Janet, I got the 90 FE lens this morning, and just tested it with the scope. As you're reporting, it is much improved over the Zeiss 100/2 at high EFCS shutter speeds. I posted 'scope photos here .
What's going on?
I sure don't know.
Jim
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
Fastest EFCS shutter speed usable on the a7RII?
- Thread starter JimKasson
- Start date
-
- Tags
- nikon d810 sony a7 sony a7r
What's going on?
I sure don't know.
Jim
No. I was using f/4 with the Zeiss 100 and f/2.8 with the Sony 90. I can go back and reshoot at f/2.8 with the Zeiss if you think it might make a difference.
If the aperture were changing during the exposure, we'd see brightening or darkening of the trace, wouldn't we? With the 'scope, we can separate time and intensity.
Maybe more lit-phosphor-density would help. I can turn up the frequency some more on the Wavetek, and turn up the beam current on the 'scope, and rerun the tests at the same aperture.
Jim
l_d_allan
Veteran Member
I did look some more, and found this link on the Sony web-site:
http://helpguide.sony.net/ilc/1520/v1/en/print.pdf
Last edited:
jalywol
Forum Pro
Tried both the Sony 70-300mmG, and a Minolta AF 100mm f2.8 macro.
On both of these lenses, there is no obvious degradation using EFCS below 1/2000, and even at and above 1/2000, what shading does occur is really minor. So, in spite of Sony putting a warning in about Minolta lenses, the Minolta AF and Sony A mount lenses really don't seem to suffer much from the shading effect.
I think I will run these again tomorrow, though, as the light was fading a bit as I was doing the test, just to make sure, but it looks like the only lenses with significant issues are the fully manual ones with no electronic connection. I would be most interested in how Canon lenses used with the full electronic coupling adapters behave, too, but I don't have any, so I can't test those out.
-J
On both of these lenses, there is no obvious degradation using EFCS below 1/2000, and even at and above 1/2000, what shading does occur is really minor. So, in spite of Sony putting a warning in about Minolta lenses, the Minolta AF and Sony A mount lenses really don't seem to suffer much from the shading effect.
I think I will run these again tomorrow, though, as the light was fading a bit as I was doing the test, just to make sure, but it looks like the only lenses with significant issues are the fully manual ones with no electronic connection. I would be most interested in how Canon lenses used with the full electronic coupling adapters behave, too, but I don't have any, so I can't test those out.
-J
l_d_allan
Veteran Member
Agree. I've only had my a7Rii since late Monday, and have only used the native Sony FE 28mm f2 so far (only native lens I have ... reasonably priced fast prime).
I haven't yet opened the FotodioX adapter to see how my Canon EF lenses work.
Yeah, I saw a difference with my A7s - smaller apertures being worse regarding evenness of exposure.
I cranked up the beam current and the function generator frequency. I set both lenses to f/2.8. The Zeiss Makro is not an internal focusing lens, and the Sony 90 FE is, so the Sony probably ended up at 80mm focal length or less after I filled the frame with the 'scope. I had to back up a lot when I put the Zeiss on.
I didn't get a good image at 1/1000 with the Sony, but we know that's good.
1/2000, Zeiss
1/2000, Sony
1/4000, Zeiss
1/4000, Sony
1/8000, Zeiss
1/8000, Sony
Jim
--
http://blog.kasson.com
I didn't get a good image at 1/1000 with the Sony, but we know that's good.
1/2000, Zeiss
1/2000, Sony
1/4000, Zeiss
1/4000, Sony
1/8000, Zeiss
1/8000, Sony
Jim
--
http://blog.kasson.com
Attachments
Last edited:
Based on Sony's warnings about EFCS and non-native lenses we've been speculating in this thread about what the camera might be doing differently behind the scenes for native lenses. To help discover that it would be useful to know how exactly an EFCS exposure changes when a native lens is mounted vs that same lens mounted when the camera is not aware of the lens model attached. There's no way to accomplish that for E-Mount lenses but it can be done for Canon-adapted lens mounted on a Metabones adapter. These lenses appear to the camera as native E-Mount lenses; the only difference is that Sony's firmware wont have any database knowledge for Canon lenses, which matters in case EFCS behavior is based not just on the camera's awareness of the dialed-in aperture but also on the lens model attached. But an experiment with a Canon lens is still useful because it'll at least let us determine if aperture awareness does in fact affect an EFCS exposure.
The method we can use to compare a Canon-mounted lens with aperture awareness vs non-awareness is by first mounted the lens normally on the Metabones and taking our EFCS ON vs OFF exposures. Then, to test non-awareness I twist the Metabones adapter just enough for it to lose electrical contact with the body - a nice aspect of Canon EF lenses is that their aperture stays stuck to the last electrically-commanded aperture, so as long as I twist the Metabones adapter while there's still power to the camera the lens will remain at the last aperture used, which then allows us to do our EFCS ON vs OFF captures for the non-aperture awareness case. Note that after twisting the Metabones to lose the connection I then cycle power on the camera before taking the non-aperture aware exposures, to make sure there's no memory of the lens or aperture those for non-aperture aware exposures.
