Measuring rolling shutter: put a number on this issue!

Thanks!

You don't need to see the numbers to know that it's a very, very, very good rolling shutter, but here we go: 8.3ms (three measurements: 7.8-8.5-8.6)
 
Thanks! Fuji XT2 added to the table. 29.5ms in 4k and 16.9ms in 1080p. Boost mode on or off doesn't make a difference (neither at 4k nor at 1080p).
 
Marc Jones sent me a6000 and G7 clips, but they're not shot with the best possible setup so measures are a bit less accurate than usual. I'll post the results here but won't add them to the table on page 1.

a6000 is pretty good: somewhere between 11 and 14 ms

G7 is not too bad in 4k mode (somewhere between 26 and 29 ms) and also prety good in 1080p mode (15-17 ms). I'd say it's the same as the GX85, which you can find in the big table.
 
Did anyone ever submit A7RII FF 1080p footage for measurement? I see numbers for every mode but that (4K both sizes, 1080p crop).
 
I didn't get that clip, no
(no idea why but maybe they didn't send it because of aliasing/moire? I really don't know if it's there but it wouldn't surprise me).
 
Samuel H said:
I didn't get that clip, no
(no idea why but maybe they didn't send it because of aliasing/moire? I really don't know if it's there but it wouldn't surprise me).
Click to expand...

1080p Full Frame is decent, crop mode has very bad image quality. Will send at test of FF mode shortly.
 
Samuel, I just got the 1DX mkII, I can provide clips if you want :)

What do you need? Straight, contrasty line and tripod, ramp up the panning speed? All I've got is a 50mm right now, is wider needed?
 
I'm not particularly interested (it's out of my budget) but I wouldn't mind adding it if you can shoot the clips. Straight vertical line and constant-speed pans on a tripod are best. 50mm is OK, focal length doesn't matter (whatever it is, my measurement will compensate for that).
 
hempo22 said:
Here you go: https://www.dropbox.com/s/4b86rd1cpgr6n7n/Canon 1DX2_RS test.zip?dl=0
Click to expand...

Thanks, that was a very well shot test.

The 1DX2 is a very good performer: 14.7 ms in 4k (both 24 and 60 fps), 11.1ms in 1080p24, and 6.7 ms in 1080p 60 and 120 fps.

Also: I found a typo on my excel file: the last 5 measurements were slightly wrong because I used 25 fps in the excel but they were shot at 24 fps. Not a huge difference, but well, I'm updating the table in the first page to fix that.
 
GH5

GH5

If you look at page 7 of this link:
http://www.personal-view.com/downloads/GH5_Presentation.pdf
It gives the GH4 4k readout time as 22ms (a close agreement with the measured 22.5ms) and says the GH5 will be about 15ms.
It also says "Suppressed Rolling Shutter Distortion when using an electronic shutter or while video recording" (I was told this is in software) which might make measuring the rolling shutter a bit more interesting.
 
Glad to have another confirmation that my measurements are accurate, I'll add it to the table.
I also added the GH5 number, though I'd wait for people to have the GH5 in their hands before trusting any marketing claims. If it's 15ms it's very good indeed. The software thing looks like something I'd rather do in post, though, unless it's using an accelerometer and gyroscope to get an idea of how it has to correct the image.
 
By software I meant inside the camera, in case it wasn't clear. However I assume they can only fix camera movement, which I wonder how much point there is to, with a fast read-out and a limit to how fast people pan who aren't doing rolling shutter testing? Plus presumably they need some extra pixels to the side of the image to do that and it's supposed to be doing full-sensor readout? (Of course even if they could judge subject movement, rather than just camera movement, the background won't be moving so presumably they are stuck trying to fix that.)
 
You must log in or register to reply here.
Back
Top