistic shore and add some cool shader to it.

Ill create a little example scene what I would like to archive.

And about pooby - yea, would be great to have the archive somwhere saved.
I used to download most videos since I have sat internet and it always stops preloading. Bbut long time ago jDownloader did not work with Vimeo anymore and the videos are somewhere on a backup. If I find them and with his permission Ill send them to the admin here - but hopefully, he still has the videos.

[sant0s]

here is the testscene with animated river.
that, what I am doing now with the lattice (its not visible in camera view), pushing and pulling the geometry, is what I would like to do instead with two outer curves.
animating a grid along the curve, generating two outer curves (by plotting or by hand) and the grid will than align smooth to the outer curves but still uses the middle curve to drive and is the center of the river/glacier.

river_testscene.zip