Cineversity

Hi all,

Just wondered… Is there a global timescale setting for the jiggle deformer which can increase/decrease
the speed of the oscillations as they die down. The results I am getting do not really match the scale and
speed of the animation movement.

I’ve tried: project_settings->dynamics->timescale with no perceptible effect.

I can change the cache playback speed to 200% but then I run into sync problems with the character joints
and other dynamic objects in the scene.

Changing the number of springs and iterations within the jiggle deformer is complex and unpredictable
and does not really yield the result I need - i.e. a 200% speed increase in the oscillations without
affecting the other scene parameters (i.e. scale, framerate etc).

Any ideas appreciated!

Cheers.

K

Hi keith_2,

If I get your target correctly, you like the initial effect, how far it oscillates, but the “dampening” of such should be more dynamically and progressive. Then:

The parameter “Drag” is supposed to do that. If you have not the results you are looking for with this option, increase the “Motion Scale” accordingly. This should result in an expressive initial response—without a long echo of the previous event.

If that is not working, perhaps you need to employ an additional Friction Object, inside to the Jiggle—Forces Tab. This can lead to a behavior like a sports car suspension. It takes the punch, but it is very quickly un-impressed about the last bump. This comes closer to the “Iteration” parameter combined with the Drag.

If the above is not solving it, then:

I believe the Iteration parameter alone is the key here. Think of it as the time scale, perhaps it is the one you originally wanted. If you only change the iteration is like speeding up the loss of energy. It cycles to my understanding through the complete entropy—but faster with higher numbers. As you wrote already about it, I’m uncertain what really would work for you. It is all dependent on each other. But the Iteration parameter doesn’t eat the energy and limits the oscillations, it just moves quicker through the process.

All the best

Hi Dr Sassi,

Thanks for the quick feedback with my problem.

Sadly, I have tried the drag settings, and motion scale (although this is obviously only available with the ‘on stopped’ option active).

I would say that the appearance and amplitude of the jiggle in the scene is perfect for the application, but it is the timescale that is wrong!
The oscillations simply look too slow (half speed really). Altering the springs, iterations, stiffness etc simply changes the character of
the jiggle motion but not really its oscillation time period!

I can’t think of any other options other than messing around with the frame rates, but this tends to knock out all the synchronisation
with other elements in the scene (hair dynamics etc).

It would be really great if there was some timescale control available for jiggle in the future, as I seem to run into this problem quite
frequently at the moment.

Any other thoughts/ideas appreciated!

Cheers.

K

Thanks for the feedback,

The Iteration parameter does exactly this. If you increase the value here it is like increasing the frame rates. I made some tests, based on you questions, with long movements, so I was able to count those. With the Iteration parameter up, I get the same moment but faster. Yes, I was on the on stop setting.
To increase the amount of “swings” the springs needs to be less strong to allow for the many swings, the Drag needs to be low (so, no friction as well), to not stop that swinging. But the iteration must be high to allow for a larger amount of oscialltions, but not long. This would be the logic I get so far between how I understand your target and the options of the Jiggle.

All the best

Hi again,

Certainly, I will try your suggestions and further experiments with the iterations parameter, but I do feel
that altering this parameter changes the character of the jiggle in question. Perhaps it is the complexity
of the model and scene in question?

I started with 2 springs and 10 iterations (with v low drag) and the results were visually correct. Having
changed the iterations to 40+ the speed of the oscillations, as you suggest, are increased, but the
character of the jiggle is also changed. I find it difficult to get back to the original appearance by changing
springs, drag, stiffness etc which is problematic if a client wants exactly the same appearance but at
a different speed!

A simple ‘speed dial’ would be great if it was possible to incorporate somehow.

I’ll keep experimenting with your suggestions though!

Thanks.

K

Yes, it is complex or it appears to be, and especially the manual gives you here not a lot of that interaction: case in point here the number of springs. What springs are used if the number is small? The only reasonable answer I can up with is, the springs close to the moving joint, and the more springs involved the more the effect shows. Well, I might be wrong here, and if I’m, then certainly based on the need to interpret the manual than getting a more detailed instruction here.

So, lets lower the complexity.

https://www.amazon.com/clouddrive/share/XNah3dfO_sAyxeflLjeFfHEmDHQIO3hk9eGj4Bc-Y28?ref_=cd_share_link_copy
Have a look how each step influences the areas of the cylinder itself, how fast it gets back (and oscillates or stopped) and how far it moves).

