Fastest EFCS shutter speed usable on the a7RII?

[JimKasson]

Yesterday I looked at shutter travel time for silent shutter operation on the Sony a7RII. Incidentally, we noted that the electronic first curtain shutter (EFCS) pretty faithfully emulated the acceleration of the mechanical second curtain, at least at 1/500 second.

On the Nikon D810, EFCS stops working at shutter speeds above 1/2000. On the Sony alpha 7 cameras with EFCS, Sony advises against using the capability at fast shutter speeds, but let's the user decide how far to push it.

So, let's push it and see what happens when we take pictures of a high-frequency triangle wave on an oscilloscope with the time base set to 500 microseconds (us) per division.




1/1000, mechanical shutter




1/1000 EFCS



You can see that the EFCS image has a little less exposure at the top of the image, but you probably wouldn't notice that in a real photograph. The all mechanical image is pretty even.




1/2000 mechanical




1/2000 EFCS



Now the EFCS exposure is suffering from nonuniformity noticeably, but would probably be fine for many purposes.




1/4000 mechanical




1/4000 EFCS



The EFCS exposure at the top of the image is about 1/3 of that at the bottom. This should normally be avoided.




1/8000 mechanical




1/8000 EFCS



There is essentially zero exposure at the top of the EFCS image. Not useful.

The synching of electronic and mechanical behavior over time, temperture, wear, and sample variation is an iffy business.

I used to say that users should avoid a7x EFCS shutter speeds of over 1/2000 second. Based on these results, I'm dropping that to 1/1000.

--

the last word the last word - Photography meets digital computer technology. Photography wins -- most of the time.

Photography meets digital computer technology. Photography wins -- most of the time.

blog.kasson.com

 

[Snikt228]

Are these results any different on the normal A7? I've taken many pictures at 1/2000 and 1/4000 with EFCS enabled and I've never noticed any issues

 

[JimKasson]

Are these results any different on the normal A7? I've taken many pictures at 1/2000 and 1/4000 with EFCS enabled and I've never noticed any issues

I don't know, and I don't have a visible light a7 to test. Sorry.

You could do something as disposive by taking a picture of an evenly lit white wall at fast shutter speeds and reading the average values of raw channels in 200x200 squares at top and bottom. Compare with pictures made with the mechanical shutter and/or at slow shutter speeds.

That way you wouldn't need a 'scope to do the experiment.

A key element is how much tolerance you have in your images for non-uniform exposure. I imagine that darkened tops could improve some images.

Jim

 

[exdeejjjaaaa]

Sony advises against using the capability at fast shutter speeds, but let's the user decide how far to push it.

but Sony does not provide any automation (like a setting in menu/setup saying disable EFCS starting from 1/xxx, etc) to disable that so it is a big issue in many situations different from slow methodical shooting, does it not ? so when you face the illumination that you can't control and your range of your exposures changes wildly you can only do a switch between C1/C2 modes or so, no ? or you have to dig into menu settings = loss of time

 

[sybersitizen]

You might find different results depending on the lenses used and their particular telecentricity characteristics. See the theory offered in this article:

http://www.mhohner.de/newsitem2/efcs-update

 

[rrccad]

Sony advises against using the capability at fast shutter speeds, but let's the user decide how far to push it.

but Sony does not provide any automation (like a setting in menu/setup saying disable EFCS starting from 1/xxx, etc) to disable that so it is a big issue in many situations different from slow methodical shooting, does it not ? so when you face the illumination that you can't control and your range of your exposures changes wildly you can only do a switch between C1/C2 modes or so, no ? or you have to dig into menu settings = loss of time

yeah I'm not a fan of nothing having limiters or having it automatically turn off.

a good case for that is during AEB, where your shutter speed on your bracketed shot may creep up to 1/4000th of a second and you have ECFS enabled to handle the slower shutter speeds.

in Auto-ISO mode it shouldn't allow for extremely high shutter speeds I would hope.

 

[JimKasson]

You might find different results depending on the lenses used and their particular telecentricity characteristics. See the theory offered in this article:

http://www.mhohner.de/newsitem2/efcs-update

Good point. I used a Zeiss 100mm f/2 Makro-Planar, focused quite close (it's a five inch 'scope), so the incident rays were fairly perpendicular to the sensor surface. I don't have a convenient way to do this test with a symmetric short lens.

Jim

 

[Lightshow]

I've had zero issues using EFCS on my NEX-7 the entire time I've had it, I've only used the mechanical shutter for testing, I know some users had visible falloff with theirs, I haven't seen any with mine.

I'm disappointed they still haven't got a usable high shutter speed yet for FF.

 

[jalywol]

I was going to ask about the lens component. I've noticed a problem with legacy manual lenses going above 1/2000, but with native FE lenses, I've gone higher than 1/2000 (when I wasn't paying attention ) and had no easily noticeable degradation.

It would be interesting to test this....

-J

 

[JimKasson]

I was going to ask about the lens component. I've noticed a problem with legacy manual lenses going above 1/2000, but with native FE lenses, I've gone higher than 1/2000 (when I wasn't paying attention ) and had no easily noticeable degradation.

It would be interesting to test this....

