PDA

View Full Version : Endomorph Point Coordinate Editing



PabloMack
02-04-2010, 01:12 PM
This suggestion is for LW[8] and I don't know if the behavior I am about to describe has been changed for LW9. I have used endomorphs a lot and I like to calculate coordinates explicitly to get geometry right. I noticed that, when you enter coordinates directly, the value you enter when you have an endomorph selected, is not taken as the absolute coordinate of where the point ends up in space with the endomorph active. Instead, it seems to become the displacement caused by the endomorph movement. A much better behavior is for LW to compute what the displacement would be to arrive at the absolute coordinate you enter. I think artists don't compute coordinates very often and this is why there hasn't been much complaint about this behavior. Artists like to do free-hand changes but I need absolute coordinate editing because I do more engineering type work than artistic work.

SplineGod
02-04-2010, 02:34 PM
I believe that there are two types of morphs in LW: absolute and relative. The only thing Ive seen that supported absolute endomorphs was an old plugin from Lernie Ang.

UnCommonGrafx
02-04-2010, 03:04 PM
Smartmorph does and rather smartly!
http://www.polas.net/smorph/index.php

And I am really not sure what the initial concern is. Or, I'm not sure what the point is.

Edit: an example for the need of this would assist my clarity, should it matter.

Dodgy
02-04-2010, 09:51 PM
You can get them to work with nodes. Try this out

UnCommonGrafx
02-05-2010, 02:29 PM
You wouldn't happen to have a funky scene showing this as a morph, now would ya?

PabloMack
02-15-2010, 08:53 AM
I believe that there are two types of morphs in LW: absolute and relative. The only thing Ive seen that supported absolute endomorphs was an old plugin from Lernie Ang.

Don't be misled by my usage of the term "absolute". What I am talking about is the behavior of the "relative" endomorph mode.

PabloMack
02-15-2010, 09:14 AM
Smartmorph does and rather smartly!
http://www.polas.net/smorph/index.php

And I am really not sure what the initial concern is. Or, I'm not sure what the point is.

Edit: an example for the need of this would assist my clarity, should it matter.


Smartmorph does and rather smartly!
http://www.polas.net/smorph/index.php

And I am really not sure what the initial concern is. Or, I'm not sure what the point is.

Edit: an example for the need of this would assist my clarity, should it matter.

I just tried it on LW9.6 and the problem is even worse than I described. Try making a cube that is 1m on each side. Now create an endomorph and make it active. If you then move one of the cube's corners using the t command, LW computes the relative movement from the base position and that becomes part of the endomorph. In other words, you move the absolute position and LW computes the relative movement necessary to get it there.

Now if I, instead, edit the cube's corner coordinates (using the i command), IT CHANGES THE BASE MORPH's POSITION and it does not become part of the endomorph even though the endomorph is active. This is a serious problem! The number I type in should become the new position that I see in the view port as a result of the endomorph and should be equivalent to having moved it there with the t command. BUT NO!!!!!!!!!!!!! IT DOES NOT! It becomes the new base coordinate! My endomorp still does nothing! This needs to be fixed!

I think I know why this problem has never been fixed. The over-whelming user base for LW are artists. Artists just move stuff around by hand until it "looks good enough". But some of us use LW for engineering work and we need to get dimensions exactly right (not just looks good enough). So this problem has never been fixed.

UnCommonGrafx
02-15-2010, 09:59 AM
I can appreciate your need for precision.

I'm now curious: would an absolute endomorph do the trick for you? This is the FIRST time in all these years that some light has been shed on the naming of the two endos we have. And it's not light as much as a bounce.

So, Pablo, have you tried to do this with absolute morphs? Do they live up to their name? That is to say, do they allow you to have exact placement of points compared to relative morphs? Or are they still 'relative' to the base and not actually absolute in their creation?

UnCommonGrafx
02-15-2010, 10:03 AM
Dodgy,
Mind a hint as to how to animate moprh to morph? Still trying but it ain't grockin' here. All I get are scaling changes.
Or anyone...?

You can get them to work with nodes. Try this out

PabloMack
02-15-2010, 10:21 AM
I can appreciate your need for precision.

I'm now curious: would an absolute endomorph do the trick for you? This is the FIRST time in all these years that some light has been shed on the naming of the two endos we have. And it's not light as much as a bounce.

So, Pablo, have you tried to do this with absolute morphs? Do they live up to their name? That is to say, do they allow you to have exact placement of points compared to relative morphs? Or are they still 'relative' to the base and not actually absolute in their creation?

I learned using endomorphs from Dan Ablan's book on LW. According to him, the Absolute mode doesn't work. I would have to know what NewTek had in mind when they created the absolute mode before I could make a judgement on whether its behavior was correct or not. However, the edit coordinates should be just another way of making changes to coordinates in relative mode. I should be able to make changes to some points one way and to other points another way. If all of the points in a mode have to be changes only using one way, there is a serious problem. I stick with relative mode because Dan says that is the only mode that works.

