How to deform a transform node using lattice?

[roham]

Hi guys,

As you know we can deform the transfom nodes using another transform or just one joint.

I was wondering if there is a way to deform the transform nodes using lattice or multiple joints?

I'd appreciate any answer.

[joojaa]

Yes, you can do this. But essentially this is 2 different questions, other which is easy and the other is well a bit convoluted.

Joints well meashes dont bind to many joints but rether to a joit cluset that binds to many joints. But essentially theres a similliar tool for transforms that makes weighted avarage possible and thats the parent constraint.

Laticce a bit of a more problematic, you can bind the latice tio slmall mesh thats bound to the transofrm with a rivet or something. Essentially rpoblem is making a vector a matrix but the rivet script makes a possible fittingthat may or may not fit the bill.

[roham]

Thanks Jooja, I really appreciate your answer.

For the first solution, honestly I didn't understand. But for the second one, using rivet is a good way but it doesn't really deform the space of the joints, it just matches the position of the joint to the right place.

Maybe you could explain it more in this example:    How to put the spine joints in a lattice?

[joojaa]

For the first solution, honestly I didn't understand.

hit f1 and read on parent contraints

using rivet is a good way but it doesn't really deform the space of the joints

But then again thats not what you asked for is it? You just asked if i a transform can be driven by a joint or a lattice, and they can. Not whether you could actually use it for anything of interest, or any other assumptions.*

This s is a bit different kind of question, tough. Now id have to understand what you think the world is composed of (as opposed to what it really IS, i can relate to that). I mean is the space of a transofrm (and joints because they are transforms) its parents space of its own space? If it is its own space then howcome you could dirve numbers that are measured in the parents space (see transforms live in the space of their predecessor not their own).

Then theres a another can of worms with interactivity, once you bind the thing to the thing to a laticce its stuck there no moving directly anymore, thats how maya works. Sure a constraint could be deviced but not really in the scope of any menainfull answer.

it just matches the position of the joint to the right place.

well thats what a laticce does, there is no rotation in a latice. Points are vectors they dont know of such things as rotation, however they do know of a differential rotation tough.

Transforms on the other hand are 4 vectors, so you need atleast a 3 point mesh cluster to handle the indirect bind. Now the big problem is that the laticce is not rotating anything just translating the space, so the size of the cluster affects how things behave.

Maybe you could explain it more in this example:    Normal people would use a How to put the spine joints in a lattice?

normal people would use a spline joint for spines, then you'd drive the spline with the lattice, much more productive way. Alternatively id use a pathanim node if i really wanted the torrision control.

But yeah i can make a example of this once i find the time to sit down on my desktop, and have free time. Just that its not really all that useful compared to the other methods that do exactly the same thing.

*its a bit akin to asking somebody to build a vehicle and then complain when its not a pickup truck, when they shipped you a power boat.

[roham]

You just asked if i a transform can be driven by a joint or a lattice

No, this was not my quesiton, this was: Is there a way to deform the transform nodes using lattice or multiple joints? by that I meant if there is a way to deform the space of the tranform node so that if there is anything under this transform node would be affected by those deformation.
As an example, we can scale a parent and the child space will be deformed so that if we rotate the child, the rotation happen in a deformed space cause by parent. This deformation is what I mean not just position of the transform node.

normal people would use a spline joint for spines, then you'd drive the spline with the lattice, much more productive way.

Thanks for calling me a not normal person.

[joojaa]

Yes you can do that (but pleasre note my yes is not the same thing as yours), but if you must use existing maya nodes it gets a bit tedious on the range of 8 nodes and 24 connectionsper node*. And im still not sure it qualifies as working according to the way you THINK things ought to work.**

Is there a node to do this, no. Does it work the same way as you think it should, probably not, hard to say. Is it usable, perhaps. Can you connect it just as a parent, hell NO, but that does not mean you cant get the same result with the same gui.***

• looking into how i would package it so it makes any sense to anybody whatsoever, and stramlining it a bit as i have the most brain dead 15 minute approach going on now.

**most development processes fail (about 70-90% of all worldwide) because people imply a lot with words. A classic example form my distant past, a client wanted dxf export form a esoteric proprietary app. how much more specific and exact can a requirement get? not much right, i mean its a freaking defacto standard after all.

 Well so i got hold of the dxf standard got a validator and dug up all the applications on my machine that read dxf files. Did the export in 2 days validated and tested against 25 differnet apps, and validator. So now i had pretty much exhaustively tested the app against a bunch of industry standard apps.

Client wasn't happy, told me that it wasnt a proper dxf file. It wasn't? I was totally surprised. 2 days digging around and turns out my client defined dxf as the kind of file microsoft visio reads in (yeah there was 2 pr departments between me and my client). He had just picked the file format that he tought would have been the easiest to implement. Turns out that the particular version of visio needed a piece of export config code that was proprietary to some versions of auto cad. Which subsequently visio ignored fully but they needed to be there because otherwise the parser would just invalidate the file. Well the client said that the dxf importer has to be fixed, even tough i explained that the alternative solution is available now, reverse engineering 2 days.

Thing is had the client actually bothered to tell me he wanted to import the file into visio instead of specifying DXF i would have shipped the solution for him for free on the hour of his initial contact.

This is the reason why most developers think clients are not worth listening to, I sort of disagree but you really have to interview your clients a lot before you start working. They actually teach engineers not to really listen to clients, because few hunderds of years of experience tells them this is true. The companies that ignore clients in this way are by the way quite successful, Apple, and Google to name a few.

But your not my client.

*** DAG nodes (transforms) are a subset of things called linear math, and can not transport non linearity. But that does not stop the dg form doing it for them.

[roham]

You're amazing, thanks for your time and answer.