a7s rolling shutter test please

Yes, could someone please, pretty please give Samuel what he needs to put a number on the rolling shutter. That way maybe we can settle the subjective it's worse than, better than opinions once and for all.
 
Keep an eye on the email address listed in your RS thread, I'll be sending you a dropbox link shortly. Footage is sooc, XVAC, 24p, 1/50 with the "standard" profile.
 
Yes FF, I was using the FE 28-70, zoomed to about 50mm, and F/16 IIRC - because its really bright and I'm revamping my ND situation to use threaded filters with xume adapters since I hate Vari-ND's. Aperture shouldn't impact anything though.
 
So, thanks to photochris, we have numbers.

By the power invested in me, I hearby declare the a7s a slow motherf###er. But not horrible. 30.5ms, the slowest RS I've measured so far, about the same as my 5N but actually slightly worse. Close but not as bad as the D90, which was 33ms according to other sources.

I think I can live with this pretty happily. If I ever need faster RS I can always use APS-C (which I think will be 20ms), since I'm unlikely to need both fast readout and lack of aliasing/moire.

Full table here

Also: remember when I measured RS on the a7s based on a ghosted still from a shot of a shot from NAB? that came out at 30.9ms, so pretty damn accurate.
 
Thanks Samuel and Photochris. Nice to have a real number on it, even though it was even worse than I feared. Not sure I can deal with another D90 style RS. Still have that camera and at that level of RS I feel it does become limiting for hand held shooting without a solid rig.
 
Good to know. For handheld I'm going to be using my cheesycam modded steadicam smoothie, but I don't shoot a lot of action so RS has never been an issue for me.
 
Shouldn't the RS at 4K be the same as at 1080p, since the camera does a full sensor readout with no line skipping in both modes? Or am I missing something?
 
Samuel H said:
So, thanks to photochris, we have numbers.

By the power invested in me, I hearby declare the a7s a slow motherf###er. But not horrible. 30.5ms, the slowest RS I've measured so far, about the same as my 5N but actually slightly worse.
Click to expand...
.

That's not good. I won't be selling the MK3 anytime soon.
 
I've gone back and forth on this thing since NAB. Changing my mind almost everyday. I am beginning to think the cons are outwaying the pros here. It's great looking and a nice piece of tech for sure, but the quirks it comes with is starting to look equality as annoying. I was really excited when it was first announce because it sounded like a hurdle past other cameras limitations and really brought a lot of desired elements together. As usual though, it's a trade off. I know every tool for every job. This particular tool is starting to become clear of what it can and can't do. So far, not pulling the trigger yet.
 
I don't think I could rely on it as an A cam or an only cam but I haven't cancelled my preorder because I think I will need it for its clean high ISO. It's a lot cheaper than a white truck. The rolling shutter won't be a problem because the shots I'll be using it on are mostly wide on a tripod or steadicam. I hate rolling shutter but I had a D90, a 5D MkII, and a Nex 5n so I know how to avoid it.
 
philiplipetz said:
Samuel, still excited about the A7s?
Click to expand...
Definitely. My 5N is just as bad in this aspect and I'm really happy with it, this one is sharper, has a cleaner image (no aliasing/moire), and uses a better codec. And, above all, it has hugely better DR thanks to slog2.

Plus, as I said, whenever I need faster RS, I can just shoot in APS-C mode, which should be 20ms. (BTW, 4K should be 30.5ms too, that is the sensor integration time, it doesn't depend on what happens later on with the information; the only reason for it to be even slower in 4K would be if 1080 uses in-sensor binning, which seems not to be the case, since that would put 4K integration time at an absolutely unbearable >60ms)

I can't think of a scenario that requires fast RS but also will show the IQ issues of the APS-C mode. A shot that starts stationary and then turns into shakycam without a cut in the middle? Odd.
Tha a7s is going to be my A-cam. Then my B-cam with good RS will be the a7s-APSC. They match each other pretty well :)


jideobi said:
I am almost certain that rolling shutter is better in APS-C crop mode
Click to expand...
Care to shoot a test and post the original file somewhere so I can check it exactly?

* 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)
* 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)
 
Last edited:
Cosimo Bullo said:
Worse than the 5d2... don't want to move backwards. Do you guys have a measurement for the D90? Maybe it was around 40?
Click to expand...

The D90 was about 33 ms, the 5DII about 25ms, and the A7s is 30.5ms. I'm frankly a little surprised that Sony deemed that good enough to produce a camera in 2014. If it had been roughly like the 5D III (20ms), I would have been perfectly happy.

Some Sony representatives apparantly said at one of the trade shows that the RS should be tuned to at least match the 5D III in the final production/ firmware version. I guess that was either just to deflect the issue, or perhaps any faster sensor readout turned out to generate to much heat.
 
You must log in or register to reply here.
Back
Top