There are two sets of Animated GIFs below, one at f/1.2 and one at f/4, all with the A7s at 1/8000. The first image in each set demonstrates EFCS ON vs EFCS OFF with the body being aware of the aperture. The second image in each set then compares EFCS ON with aperture awareness vs non-awareness. The first image is to establish the baseline of what EFCS ON vs OFF looks like with aperture awareness; the second image then lets us see what EFCS ON looks like with and without awareness. To make sure that the non-awareness case of twisting the Metabones doesn't alter the image I also took samples with non-awareness EFCS OFF and made sure they matched the EFCS OFF awareness exposures - they did so I wont show those control samples.
f/1.2:
Canon 50L @ f/1.2, Aperture Aware, EFCS ON vs OFF
Canon 50L @ f1/.2, EFCS ON, Aperture Aware vs Not-Aware
f/4:
Canon 50L @ f/4, Aperture Aware, EFCS ON vs OFF
Canon 50L @ f/4, EFCS ON, Aperture Aware vs Not-Aware
The method we can use to compare a Canon-mounted lens with aperture awareness vs non-awareness is by first mounted the lens normally on the Metabones and taking our EFCS ON vs OFF exposures. Then, to test non-awareness I twist the Metabones adapter just enough for it to lose electrical contact with the body - a nice aspect of Canon EF lenses is that their aperture stays stuck to the last electrically-commanded aperture, so as long as I twist the Metabones adapter while there's still power to the camera the lens will remain at the last aperture used, which then allows us to do our EFCS ON vs OFF captures for the non-aperture awareness case. Note that after twisting the Metabones to lose the connection I then cycle power on the camera before taking the non-aperture aware exposures, to make sure there's no memory of the lens or aperture those for non-aperture aware exposures.
There are two sets of Animated GIFs below, one at f/1.2 and one at f/4, all with the A7s at 1/8000. The first image in each set demonstrates EFCS ON vs EFCS OFF with the body being aware of the aperture. The second image in each set then compares EFCS ON with aperture awareness vs non-awareness. The first image is to establish the baseline of what EFCS ON vs OFF looks like with aperture awareness; the second image then lets us see what EFCS ON looks like with and without awareness. To make sure that the non-awareness case of twisting the Metabones doesn't alter the image I also took samples with non-awareness EFCS OFF and made sure they matched the EFCS OFF awareness exposures - they did so I wont show those control samples.
f/1.2:
Canon 50L @ f/1.2, Aperture Aware, EFCS ON vs OFF
Canon 50L @ f1/.2, EFCS ON, Aperture Aware vs Not-Aware
f/4:
Canon 50L @ f/4, Aperture Aware, EFCS ON vs OFF
Canon 50L @ f/4, EFCS ON, Aperture Aware vs Not-Aware
Last edited:
**** f/1.2 ****
Aperture-Aware EFCS ON vs OFF
Top: EFCS has a modestly dark band vs non-EFCS
Bottom: Looks about the same
EFCS ON, Aperture-Aware vs Not-Aware
Top: Band is darker for aperture-aware vs not-aware
Bottom: Opposite of top - not-aware has darker band than aware or EFCS OFF
**** f/4 ****
Aperture-Aware EFCS ON vs OFF
Top: EFCS has a much darker band
Bottom: EFCS has modestly brighter band vs non-EFCS
EFCS ON, Aperture-Aware vs Not-Aware
Top: Band is significantly darker for aperture-aware vs not-aware
Bottom: Aware has brighter band, non-aware has normal band (matches EFCS OFF)
Initial Conclusions
For aperture aware EFCS exposures, the top of the image is darker than normal/EFCS OFF. For non-aperture aware EFCS exposures, the bottom is normal for f/1.2 but at f/4 is darker than normal/EFCS OFF (and darker than aperture aware EFCS).
Focusing on just f/4 where the differences are more noticeable, the area of EFCS exposure errors reverse when the camera is aperture aware vs no aware. First, this tells us that the camera is in fact doing something different for EFCS exposures when it knows what the aperture of the lens is during the exposure. Second, it tells us that whatever adjustment the camera is making it helps one end of the frame to detriment of the other...at least for Canon-adapted lenses.
My best guess is that the camera is altering the initial timing of the EFCS reset based on the dialed-in aperture. This alteration causes the start of the frame to be too dark or too light, then causes the end of the frame to be the opposite, too light or too dark. For example, for the case where the top of the frame is too dark, the camera is resetting the rows too late, not allowing enough time/exposure between the interval of the reset and the arrival of the second curtain; when the exposure reaches the opposite end of the frame, the camera is resetting the rows too early, allowing too much time/exposure between the interval of the reset and the arrival of the second curtain. In other words, it's making an initial, coarse correction to the timing that is benefits one edge of the frame but hurts the other, due to perhaps edge differences of the mechanical shutter that affect one end of the frame differently than the other (as was discussed earlier in the thread by Antisthenes).
