GH4 GH5 4K/60P stabilized and slowed

J

joe12south

Veteran
[video]http://blog.josephmoore.name//luts/neuman/slomomotiontest1080-24p.mp4[/video]

Many thanks to Luke Neumann for the sample footage!!!

I wanted to see how far I could push the GH5's new 4K/60P recording and the new IBIS stabilization. This handheld clip seemed like a decent torture test for both. Using FCPX, I slowed it down 25% (optical flow) and pushed stabilization effect to the max.

Artifacts are visible, but to me it's quite usable. If it had been shot with a shutter speed planning to be slowed down even more, it would have worked out even better. There are also better stabilizers and frame interpolators, as well.
 
Is there a reason why you just didn't drop the footage in a 24p timeline instead of trying to use optical flow to time remap? I think the quality would have been better this way, as there is no chance of optical flow artifacts.
 
estarkey7 said:
Is there a reason why you just didn't drop the footage in a 24p timeline instead of trying to use optical flow to time remap? I think the quality would have been better this way, as there is no chance of optical flow artifacts.
Click to expand...
I did. And then I slowed it even more. Just like I took in camera stabilized footage, and stabilized it even more. That was the point of the test, to see how much further it could be pushed in post when starting from a better place than the GH4.
 
I really prefer to stay at 4K when possible, so if I can synthesize 120fps from 60fps, it's a useful trick. And if I can synthesize 360fps from 180fps in 1080p, well then that really will be something.
 
[video]http://blog.josephmoore.name//luts/neuman/360fps_test.mp4[/video]
I was playing with Luke Neumann's 1080/180fps footage and thought it might be interesting to share:

1. Slowed down to 360fps. Note how optical flow really loses it once the motion speed exceeds what the shutter can capture. If you're shooting with the intention of slowing in post, always got to remember to set the shutter for the target fps, not the real fps, for best results.

2. Heavily graded. Note how the internal 8-bit V-LOG L doesn't hold up to grading very well. The same h.264 encoding errors (look at the back wall) that make V-LOG L so hard to work with on the GH4 haven't magically gone away. Save V-LOG L for 10 bit recordings.

3. Hard to say definitely from these samples, but resolution does seem to be better than the VFR modes on the GH4. Until we have samples with charts or other high frequency data, the jury is still out as to what amount of detail we lose in these modes.
 
You must log in or register to reply here.
Back
Top