Strange Motion Blur Sampling

[profbetis]

I know typically people come across a problem and it's fixed by increasing that object's motion blur samples, however this is not that problem.

Firstly, this is an image from a series I'm doing about car wheels in motion. The result is spectacular and exactly what I needed (granted I had to use 128 motion blur samples).


But as I started messing with the kernel's motion blur duration setting, I noticed something strange.



It appears as though octane is taking it's samples during the entire frame, and then just using less of them based on the duration. I think this is a shoddy way of dealing with this. I think that instead (if possible) the motion blur samples should be distributed only amongst the time needed.

[Karba]

I know typically people come across a problem and it's fixed by increasing that object's motion blur samples, however this is not that problem.

Firstly, this is an image from a series I'm doing about car wheels in motion. The result is spectacular and exactly what I needed (granted I had to use 128 motion blur samples).


But as I started messing with the kernel's motion blur duration setting, I noticed something strange.



It appears as though octane is taking it's samples during the entire frame, and then just using less of them based on the duration. I think this is a shoddy way of dealing with this. I think that instead (if possible) the motion blur samples should be distributed only amongst the time needed.

I will check this case

[profbetis]

Thank you Karba! I know you are busy so I appreciate this. I know a majority of users may not benefit from this, but those of us pushing the renderer to its limits are!

[bcravin]

Here is a quick test that I ran to show some of the issues that I'm seeing.

Attached are a few stills and a link to a short animation. I'm using Octane 2.06 and 3ds Max plugin.

The scene is simple, it is a small rectangle that is rotated at a high velocity. and then linked to a box that is moving forward is space with high velocity. I have a more complicated scene for a project that I'm working on that actually shows much worse artifacts than this. But this will give you a good idea.

As for motion blur settings: Duration .5 (180 degree film standard).
Motion blur samples: 4 (I saw no difference when going to higher sample rate)
Movable proxy: Checked
vertex motion blur: UNchecked
object motion blur: Checked

https://www.youtube.com/watch?v=LnSuleJ7bMo

Can someone at Otoy please take a looks at this? or at least respond? I've posted about this in a couple other locations with no response.

Thanks!

[Karba]

Here is a quick test that I ran to show some of the issues that I'm seeing.

Attached are a few stills and a link to a short animation. I'm using Octane 2.06 and 3ds Max plugin.

The scene is simple, it is a small rectangle that is rotated at a high velocity. and then linked to a box that is moving forward is space with high velocity. I have a more complicated scene for a project that I'm working on that actually shows much worse artifacts than this. But this will give you a good idea.

As for motion blur settings: Duration .5 (180 degree film standard).
Motion blur samples: 4 (I saw no difference when going to higher sample rate)
Movable proxy: Checked
vertex motion blur: UNchecked
object motion blur: Checked

https://www.youtube.com/watch?v=LnSuleJ7bMo

Can someone at Otoy please take a looks at this? or at least respond? I've posted about this in a couple other locations with no response.

Thanks!

What is wrong with this video?

[Karba]

I know typically people come across a problem and it's fixed by increasing that object's motion blur samples, however this is not that problem.

Firstly, this is an image from a series I'm doing about car wheels in motion. The result is spectacular and exactly what I needed (granted I had to use 128 motion blur samples).


But as I started messing with the kernel's motion blur duration setting, I noticed something strange.



It appears as though octane is taking it's samples during the entire frame, and then just using less of them based on the duration. I think this is a shoddy way of dealing with this. I think that instead (if possible) the motion blur samples should be distributed only amongst the time needed.

I am trying to reproduce the issue.
Could you send me the scene (just with this wheel)?
[email protected]

[bcravin]

Please refer to the still frames from the video that help show the problem

[Karba]

Please refer to the still frames from the video that help show the problem

And what is the problem with them?

[bcravin]

If that doesn't look wrong to you, I will post some stills from my current project to better illustrate the issue I'm having. But I keep running into the same issue where object with rotation that are moving in a straight line do not have proper motion blur. They have a weird curved motion blur.

You can see it in this picture I posted a few months back as well. The object is moving in a perfect straight line. no reason for the motion blur to look like that.

http://render.otoy.com/forum/download/f ... &mode=view

[Karba]

If that doesn't look wrong to you, I will post some stills from my current project to better illustrate the issue I'm having. But I keep running into the same issue where object with rotation that are moving in a straight line do not have proper motion blur. They have a weird curved motion blur.

You can see it in this picture I posted a few months back as well. The object is moving in a perfect straight line. no reason for the motion blur to look like that.

http://render.otoy.com/forum/download/f ... &mode=view

Are you expecting straight lines in motion blur of rotating objects? LOL