Motion blur without decay?

Hi guys,

I’m trying to automate a trail for a simple object (in this case a yellow star). I would like to get the following (this is a faked mockup) where a yellow star follows the black path:

I think the best option would be using “motion blur” layer, but I can’t control the decay or transparency. I used the constant subsampling type and I’m so close to get what I need but as I can’t control the decay.

I found a workaround but make the animation rendering so slow: if I use a “duplicate” layer on top of the motion blur and set the copies (“To”) up to 50 then I get the opaque effect I need but, as said, it’s not very smart…

Any idea?
Thanks!

1 Like

Are you talking about something similar to this one? Was bored, so doodled a girl in synfig for you guys - #85 by Kai

No exactly, this is what I want to get:
example

I would like to control the number of “subsamples” and their separation. The motion blur layer is perfect for this task but as I can’t control the transparency I need to “hack” it by duplicating the layer as many times as needed to get an opaque result (I automate it with the “duplication layer” but I said it makes rendering so slow).

example.sifz (2,4 KB)

1 Like

That looks like this. Was bored, so doodled a girl in synfig for you guys - #55 by Kai
I’d created this using duplicate layer.

Yes, I think that one is what I’m looking for. How you made Duplicate layer follow the path?

I linked the group layer to the spline, opened the transformation parameter of the group layer and converted the amount parameter to add and then linked the index of the duplicate layer to the addition sub-parameter of the converted amount parameter. That’s it.
image
nnmdragontail.sifz (6.3 KB)

1 Like

Hi @Kai,

I’ve been playing with your workaround, it works but it doesn’t behave like a real trail, for instance, in my example the star goes forth and back and the “subsampled stars” follow what the parent star did. In your solution, it doesn’t. It’s also pretty hard to make it work as you want so, in the end it makes easier to go back to my solution even if it increases the computer calculations to the moon, hehehe.

BTW, I’m going to suggest as a feature request to have the option to set the subsamples opacity when using “constant” subsampling type, I think it’s something that makes sense and I hope not hard to code.

Anyway, any more ideas?

Thanks!

Maybe this one is what you are looking for?

onion

1 Like

Thanks @bjyitu, it looks like it’s the same approach as @Kai, it has the problem that you can’t make your object bounce from the last point of the path to the initial one. Look how my “example.sif” star moves from left to right and right to left, the subsamples always are behind the parent star.

FYI, I ended up with a feature request and @rodolforg was so fast developing it: feat: layer Motion Blur now can create a non blurred trail by rodolforg · Pull Request #3173 · synfig/synfig · GitHub
I hope it gets approved soon. :clap:

Thanks all!!

looks like this one?

onion_cycle.webp

No exactly: in your example the circle “waits” in the right side to the subsamples to “arrive”, then it moves back to the left side.

In my example, the parent star (or circle in yours) bounces in the right side and heads left but the subsamples still go right and start going left then it’s time to. Do you get the difference?

got it. that’s not the same. :grinning:

1 Like

not smart to show the effect,that’s need a lot path to keep loop :grinning:

-

onion_cycle2.webp
onion_cycle2.sifz (5.3 KB)

Yes, I also made this option when playing with @Kai’s suggestion, but as said it’s not smart, hard to draw and animate… the best option is going to be the one with the new feature, it will allow to “echo” anything you want effortless and it will always follow any path or animation the parent does :wink:

1 Like

Duplicate layer isn’t that heavy on computation,

This example you provided is heavy for computation because it is using Duplicate layer which has Motion Blur layer below it. Motion Blur is very slow, I avoid using it.

@bjyitu has already kinda found out the solution, here’s a modified version which may solve, your (@pablogil ) query. bjyitu your example for onion skinning didn’t use Time Offset parameter, using TimeOffset parameter you can create true onion skining and not worry about linking individual parameters to duplicate index value, just delay the animations using Time Offset parameter by connecting it to the Duplicate Layer’s index value.
Some times things don’t work good with TimeOffset, Speed and Duplicate Layer (more)

onion.sifz (2.7 KB)
And your original
example.sifz (2.7 KB)

I think type of interpolation is important here, also the time delay between “subsamples” also makes a significant impact, I believe your effect and shown by bjyitu are same, just different implementation of same thing. How you animate the shapes is also important.

1 Like

:open_mouth: :open_mouth: :open_mouth: i see.That’s how it works.