Rolling shutter for stills Sony A1/Nikon Z9/Canon R5 & R3

dpockett

EOS M50
Feb 23, 2020
30
11
Trying to get my head around the differences between A1/Z9/R5/R3 readout speeds.

My R5 has too much rolling shutter to depend on using the electronic shutter.

Does anybody know how all 4 stack up? And wondering what a rumoured R1 will deliver...
 

Joules

doom
CR Pro
Jul 16, 2017
1,734
2,140
Hamburg, Germany
Trying to get my head around the differences between A1/Z9/R5/R3 readout speeds.

My R5 has too much rolling shutter to depend on using the electronic shutter.

Does anybody know how all 4 stack up? And wondering what a rumoured R1 will deliver...
We should wait until all of these are out on the market. Then you can try to find a comparison between the modes that are relevant to you.

I guess you are talking about video. But for also for stills, the resolution and bit depth matter a whole lot. For 8K video for example, you can check the section about rolling shutter on this site and see that the R5 is a bit better than the A1 with a read out speed of 15.5 ms vs 16.6 with the A1 - even though the R5 is delivering the higher resolution DCI format in that test and the A1 isn't.

But without detailed tests for all the different modes, that alone doesn't say much. Theses cameras offer so many compromises between speed and rolling shutter already, and the Z9 throws in even more with its JPEG only modes.
 

dpockett

EOS M50
Feb 23, 2020
30
11
We should wait until all of these are out on the market. Then you can try to find a comparison between the modes that are relevant to you.

I guess you are talking about video. But for also for stills, the resolution and bit depth matter a whole lot. For 8K video for example, you can check the section about rolling shutter on this site and see that the R5 is a bit better than the A1 with a read out speed of 15.5 ms vs 16.6 with the A1 - even though the R5 is delivering the higher resolution DCI format in that test and the A1 isn't.

But without detailed tests for all the different modes, that alone doesn't say much. Theses cameras offer so many compromises between speed and rolling shutter already, and the Z9 throws in even more with its JPEG only modes.

Thanks Joules, no actually I am talking about stills (as per the title). But wouldn't the video RS equate to stills also?

I find the rolling shutter too severe for stills shooting on the R5, so I don't use the electronic shutter at all. But I was under the opinion A1 (approx 1/240 sec) was better than R5 (approx 1/60 sec). And from what I had gathered the R3 would be approx 1/180 sec? And Z9 claiming to be even better than A1 (approx 1/250 sec)?

JPEG only modes don't bother me, I wouldn't shoot high fps RAW files, I only use JPEG for sport.

I have a deposit on an R3, but with Nikon's looking better for less $$ and then the though of R3 becoming redundant in 12 months (R1 for more $$) I am unsure on the R3.
 
Last edited:

Joules

doom
CR Pro
Jul 16, 2017
1,734
2,140
Hamburg, Germany
Thanks Joules, no actually I am talking about stills (as per the title). But wouldn't the video RS equate to stills also?
Oh, sorry, I didn't read the title well enough.

Nonetheless, I don't think one can reliably conclude the read out speed for stills from the one for video. At lower bit depths, read out is sped up. As video often uses 12, 10 or even 8 bit it is at an advantage compared to stills.

Also, stills will read the full sensor height and therefore take more time, as they do not use the 16:9 format.

And with different cameras, the specific modes matter. I have not been able to find a good source for the bit depths of the Nikon Z9 formats. But if they were to be 14 bit instead of the 12 bit of the R5 for example, that should be factored into a comparison, not just the pure speed.

I find the rolling shutter too severe for stills shooting on the R5, so I don't use the electronic shutter at all. But I was under the opinion A1 (approx 1/240 sec) was better than R5 (approx 1/60 sec). And from what I had gathered the R3 would be approx 1/180 sec?
Do you mind sharing where you got those numbers? I find very few sites actually test the read out speeds at all and tests for stills are even more rare.

