8K and Gimbal Stabilzers. Changing It Up?

[Terry VerHaar]

With all the talk about 8K and with the realization that it is either here, or coming very soon, for many RED shooters, I got to thinking. There are many reasons to like 8K. Of course, the most obvious benefit is the absolutely stunning images that come from downsampling 8K acquisition to HD or UHD delivery.

In the new RED TECH series, the difference between shooting 8K and delivering either HD or UHD is dramatically illustrated. But, they also mention the obvious fact that you can crop and/or reframe the image if desired. With an image that is, according to the video, 17 times greater than HD and several times larger than UHD, you can even create "fake" pans, tilts and zooms/dollies in post. Not necessarily the ideal way to get these things but certainly available if it can enhance the telling of the story and you didn't have the time, budget or foresight to do it on set.

Now, the other potential benefit is the ability to stabilize shaky footage and that brings me to my question...

With all of this extra real estate to work with for doing post stabilization; and given the seemingly improving quality of the stabilization software, are we getting to a point where gimbal stabilizers (Movi, Ronin, Helix, etc.) are not necessary - or at least needed less often? I guess the same question even applies to steadi-cam as well.

If one can handhold a move as we follow someone walking, or moving through a maze, or following a kid on a Big Wheels tricycle (I saw the Kubrick exhibit yesterday in San Francisco - LOL), with "reasonable" steadiness at 8K, can we then just go into post and smooth out the footage?

I am sure there are arguments against this approach - not quite as good, giving up the resolution for downsampling, etc. And I am not suggesting a project is going to go out of its way to shoot 8K just to save the rental cost of a stabilizer. But, if you have other reasons to shoot 8K (or even 6K), is this a bonus benefit?

I am interested what others think and what approach people are inclined to take. Would love to hear commentary.

Cheers.

 

[Minu Park]

You are super right, but. I prefer to have solid stabilized 8k 2 hour footage with movi or Steadicam than, 2 hours of HD image with post works that will cost me similar and spending similar memory.

It's just an option, but I actually tend to shoot straight instead of relying on post...

I don't know... I may be very wrong...

 

[Andy Roberts]

One thing post stabilization doesn't fix is the blur caused by the unstable movement of the camera. Inherently stable shots are sharper. Stabilizing in post is generally a band-aid for worst case scenarios.

 

[Josh Becker]

One thing post stabilization doesn't fix is the blur caused by the unstable movement of the camera. Inherently stable shots are sharper. Stabilizing in post is generally a band-aid for worst case scenarios.

If it's intentional and planned, though, you could increase your shutter speed to minimize motion blur, then re-add it using a plugin like ReelSmart MotionBlur.

I actually found this music video the other day that pulls off a pretty cool concept with what appears to be pan+scan on a wide 8K shot. Excited for the new creative doors that could open from a camera with capabilities like this!

https://vimeo.com/183836472

 

[paulcurtis]

Stablisation doesn't work too well with lots of different movement within the scene, foreground moving one way, background the other. To stabilise properly you'd need to separate different layers of movement and of course there isn't really enough information in a 2d frame to do that off a single frame, so you're looking at adjacent frames to interpolate from and frankly unless the shot is pretty simple it's a right pain.

For example on a wide lens, hand holding low with a wide scene in front, cliffs and sea you'd think that would be an easy task? But the wide lens emphasises perspective and stabilisation is really a 2D process so very quickly it starts falling apart, the vast majority of solutions basically corner pin and warp a frame. But in scenes with a lot of parallax depth that's just not enough.

The trick of increasing shutter speed to counter the cameras own motion blur can work in controlled circumstances but at the same time adding motion blur after can be tough with lots of movement in different directions (so like the example above foreground and background moving counter to each other). The software has to build up an image of how the pixels and vectors are flowing and the edges of these opposing movements are usually not handled that well.

So stabilise at source will *always* be a better solution

cheers
Paul

 

[Terry VerHaar]

Thanks for the thoughtful responses, guys. I agree that it is ALWAYS better to shoot exactly what you want whenever you can. I guess I am thinking in terms of flexibility and what methods might be available as backup measures if the day gets pushed and you can't get the shot you want, the way you want it. Certainly, it would work in some situations more than others. I was imagining something like just walking slowly next to or behind the talent, following them as they walk; or something like the Big Wheels scene I mentioned in The Shining.

I liken it somewhat to "punching in" for a close-up with a 5K or 6K resolution when you shot a wider shot. It's always an option but not necessarily the ideal solution.

One thing I can say is that I am looking forward to testing and experimenting when/if I get my hands on 8K!!

Cheers.