Difference between shooting slow motion with higher fps vs during post production

[swan n]

What are the advantages and disadvantages of shooting with higher fps vs making the footage slow motion in post production. If there's a link to this subject already, please let me know.

 

[Les Dittert]

Software slow-motion easily gets tricked and confused by one object moving over another, or one object being revealed in front of another.

 

[Phil Holland]

There are many advantages to capturing it right in camera. While software solutions have really evolved in the last few years, they aren't good for every situation. Specifically when it comes to high spatial frequency details and more complicated scenes (like what Les described). Also when dealing with slow motion past 60% I find it can get hairy and can produce some gnarly artifacts. Sometimes software gets "confused". Mainly due to the fact that the algorithm that is trying to create new frames in between existing frames can't "predict" a good result.

By capturing the slow motion in camera you have more data, detail, and information. You also don't have to spend time later trying to fix the shot in post via applying paint fixes, which can be time consuming.

Two tests I did a while ago that produced really bad results via post slow motion were:
- slow motion of fire remapped to 20% captured speed
- a girl spinning in a long dress with a "wavy" pattern slowed to about 35%

Both of those created unwanted artifacts and the "organic" nonlinear nature of the movement was something that the software couldn't recreate.

For this type of work I always try to get it in camera.

 

[Emin Mihat Hye]

Exactly as Phil said. Capturing in camera gets the actual full frames of the motion (i.e. 60 frames, slowed down to 24 as the final output), whereas to get the same effect in post with 24p footage, the program (such as Twixtor in AE) is actually guessing the filled in frames based on vectors in the images before. Hence the reason we get the artifacting. As stated before, the only times it works is when there are solid colored backgrounds with simple silhouettes moving in the foreground (i.e. skateboarders on a flat sky background such as in the following video: http://vimeo.com/13557939).

 

[Jeffrey Loewe]

Exactly as Phil said. Capturing in camera gets the actual full frames of the motion (i.e. 60 frames, slowed down to 24 as the final output), whereas to get the same effect in post with 24p footage, the program (such as Twixtor in AE) is actually guessing the filled in frames based on vectors in the images before. Hence the reason we get the artifacting. As stated before, the only times it works is when there are solid colored backgrounds with simple silhouettes moving in the foreground (i.e. skateboarders on a flat sky background such as in the following video: http://vimeo.com/13557939).

Without even going frame by frame I can still pick out a artifact in nearly every shot in that video. You can tell he used some Blur effects to help fake a focus pull to cover some up. Only on a few shots but it helped.

I've done numerous tests with the Scarlet and never found anything remotely pleasing while trying to twixtor it. Worth renting or buying the right camera for the job.

 

[Emin Mihat Hye]

Ya, there is some artifacting, but with H.264 compression and compressed 7D footage its understandable, especially considering it is a low-budget alternative. Myself and plenty of others find the footage incredible. He even did another piece and used Particle effects to cover up artifacting in his newer video: http://vimeo.com/17439665

I have yet to do any tests on the Scarlet, but I had fiddled around with Epic footage a year ago with some decent results. However, not even close to using a high speed camera. Definitely worth renting/buying the right camera for the job.

 

[Jason L]

I'd recommend to always do it in camera, but the downfall to that, is you'll need a lot more light because you'll have to increase the shutter to go with the increase of frames in the overcrank.
It can be manageable to do it in post, but like others said, there will be artifacts after a certain point.

 

[Pablo Moreno]

All of the above could be true, but also false. The people behind the Matrix saga, used to use slo-mo software emulators with pretty good results back in the nineties, and after that, it came Twixtor plugin and some others, that can even simulates something as difficult as water in motion and fire. But of course, it depends a lot on your footage and the way you shoot. If you have to many blurred frames on your footage, that the software cant resolve, then it cant do miracles. So if you plan ahead and be creative, you can get fantastic results. The two videos below may be a little old, but shows what Twixtor plugin can do.

http://vimeo.com/13557939#

http://vimeo.com/8352041

EDIT: There is one more here, I guess should be better software today..

http://vimeo.com/21050859#