This German site has good numbers on the read out speed in milli seconds for video with the R5 and A1, but those show how close they actually are and do not reflect the numbers you quote (for stills?) :

8K 30p4kcrop
R51616 @ HQ, 10 @ 60p, 8 @ 120p/
A115.58.1 @ 60p, 6.2 @ 120p11.3 @ 24p, 10.5 @ 60p
Sources:

R5 Test Slashcam

A1 Test slashcam
 

Codebunny

Elil
Sep 5, 2018
995
1,044
Scotland
The stacked sensors will have significantly less, or no rolling shutter compared to a traditional sensor. The Z9 sits at the top just now with the A1 next then the R3, however, the R3 should have no perceivable rolling shutter in most situations. I have bought the Z9 myself, but If you are heavily invested in Canon glass it seems expensive to switch. Perhaps rent/borrow a R3 and see if it meets your needs before jumping ship?
 
  • Like
Reactions: EricN

dpockett

EOS M50
Feb 23, 2020
30
11
Oh, sorry, I didn't read the title well enough.

Nonetheless, I don't think one can reliably conclude the read out speed for stills from the one for video. At lower bit depths, read out is sped up. As video often uses 12, 10 or even 8 bit it is at an advantage compared to stills.

Also, stills will read the full sensor height and therefore take more time, as they do not use the 16:9 format.

And with different cameras, the specific modes matter. I have not been able to find a good source for the bit depths of the Nikon Z9 formats. But if they were to be 14 bit instead of the 12 bit of the R5 for example, that should be factored into a comparison, not just the pure speed.


Do you mind sharing where you got those numbers? I find very few sites actually test the read out speeds at all and tests for stills are even more rare.

This German site has good numbers on the read out speed in milli seconds for video with the R5 and A1, but those show how close they actually are and do not reflect the numbers you quote (for stills?) :

8K 30p4kcrop
R51616 @ HQ, 10 @ 60p, 8 @ 120p/
A115.58.1 @ 60p, 6.2 @ 120p11.3 @ 24p, 10.5 @ 60p
Sources:

R5 Test Slashcam

A1 Test slashcam

No worries, I thought you must have missed it.

A lot of factors, I see. Yeah it would be better waiting until they have been tested properly, as you mentioned.

Oh I have just found things here and there; dpreview forum, fstoppers, a couple of other sites I can't remember. But these aren't all confirmed so that's why I wanted to ask the question in here, if anybody knows accurate numbers.
 

dpockett

EOS M50
Feb 23, 2020
30
11
The stacked sensors will have significantly less, or no rolling shutter compared to a traditional sensor. The Z9 sits at the top just now with the A1 next then the R3, however, the R3 should have no perceivable rolling shutter in most situations. I have bought the Z9 myself, but If you are heavily invested in Canon glass it seems expensive to switch. Perhaps rent/borrow a R3 and see if it meets your needs before jumping ship?
Yes would like to try before buying, problem is stock will be limited so need to pre-order to get my hands on one. Yeah heavily in Canon, so switching isn't what I would want to do. I just hope the R3 doesn't break my trust like the R5 did with electronic shutter for sport.
 

Codebunny

Elil
Sep 5, 2018
995
1,044
Scotland
Yes would like to try before buying, problem is stock will be limited so need to pre-order to get my hands on one. Yeah heavily in Canon, so switching isn't what I would want to do. I just hope the R3 doesn't break my trust like the R5 did with electronic shutter for sport.

I honestly can't imagine you being disappointed by the R3, from the preview videos so far you may get a few squished looking balls but nothing like what the R5 produces. Will CPS not rent you one? Is there a 'famous' shooter on the Youtube that already has one in their hands shooting your subject and posting sample images?
 

Codebunny

Elil
Sep 5, 2018
995
1,044
Scotland
As you can sort them in such a strict way, you must have the read out speeds in mind. Could you please share them with us and mention where you got them from?