OK, testing the thesis, with many joints along a Cylinder. The more springs are involved the more detailed the effect is, from 1 to 4 springs for example. So, if you found a good number, leave it there. (To make things more visible, crank up the Motion Scale for the time being.)

Now change another value, the Iteration. Higher numbers will speed up the process, and if you miss details then they are hidden based on the fps (e.g., Nyquist Frequency), but it moves along the initial impulse and its consequences—just faster.

To get the change, and I know you wrote about the springs that you don’t like to go there. But lets put it into a nut-shell for now. Stiffness and Structure and the two springs. Stiffness stands for the ease how it can be influenced and how fast it goes back to the initial state. The Structure is more a parameter how far the change goes. (I stick all the time to the cylinder deformed first by joints)

I have several backgrounds, many degrees, e.g., in fine art as well in science, so I understand your idea how it should look like (I hope at least I do), but I also know that ignoring parts of the equation leads to the feeling that the subject must be complex (leaving out the springs or changing them with other stuff at the same time). I strongly encourage to build the cylinder with many joints and change always ONLY ONE parameter. If you change two or more, the results are useless, except you got what you want accidentally.

I have attached the little set up, I’m certain you have no problems to set this up on your own, but just for communications sake.

On the other hand, and I have said it often here in the past decade, any automatic or kind of dynamic procedural, has the complexity you mentioned. Based on that, it is not always the best way to get the expression you as an artists are after. Yes, it is more work to set it up otherwise (e.g., PoseMorph) but that separated the average from great artists, to go that extra mile, if comfortable options show limitations. If you are in full control, the results are always closer to the vision you had about your scene.

My best wises

Hi again!

Thanks for the detailed examples you have presented. They do indeed show the effect that increasing iterations have on
the cylinder in question. Unfortunately on the 1358945 polygon model (correction: overall scene size!) I am working on I find the
increase in one parameter (iterations) does not seem to have the effect I am seeing in your cylinder example.
(The jiggle effector is restricted to a smaller number of points on the surface).

The best I have achieved is something akin to a slight apparent ‘speed up’ effect in terms of the oscillation frequency,
of perhaps no more than 15-20%.

I really have tried most options, including keeping the springs at 2 but just increasing the iterations. This does not
have a linear effect in increasing the oscillation frequency. A doubling of the iterations just seems to make a small,
almost negligible difference. Putting the iterations up to 100 merely increases the harmonic vibrations within the model
without increasing the overall apparent ‘timebase speed’.

The only way that I can see the visual result I need is to play back the preview renders at 50fps as opposed to 25fps, but
then all the other animation elements would have to be scaled to fit - hence the problem!

Since it’s getting rather late here, I’ll probably look at this again tomorrow with a fresh mind!

Many thanks!

K

I never had a 1,358,945 polygon large object with a Jiggle Deformer (even partially). It sounds very large, but you might have your reasons.

IF the area is limited, the Structure parameter has less influence, on part of the equation that has already less power then, sounds easier (kind of).

Perhaps (just a guess) you set up the Scene fps to 50 and then set the frame step to—two—in the render-settings. Even if I think we play here with numbers only…but if it delivers what you need, why not.

Alternative to this, you might need to find a more manual way, to get what you like to have, or use your workaround.

I’m sure with your shown intention to get what you want, you will find the right way.

All the best

Hi again,

I’m afraid the number of polygons are rather fixed due to the nature of the project in question, double checking the info tab it does say
I have over a million polygons in my scene overall (doh! tiredness does this sometimes) - but even then it is a very substantial model being animated.

Despite this jiggle was working very well on it with 2 springs and 10 iterations - except for the timebase issue as discussed.
I would like to have continued using jiggle but it will have to be abandoned just for the moment…

I think I will probably try a workaround using dynamics. By attaching nulls to the character rig and then binding them to the
character mesh with with carefully painted weight maps I can apply hard-body dynamics to the nulls and use the follow position/rotation
settings in the ‘force’ tab to create a ‘wobbly’ lag to the selected areas of the mesh.

The timebase can be controlled using the ‘timescale’ settings in the project_settings->dynamics pane, so I can probably get the
correct timing via this method. It won’t look exactly like the jiggle version, but the frequency of the wobble is more important
at this moment in time. I’ll give it a go tomorrow…

Once again, many thanks for your advice. Hope it’s useful for others encountering these issues also.

Cheers.

K

Good luck with it, K.

I hope you can imagine that I’m certainly try to follow, but I have never seen a scene file. Which leaves me to say,

my best wishes with it

:o)