Measuring rolling shutter: put a number on this issue!

Yes, it's not as bad as some people think. If you ask me: those people only use Canon cameras (among those, it's clearly the worst). But yes, Samsung and, above all, Sony, don't seem to mind so much about slow rolling shutter.
 
Samuel, I once again must state I love what you have created with this thread. Prevented me from getting a A6300. No way I'm getting a 4K camera that has worse rolling shutter than my NX1.

Will send you D750 shortly. It's another good performer I think.
 
Thanks. The funniest part is that I dislike rolling shutter much less than most people, but nobody was doing this, so, it might as well be me, since I also see the usefulness of the table, and well, each line on that table only takes me 10 or 15 minutes. So: glad to help :)
 
In order to get a half-decent measure of this, I would need to get a very short clip from each camera, shot following these instructions:
* camera on a tripod, well levelled, set at 24p and 1/50s shutter
* a vertical line somewhere in the frame, crossing it completely from top to bottom (it can be a pole, or a corner, or a door, etc; if it is clearly tilted it will distort my measure)
* start recording, do a quick pan left at constant speed, then a quick pan right at constant speed, repeat, reapeat, then stop recording
(that's a total of 3 pans left and 3 pans right, which will help average off the measurement errors)

It’s not clear to me how you could measure RS without knowing the speed at which the person was panning, and what lens was used. What am I missing here?
Do you have a detailed description of your methodology? Thanks for all of your work on this.
 
I use the inter-frame displacement to put the intra-frame displacement in perspective and get a measurement in ms.

- I look at a frame, and measure the skew in it (in pixels)
- I look at the next frame, and measure how much the top of the reference item has moved between these two consecutive frames (in pixels)
- I look at the second frame again, and measure the skew in it (in pixels)
I take the average skew of those two frames (that's important because it corrects for accelerations or decelerations of the panning speed), and divide that by the inter-frame displacement. Multiply that by 41.7 (that is 1000/23.976; use a different number if your test clip is not 23.976 fps) and you have a rolling shutter measurement in ms.

Then do it again with a pan that goes to the other side, just in case your reference wasn't vertical, and average those two measurements. Then repeat everything once or twice because this reduces measurement error.

The resulting number is in ms and does not depend on focal length or panning speed (though with a faster lens and a faster pan the measurement is more accurate; on the other hand, faster pans tend to have more acceleration or deceleration in panning speed, so I don't ask specifically for fast pans).
 
I'm wondering if I could find a way to attach these numbers to a method of fixing the RS in post. What I mean is: if people could use this chart and go into Premiere's Rolling Shutter repair and know what number they could put in to do a basic rolling shutter fix. I'm going to find the numbers to fix the RS in the NX1 in 4K and 1080 and see if I can make an equation for others to use.
 
I've never used that tool, but my guess is that you're right: that should work (easy way to know if it will: does each camera always work best with a specific setting?)
 
Any chance of getting a rolling shutter reading on the Black Magic Micro Cinema Camera? Disappointed that they dropped global shutter as an option.
 
Nice.

This would be the action camera to cut with high-end cameras: tiny, very good rolling shutter, very good DR, no compression issues, sharp enough.
 
Back
Top