GH5 GH5 S - Weird stutter not sure what happen?

F

filmguy123

Veteran
Shot just a few low light shots the other night with GH5s. 29.97. Conformed to 23.976 in premiere for a very subtle slow motion.

180 degree shutter.

Looks like it was shot with high shutter. Odd. Kind of a weird motion to it. Any idea what could have happen?

Possibly related to the transform?

It was shot HD 10-bit vlog and the sequence settings match in premiere properly to 23.976
 
mixing video framerates with cine film rates have been the bane of many. they just dont play well together. it is better to go from 60 to 24 than 30 to 24, but that doesnt help you here. this is what i would do:
extract individual frames (like DNxHD) for the length of the slo mo. then import those frames as a sequence in a 24p timeline. i think this is better than forcing your NLE to stretch your footage mathematically.
 
What he is doing is a perfect mathematical calculation.

It's very common to go from 29.97 or 59.94 to 23.976. Both produce industry-standard 80% or 40% slow-motion results.

29.97 - 80% = 23.976

59.94 - 40% = 23.976

If there would be any issue(s), it would be from mixing integers/rationals (i.e. 60 --> 23.976 or 59.94 --> 24), but I doubt that would be visible.

I'm guessing it's something else.
 
Last edited:
By the way...just in case you were playing with your shutter speed at any point (while going back and forth between 30p and 60p) and you accidentally used 1/50 (if possible) while shooting 60p then this is the stutter you're referring too.

I don't remember the technical name for this, but the shutter needs to be AT LEAST 1/60 with 60p (for smooth slow-motion in a 24p timeline).

(This has burned me before...and I was using a Panasonic camera!)
 
Was hoping not to have to upload but if no one knows...

The shutter is set to 180d and I have it set to syncro scan so only 1 degree change at a time to avoid that error. Worst case I might have hit 178-182 if I rolled my thumb but I re-checked the camera and it was 180d when powering on.

The conversion in premiere is one I do all the time as Norbro said, but I hadn't yet done with the GH5S since I have shot very little with it. Never had this problem on the GH5. THe conversion was 29.97 to 23.976 -- not 30p to 23.976 or 29.97 to 24p as theorized.

The resultant footage looks just like fast shutter, but rather dramatic fast shutter, but I don't know how that would be possible given the above (I checked too, to confirm). Also given the low light scenario I don't think it could have actually been set to faster shutter without the light being gone.

Really odd.
 
filmguy123 said:
The resultant footage looks just like fast shutter, but rather dramatic fast shutter, but I don't know how that would be possible given the above (I checked too, to confirm). Also given the low light scenario I don't think it could have actually been set to faster shutter without the light being gone.

Really odd.
Click to expand...
Do you have a situation with a pan against a very light or contrasty backlight, that could cause it.
 
Filmguy, how did you do the framerate conversion?

Did you "interpret" the footage or did you actually "convert" it? Don't know if this could be the issue, but worth exploring. (You should interpret, to be clear.)
 
I did interpret footage via right click in premiere (my usual workflow which hasnt caused issue prior)

I'm wondering if this is anything specific with the GH5S?
 
filmguy123 said:
I did interpret footage via right click in premiere (my usual workflow which hasnt caused issue prior)

I'm wondering if this is anything specific with the GH5S?
Click to expand...
Why not Dropbox a few seconds of out of camera unaltered footage, without the footage all we can do is speculate.

For instance with ffmpeg you could use:

ffmpeg -i source.mp4 -ss 1 -t 2 -c:v copy -an cutfile.mp4

ss = start point
t = duration
 
Last edited:
filmguy123 said:
...

The conversion in premiere is one I do all the time as Norbro said, but I hadn't yet done with the GH5S since I have shot very little with it. Never had this problem on the GH5. THe conversion was 29.97 to 23.976 -- not 30p to 23.976 or 29.97 to 24p as theorized.

The resultant footage looks just like fast shutter, but rather dramatic fast shutter, but I don't know how that would be possible given the above (I checked too, to confirm). Also given the low light scenario I don't think it could have actually been set to faster shutter without the light being gone.

Really odd.
Click to expand...

I've done the same with my GH5S footage, with no issue.
 
Alright, I'll try and get some sample clips up here soon. I have no idea what the issue is then. Odd.

Luckily I'm not using this footage for anything, I was just playing at a friend's wedding reception to see how well the camera could handle awful low light and was super impressed... it could see things I couldn't with my eye, and with neatvideo the results are somewhat magical.

But the motion on the shots is weird and I'm still scratching my head here. More soon. Probably tomorrow.
 
Cary Knoop said:
Why not Dropbox a few seconds of out of camera unaltered footage, without the footage all we can do is speculate.

For instance with ffmpeg you could use:

ffmpeg -i source.mp4 -ss 1 -t 2 -c:v copy -an cutfile.mp4

ss = start point
t = duration
Click to expand...

Shhh! Don't tell people about all the cool things like lossless copy you can do with good old command line FFMPEG. You'll ruin it for us guys that use it as their secret weapon. :)

Jim (who may or may not sometimes batch clone the video and replace his GH5s original audio with with noise reduced audio because he's too "lazy" to link an extra audio file to a clip in his NLE) Arthurs
 
Jim Arthurs said:
Jim (who may or may not sometimes batch clone the video and replace his GH5s original audio with with noise reduced audio because he's too "lazy" to link an extra audio file to a clip in his NLE) Arthurs
Click to expand...
Also that can be done using the ffmpeg command line and lossless! :)
 
Cary Knoop said:
Also that can be done using the ffmpeg command line and lossless! :)
Click to expand...

Yes, yes of course, that's exactly what I do with it, and why I said it. I was trying to say with "humor".. :) One of the many valuable daily uses of FFMPEG.

Jim (GUI's like Handbrake for FFMPEG are feature stripped half-measures used only by people living in fear of typing a little text) Arthurs
 
Jim Arthurs said:
Yes, yes of course, that's exactly what I do with it, and why I said it. I was trying to say with "humor".. :) One of the many valuable daily uses of FFMPEG.

Jim (GUI's like Handbrake for FFMPEG are feature stripped half-measures used only by people living in fear of typing a little text) Arthurs
Click to expand...

Don't worry about people living in fear of typing text, far from the only ones. I use FFMPEG command line for years now as well, mkvtoolnix, IfoEdit, AviSynth, x264, x265, scripting, batch, compilers, simulators and debuggers, many others I can't count. Utilities like those are among the reasons I remain in the PC camp.
 
But back on point, the things I've seen as possible sources of uneven motion handling, stutter and judder are sometimes caused by cmos skew (rolling shutter artifacts), uneven cadence, 60p or 30p to 24p, fast shutter and image stabilization that jumps, either OIS or IBIS both do it. It happens to me on GH5 and GH5s.

And then sometimes, the reason is no more complicated than I bumped the tripod or monopod of the gimbal or the camera. If you don't remember doing it, can drive you crazy trying to figure out what happened.

GH5S on the gimbal with no camera or lens stabilization, you can still have micro-jitter, the gimbal is not perfect in all axis after all. With gimbal'd GH5 and stabilization off, the sensor can still bounce around and if stabilization is on, can be super smooth or a wobbled mess. Overall, I prefer the GH5s on the gimbal with its fixed sensor, and stabilize whatever micro jitters if any in post. Resolve can add temporal smoothing to fast shutter and image stabilization for shake.

Might never notice these problems in 30p/50p/60p. For 23.98 have to be extremely careful.
 
You must log in or register to reply here.
Back
Top