Long shutter motionblur, only trail behind, not infront?

[Buck]

I'm trying to do a long exposure motionblur for rendered sequence, where the motionblur lasts over 100 frames backwards in time, to be added on top of a render of the exact same object that has regular 180 degree motionblur. So for instance if a person walking would leave a transparent trail behind them that lasts for several steps back in time. As they keep walking, the trail would keep following them, by 100 frames (or however many is chosen).

I can get beautiful looking long shutter motionblur, however I can't seem to get the motionblur to NOT blur in front of the person walking, it's always infront and behind him. I feel like the interval center parameter is what would do this, but it can't be adjusted beyond -1.0

I can get it to work by offsetting the long shutter render. If my motionblur is 100 frames long, I can offset it in nuke by 50 frames, and it lines up and does what I want. Is there a way to do this without offsetting the frames?

Is this possible? Or am I just missing a parameter somewhere?

Thanks!

[hermit.crab]

Could you share some screencaps illustrating both the desired effect through the offset in post and the result after rendering?

[kosso_olli]

We had a similiar issue in 3dsMax and were not able to solve it properly. We always had the motion blur "ahead of time", so to speak.

[SonyBoy]

Can't imagine why you want to try to do this in-camera in the first place. Your render times must be off the chart. Have you looked at post options, such as Timewarp in After Effects?

[Buck]

Can't imagine why you want to try to do this in-camera in the first place. Your render times must be off the chart. Have you looked at post options, such as Timewarp in After Effects?

actually, it was faster for me to render it in v-ray, than to get something that looked as good in Nuke using Oflow. It's not a person walking, I'm just using that as an example, I can't speak about or show the actual render.

2d solutions to this also run into problems when things overlap in certain ways. And to get the blur to fade off like I can get with V-ray, I have to do many frame counts of oflow timewarp and add them together. For example, one Oflow that is 10 frames, then 15, then 20, then 30, and so on all the way up to 160 (or whatever my length needs to be). That can take quite alot of time to render, where as V-ray's blur will be nice and smooth tapering off. Maybe AE has a different way of doing this, not sure.

Here's a simple example, it takes the ball 160 frames to go from left to right, and has 160 frames of motionblur.

Using oflow, Im able to get the blur. But I cannot get it to fade out towards the end.




The example below I've added a 2d ramp to simulate the falloff that I can get in Vray (or stacking multiple oflows, like I mentioned above)

[Buck]

using v-rays motionblur, you get something similar to this. Nice smooth falloff and motion, but it's going both ways when you add the long motionblur render to the non-motionblured render, ahead of and behind instead of just behind. I know none of this is actually possible in real life, but curious if there's a way in V-ray:

[fixeighted]

In Vray IPR it seems to be a bit broken and doesn't update properly or at all when changing settings, so it's really hard to iterate to see what makes a difference.
Refresh sort of refreshes it but in fact is completely different to what happens if you hit render. Thus it's frustrating.

So I got annoyed and tried Corona and got immediate feedback and exactly what I wanted the first time I tried.
The IPR reacted according to all changes and the render matched.
In this image you can see that the only thing I have changed is the shutter speed, plus the offset to
create the backwards blur. Could not have been simpler.

I went back to Vray and tried to mimic the settings and could not achieve this result, regardless of settings
So I'm now confused as to whether it's buggy or just that in order to do it I need specific settings which,
after an hour, I simply cannot discover or in fact relate to Corona's settings, as similarly named ones do not have the same impact.

[kosso_olli]

Props for using Vlado as test object.
We had to render the classic long-exposure sparkler photograph. The setup for the sparklers was very complex and already there, so we used that.
The issue was exactly like Daniel described: The bright particles causing the streaks were always in the middle of the streaks, regardless of settings.

[fixeighted]

LOL It didn't occur that it looked like him...it's not though, just a random walker I had lying around
Just tried a very simple sparkler thing like that and of course it works fine...but only using Corona.
So I guess Vray can do it, or will be able to at some point - Corona can't be doing anything magically different can it? Or some sort of fake?

[kosso_olli]

Here is a small test image and the corresponding file. The sphere is on a path constraint. I would like to get the streak with a nice falloff: Dark at the beginning of the path, and brighest right at the end. That is the result you would get while long exposing a shot like this. Right now, the streak is evenly bright across the whole duration. At first it seemed like shutter efficiency would help, but it works so weird it makes things worse.

Warning, Max-File attached. It is the same problem there, but different DCC.



sparkler_motionblur_TEST.zip

[hermit.crab]

Buck If I understood correctly, you would like for the motion blur to be lighter at the front (where the object is), rather than all the way through? A negative Bias (Overrides>Motion Blur) parameter should be able to help in this case.

[kosso_olli]

I just tried negative bias in my scene, and it does not have any effect whatsoever.

[fixeighted]

How did you get it to fade out Buck ?
EDIT: Don't worry, I figured it out.

[fixeighted]

Bias does work but only if using a perspective view it seems.
I only found out by accident, as the docs when talking about camera motion blur then go on to give the example from a perspective view.
I hate trial and error on such things that should be very apparent.
I give the Corona example again. Simple and effective with immediate results and IPR which
updates correctly.
I will hope for an update for Vray and some better documentation for everyone to benefit from.
For now, I give up, as it is just too weird.

[kosso_olli]

I will hope for an update for Vray and some better documentation for everyone to benefit from.
For now, I give up, as it is just too weird.

Same here. Can not get it to work, even though the view is in perspective mode.
Also, one thing that always give me headache: Why are there two locations for parameters regaring the motion blur: One in the camera itself, one in the camera render settings. Some values even overwrite each other. Confusion times two.