How to Setup Slow Shutter?

[AlexeyAdamitsky]

I've been exploring and testing this for quite a while now and I can figure out how to produce slow shutter effect in Octane for Houdini. I don't know if it's even possible in other DCC integrations.

See the attached files for the examples of what I'm trying to achieve. So far I wasn't been able to do so. If I increase the shutter time value beyond 1 it just makes the motion blur stronger between frames but it doesn't capture the motion blur between a range of frames. Maybe I'm settings it up wrong but I've tried many configurations and I can't do what I need.

In photography, you just control this with shutter speed/shutter angle but I can't figure out why it's not working in Octane.

I'd love some help.

[frankmci]

I've been exploring and testing this for quite a while now and I can figure out how to produce slow shutter effect in Octane for Houdini. I don't know if it's even possible in other DCC integrations.

See the attached files for the examples of what I'm trying to achieve. So far I wasn't been able to do so. If I increase the shutter time value beyond 1 it just makes the motion blur stronger between frames but it doesn't capture the motion blur between a range of frames. Maybe I'm settings it up wrong but I've tried many configurations and I can't do what I need.

In photography, you just control this with shutter speed/shutter angle but I can't figure out why it's not working in Octane.

I'd love some help.

I'm guessing it's because units of time are an artificial, arbitrary human concept, but most software does not treat it as such. For most render engines, exposure wise, there is no such thing as a unit of time that truly lasts more than one frame. I think to get the effect you want you will need to not just increase exposure, but reduce the number of frames per second, too. If you want a 0.25 second exposure, you will need to render your animation at no more than 4 fps.

I just woke up, so I hope that makes sense. I'll come back after I have some coffee.

[skientia]

There is only a thin correlation between a real photography and a renderer. The concept of exposure differs between the two — exposure time is non-existent in a renderer, "exposure" is arbitrary.

One way to yield this result in a physically plausible manner is by accurately (enough) animating the object and/or camera and facultatively via a double exposure method, akin to real photography, for more control. The emphasis is on the animation being the most challenging and fundamental part (the accumulations of all the imperfect movements, shaking, twisting).

The double exposure approach is also reminiscent of another single-take slow shutter technique which involves a early static stabilization after engaging the shutter button, followed by a sudden movement resulting in a defined photograph complemented by motion blur "filaments" (all in one press of the shutter button) analogous to the work of "wrapped nil", pushing the aesthetic further.

[AlexeyAdamitsky]

I'm guessing it's because units of time are an artificial, arbitrary human concept, but most software does not treat it as such. For most render engines, exposure wise, there is no such thing as a unit of time that truly lasts more than one frame. I think to get the effect you want you will need to not just increase exposure, but reduce the number of frames per second, too. If you want a 0.25 second exposure, you will need to render your animation at no more than 4 fps.

Ok, I think it's a lead I haven't explored yet. I need to see what I can do with this information and if it helps. Thank you for the suggestion anyway.

There is only a thin correlation between a real photography and a renderer. The concept of exposure differs between the two — exposure time is non-existent in a renderer, "exposure" is arbitrary.

I understand that but renderers like Arnold and Redshift, for example, are following the photography conventions pretty well in language and effects, mimicking the behavior of real photography really well. Octane is pretty good at this too, except this particular case which I can't crack yet.

One way to yield this result in a physically plausible manner is by accurately (enough) animating the object and/or camera and facultatively via a double exposure method, akin to real photography, for more control. The emphasis is on the animation being the most challenging and fundamental part (the accumulations of all the imperfect movements, shaking, twisting).

I'm not familiar how to do the double exposure technique with Octane. Where can I learn more about it?

[skientia]

the double exposure technique with Octane.

There is nothing specific or exclusive to Octane as it dates from far back and essentially consists of merging two takes in one, granted the physical digital camera features this function for in-camera (two distinctive raw files processed in one double-exposure JPG). Film cameras do not require a in-camera feature as the process can be achieved mechanically.

Googling "Double Exposure Photography" should yield numerous results. None will be linked in this reply as the first page are pages containing obnoxious ads and data-seeker all over the place.

The concept can be translated in post (from two distinctive render output files, one with motion blur and the other without) with proper compositing that complies with the linear workflow (exporting in a linearly encoded EXR and by avoiding non-linear "closed domain" "destructive" operations), more on that here).

[skientia]

I'm not familiar how to do the double exposure technique with Octane.

There is nothing specific or exclusive to Octane as it dates from far back and essentially consists of merging two takes in one, granted the physical digital camera features this function for in-camera (two distinctive raw files processed in one double-exposure JPG). Film cameras do not require an in-camera feature as the process can be achieved "mechanically".

Googling "Double Exposure Photography" should yield numerous results. None will be linked in this reply as the first page are pages containing obnoxious ads and data-seeker all over the place.

The concept can be translated in post (from two distinctive render output files, one with motion blur and the other without) with proper compositing that complies with the linear workflow (exporting in a linearly encoded EXR and by avoiding non-linear "closed domain" "destructive" operations), more on that here).