I have no native FE lenses that will focus close enough to do the 'scope experiment with them. I will accept an indefinite loan of a 90 FE macro, however.

Jim

 

[jalywol]

I was going to ask about the lens component. I've noticed a problem with legacy manual lenses going above 1/2000, but with native FE lenses, I've gone higher than 1/2000 (when I wasn't paying attention ) and had no easily noticeable degradation.

It would be interesting to test this....

I have no native FE lenses that will focus close enough to do the 'scope experiment with them. I will accept an indefinite loan of a 90 FE macro, however.

Jim

--
http://blog.kasson.com

Well, if you'll treat me to round trip plane tickets from Upstate NY, you for sure can borrow my 90mm lens for the tests

-J

 

[steelhead3]

Isn't the shutter only good to 1/2000 with the nex 7?

 

[Horshack]

Was trying a bunch of experiments tonight, trying to reverse engineer some nuances in Sony's EFCS implementation that I'll discuss later. In the meantime, below is an animated GIF demonstrating how the image plane and DOF changes between the EFCS and non-EFCS modes @ 1/8000. I included a second GIF showing the same at 1/1000.

After you click on a link there will be a delay while your web browser loads the full GIF, then it'll start animating. Depending on your screen size the browser will likely downsample the GIF - click on it to see it at 1:1. These GIFs are 100% crops from the center of an A7s image. Lens is a Sony 55mm FE @ f1/.8

EFCS vs non-EFCS @ 1/8000

EFCS vs non-EFCS @ 1/1000

 

[abortabort]

Are these results any different on the normal A7? I've taken many pictures at 1/2000 and 1/4000 with EFCS enabled and I've never noticed any issues

I don't know, and I don't have a visible light a7 to test. Sorry.

You could do something as disposive by taking a picture of an evenly lit white wall at fast shutter speeds and reading the average values of raw channels in 200x200 squares at top and bottom. Compare with pictures made with the mechanical shutter and/or at slow shutter speeds.

That way you wouldn't need a 'scope to do the experiment.

A key element is how much tolerance you have in your images for non-uniform exposure. I imagine that darkened tops could improve some images.

Jim

 

[JimKasson]

+1.

Jim

 

[Petroglyph]

Yesterday I looked at shutter travel time for silent shutter operation on the Sony a7RII. Incidentally, we noted that the electronic first curtain shutter (EFCS) pretty faithfully emulated the acceleration of the mechanical second curtain, at least at 1/500 second.

On the Nikon D810, EFCS stops working at shutter speeds above 1/2000. On the Sony alpha 7 cameras with EFCS, Sony advises against using the capability at fast shutter speeds, but let's the user decide how far to push it.

So, let's push it and see what happens when we take pictures of a high-frequency triangle wave on an oscilloscope with the time base set to 500 microseconds (us) per division.

You can see that the EFCS image has a little less exposure at the top of the image, but you probably wouldn't notice that in a real photograph. The all mechanical image is pretty even.

Now the EFCS exposure is suffering from nonuniformity noticeably, but would probably be fine for many purposes.

The EFCS exposure at the top of the image is about 1/3 of that at the bottom. This should normally be avoided.

There is essentially zero exposure at the top of the EFCS image. Not useful.

The synching of electronic and mechanical behavior over time, temperture, wear, and sample variation is an iffy business.

I used to say that users should avoid a7x EFCS shutter speeds of over 1/2000 second. Based on these results, I'm dropping that to 1/1000.

 

[JimKasson]

Does crop mode get another stop of usability in your opinion? I think darkening the sky in many cases might make the shot look better so that is fortunate.

It should get more than that. I'll have to test it. No rest for the weary...

Jim

 

[JimKasson]

Well, if you'll treat me to round trip plane tickets from Upstate NY, you for sure can borrow my 90mm lens for the tests

Janet, that's a nice offer, but I've arranged access to the lens through other channels. Stay tuned.

It's concours week so it's super crowded. It's probably too cold for you here anyway.



Jim

--
http://blog.kasson.com

 

[Horshack]

More clues... Here are animated GIFs of the Mitakon 50mm f/0.95 lens at 1/8000, but the aperture varied between f/0.95 and f/5.6, and this time the full A7s image instead of just the center crop. Notice the difference in both image plane effects and exposure errors between the two apertures. FYI, the Mitakon is an FE-mount lens but with no electrical contacts, so body is unaware of aperture. Extra noise on the f/5.6 GIF due to lower exposure/ISO 25,600.

Mitakon @ f/0.95, EFCS vs non-EFCS 1/8000

Mitakon @ f/5.6, EFCS vs non-EFCS 1/8000

And here's an excerpt about the EFCS from the NEX-5N manual (didn't find this in the A7s or A7rII manual though I think it still applies):

" When you shoot at high shutter speeds with a large diameter lens attached, the ghosting of a blurred area may occur, depending on the subject or shooting conditions. In such cases, set this item to [Off]."

"When a Minolta/Konica Minolta lens is used, set this item to [Off]. If you set this item to [On], the correct exposure will not be set or the image brightness will be uneven.:

 

[steelhead3]

Are the Sony 70/200 2.8 and 300 2.8 (both KM designs) included in this warning?