I have various quotes from the manufacturers and reviewers written down. If you hunt about you'll find currently plenty (including from Canon) that puts the readout speed of the R3 just behind the A1. Then the Z9 has been widely reported aa having a faster readout speed, fast enough that it no longer has a mechanical shutter at that. The number differences where something in the region of 4ms for the Z9 and 5-6ms for the A1. Sample images from the R3 show more 'slightly squished balls' than the A1 and so far the Z9 images have shown no rolling shutter artefacts at all, but the Z9 is brand new so the cases where it falls on its arse haven't been found yet.
 
  • Haha
Reactions: EricN

Joules

doom
CR Pro
Jul 16, 2017
1,734
2,140
Hamburg, Germany
I have various quotes from the manufacturers and reviewers written down. If you hunt about you'll find currently plenty (including from Canon) that puts the readout speed of the R3 just behind the A1. Then the Z9 has been widely reported aa having a faster readout speed, fast enough that it no longer has a mechanical shutter at that. The number differences where something in the region of 4ms for the Z9 and 5-6ms for the A1. Sample images from the R3 show more 'slightly squished balls' than the A1 and so far the Z9 images have shown no rolling shutter artefacts at all, but the Z9 is brand new so the cases where it falls on its arse haven't been found yet.
Alright. It just is a shame that there doesn't appear to be any nice, apples to apples comparisons or measurements of these figures. Hunting about for various bits and pieces isn't all that rigorous and even early previews can contain varying levels of truth (lile the discrepancy in Z9 buffer depth due to card variations discussed in thr other threads) or be tainted by marketing hype or manufacturer claims rather than first hand data.

I hope as ES use becomes more widespread, perhaps we'll get some better measurements under various conditions, like we have them for other aspects of sensor performance over at photons to photos.
 

Codebunny

Elil
Sep 5, 2018
995
1,044
Scotland
I hope as ES use becomes more widespread, perhaps we'll get some better measurements under various conditions, like we have them for other aspects of sensor performance over at photons to photos.

Some sort of machine that hits a golf ball with the same force in a repeatable manner and then we can measure the bowing.
 

Joules

doom
CR Pro
Jul 16, 2017
1,734
2,140
Hamburg, Germany
Some sort of machine that hits a golf ball with the same force in a repeatable manner and then we can measure the bowing.
I guess you're just joking, but there are far better and simpler ways to measure rolling shutter that involve no physical machinery anf no need for a human to perfectly reproduce any velocity.
 

AlanF

Stay at home
CR Pro
Aug 16, 2012
8,843
12,088
With the Z9, I would have rolling shutter as its massive weight causes my arms to buckle vertically down.
 
  • Haha
Reactions: SteveC and EricN

Codebunny

Elil
Sep 5, 2018
995
1,044
Scotland
I guess you're just joking, but there are far better and simpler ways to measure rolling shutter that involve no physical machinery anf no need for a human to perfectly reproduce any velocity.

I was indeed joking. Though it seems how bent a golf club is when striking the ball is the benchmark photo for rolling shutter.
 

neuroanatomist

I post too Much on Here!!
CR Pro
Jul 21, 2010
26,049
4,613
I was indeed joking. Though it seems how bent a golf club is when striking the ball is the benchmark photo for rolling shutter.
With bonus points for an oblong ball in the image… Or maybe the physical distortion of the ball following the club strike, which compresses it in the horizontal direction, would be counteracted by the electronic distortion from the rolling shutter and result in a spherical ball. :D
 

dpockett

EOS M50
Feb 23, 2020
30
11
Got my hands on a pre-production R3 today for a little while, tested out the electronic shutter. Super fast, has a shutter sound which is a good option (as well as silent). But the rolling shutter was my main area of interest. In some erratic testing, some quick panning around which probably was faster than any practical use - a direct comparison to the R5 resulted in a decent improvement. It is still noticeable occasionally, but a lot better than the R5.

I haven't done similar testing with an A1, but would be interested to see how it compares.

Bursting on cars going past and flicking back at the images, incredible to watch it in motion, full 30 fps.
 
  • Like
Reactions: kaihp