If I were to make a suggestion for another mode, it would be Rotational Mode. Endomorphs can't do that because movement is always linear. You could approximate it by creating mutliple morphs, each doing a very small approximation of a rotation. But this is very messy and you really need rotational endomorphs to make it happen well. But you might need to define a pivot point for the endomorph which might be moved by other endomorphs that you create for the same object. This could get complicated, but no more complicated than other things that are done well in LW.

UnCommonGrafx
02-15-2010, 10:40 AM
Dan was working in the past when no one had designed anything to work with absolute morphs. As has been stated above, that information is outdated, at best, erroneous, at the worst.

Hmm, Endos don't need to rotate if the object itself can be rotated.
And I now have to see if rotation works with smartmorph... Now I'm curious. The reason for this is that smartmorph is one of the FEW tools that interpolates endos in a non-linear fashion. This has alleviated the need to step through numerous morphs.
See if there is a demo to try out. I'm curious as to what you come up with.

PabloMack
02-15-2010, 12:53 PM
Dan was working in the past when no one had designed anything to work with absolute morphs. As has been stated above, that information is outdated, at best, erroneous, at the worst.

Hmm, Endos don't need to rotate if the object itself can be rotated.
And I now have to see if rotation works with smartmorph... Now I'm curious. The reason for this is that smartmorph is one of the FEW tools that interpolates endos in a non-linear fashion. This has alleviated the need to step through numerous morphs.
See if there is a demo to try out. I'm curious as to what you come up with.

I had a case where there were some turbine blades that were slightly arced. I wanted them to expand their blades along an arc so that they became more scooped (and with more surface area along the arc) as the endomorph action was increased. Because relative endomorphs only do linear displacement, you can't do it using only one of them. And since endomorphs change the shape of the object (not just moving the existing shape), rotating layers can't do it either. Remember that endomorphs don't act on all vertices equally. Some vertices are not altered at all. Others move differently to change the shape of the object. If the movement of a point needs to travel along an arc, that would be a rotational endomorph. To keep it from becoming excessively complicated, I would probably define a single pivot point and orientation that would have to be used by all points modified by the endomorph. Then you just store an angle for every point. Those that were unaffected would have 0 degrees associated with the movement. Others would move around the axis of the pivot point by its specific amount. This way vertices could spread out along an arc or contract.

UnCommonGrafx
02-15-2010, 01:12 PM
My case was a lack of patience to do multi-morphs to get a non-linear result. I could (can) live with having to do three morphs to go in a circle; more than that and I was... unappreciative of the task.

Go back to my post, #3, and try the demo. I can't suggest it enough as I am extremely curious as to whether or not it will address your needs.
Barring that, you may be waiting a VERY long time for this to be implemented for LW. VERY as in maybe never...

PabloMack
02-15-2010, 02:49 PM
My case was a lack of patience to do multi-morphs to get a non-linear result. I could (can) live with having to do three morphs to go in a circle; more than that and I was... unappreciative of the task.

Go back to my post, #3, and try the demo. I can't suggest it enough as I am extremely curious as to whether or not it will address your needs.
Barring that, you may be waiting a VERY long time for this to be implemented for LW. VERY as in maybe never...

I really don't expect NT to add a rotational endomorph any time soon (maybe never). But the VERY BASIC point editing operation not working while endomorph is active should be fixed. As far as I know, it has always been broken as long as there have been endomorphs in LW. I know that this forum category is for new features, but there was no bug fix category and that is what this is. It is my experience that someone needs to have the ear of a key person on the development team to get things like this done. Does anyone at NT regularly read this category?

I have downloaded the SmartMorph demo and plugin. Sounds like a very interesting option. Thank you for the information.

UnCommonGrafx
02-15-2010, 03:42 PM
And yes, it will do rotational endos.

Round and round...

evenflcw
02-15-2010, 05:37 PM
But the VERY BASIC point editing operation not working while endomorph is active should be fixed. As far as I know, it has always been broken as long as there have been endomorphs in LW.

Nothing in your description sounds broken. You're editing base coordinates in the Point Info panel and base coordinates are updated. What's broken? Endomorphs are vmaps and edited elsewhere. RTM? (Although I doubt it is thorough enough to state things explicitly). See attachment - editing of absolute endomorph, with absolute values, in the PointInfo panel. That's what you want, no? Need relative endomorphs in Layout, then download a script to convert when your done... or use Node Editor as suggested. The only merits to your case is that one might want to edit relative endomorphs in absolute coordinates. But I can think of better workflows/tools for enabling that than to change the PointInfo panel.

PS. to get the vmap values to edit you obviously have to create them first. So just create the endomorph. Deselect all. Use move tool to offset x millimeters and then offset back x mm. This generates a vmap assignment to every point so you can edit them later in PointInfo. The only issue I can think of is that you'll likely need a script to crop out the unneeded assignments.

PS2. You could use Set Value command to edit relative endomorph values with exact absolute coordinates. There might be some more convenient 3rd party versions of that.