Aperture-Aware EFCS ON vs OFF
Top: EFCS has a modestly dark band vs non-EFCS
Bottom: Looks about the same
EFCS ON, Aperture-Aware vs Not-Aware
Top: Band is darker for aperture-aware vs not-aware
Bottom: Opposite of top - not-aware has darker band than aware or EFCS OFF
**** f/4 ****
Aperture-Aware EFCS ON vs OFF
Top: EFCS has a much darker band
Bottom: EFCS has modestly brighter band vs non-EFCS
EFCS ON, Aperture-Aware vs Not-Aware
Top: Band is significantly darker for aperture-aware vs not-aware
Bottom: Aware has brighter band, non-aware has normal band (matches EFCS OFF)
Initial Conclusions
For aperture aware EFCS exposures, the top of the image is darker than normal/EFCS OFF. For non-aperture aware EFCS exposures, the bottom is normal for f/1.2 but at f/4 is darker than normal/EFCS OFF (and darker than aperture aware EFCS).
Focusing on just f/4 where the differences are more noticeable, the area of EFCS exposure errors reverse when the camera is aperture aware vs no aware. First, this tells us that the camera is in fact doing something different for EFCS exposures when it knows what the aperture of the lens is during the exposure. Second, it tells us that whatever adjustment the camera is making it helps one end of the frame to detriment of the other...at least for Canon-adapted lenses.
My best guess is that the camera is altering the initial timing of the EFCS reset based on the dialed-in aperture. This alteration causes the start of the frame to be too dark or too light, then causes the end of the frame to be the opposite, too light or too dark. For example, for the case where the top of the frame is too dark, the camera is resetting the rows too late, not allowing enough time/exposure between the interval of the reset and the arrival of the second curtain; when the exposure reaches the opposite end of the frame, the camera is resetting the rows too early, allowing too much time/exposure between the interval of the reset and the arrival of the second curtain. In other words, it's making an initial, coarse correction to the timing that is benefits one edge of the frame but hurts the other, due to perhaps edge differences of the mechanical shutter that affect one end of the frame differently than the other (as was discussed earlier in the thread by Antisthenes).
Last edited:
steelhead3
Veteran Member
I don't if this has anything to do with anything but lens rental can't test Sony lenses because of their behavior of resetting themselves when powered off.
Canon-adapted lenses on Sony bodies do the same. Nikon lenses as well, as a function of their aperture constantly being wedged by a mechanical lever that disconnects when you dismount the lens. Not sure about Nikon's newer electronically-controlled E lenses.
Last edited:
l_d_allan
Veteran Member
Would you consider a Canon EF 70-200mm f2.8L II to be "legacy"? Is "legacy" a synonym for "anything other than a Sony E/FE lens", including the Sony Alpha lenses?
This "Canon defector" thinks of "legacy" as manual focus, non-electronic Leica's, Zuiko's, Canon FD's, etc. from the film era.
The range of usable shutter speeds seems to be shrinking:
- initially ... 1/2000 sec should be ok with EFCS
- hmmmm ... maybe 1/1000 or slower should be ok
- hmmmm ... maybe 1/400 or slower should be ok
steelhead3
Veteran Member
It works fine on my A mounts lenses and I never worry about shutter speed and leave it on all the time. Much of this discussion is theoretical...EFC on or off is the least of my parameter adjustments that I have a concern.
Petroglyph
Veteran Member
- Messages
- 6,096
- Solutions
- 9
- Reaction score
- 1,384
My mother used to say to me: "My mind is made up; don't confuse me with facts."
Jim
I've thought about it, and I disagree with your contention that much of the discussion is theoretical. In the scientific yin/yang of experimental versus theoretical results, I think this discussion has been skewed towards the experimental. I'm not complaining about that. While I appreciate the theory that has been introduced -- I've certainly learned some things -- we've got at least three people designing and conducting experiments aimed at understanding the effects and figuring out what Sony is doing with the native lenses. We don't have this thing figured out yet, but we're closer than we were a couple of days ago.
Jim
I performed a new experiment tonight which I believe gets us very close to solving the EFCS riddle, if not solving it completely. For the past day I've been trying to devise a method to capture the speed differences between the EFCS and MSCS (Mechanical Second Curtain Shutter), since sync/speed mismatches seemed like the most likely cause of the exposure errors at various shutter speeds. It finally occurred to me to use a flash strobe at various shutter speeds above x-sync - that way we could use the outline of the MSCS to measure the timing difference of the MSCS between EFCS exposures and non-EFCS exposures. The problem with this method is that we would only be able to measure shutter speeds between 1/250 and 1/800, the range of speeds at which the MSCS would be visible at various stages during the strobe exposure. But still I hoped this would provide some useful information even if it excluded the more critical shutter speeds like 1/8000.
Experimentation sometimes yields accidental discoveries and the big one here is that different lens apertures shape the field of light in a convex projection during an exposure. The larger the aperture the deeper the convex projection.
Before further discussion, here are results from the experiment to demonstrate what's happening:
A7s Strobe Shutter Blade Expierment - 55mm FE and Mitakon 50mm
This convex projection likely complicates the implementation of an EFCS significantly because it implies that to properly sync the EFCS reset to the MSCS requires more than just simple row-based pixel resets. It would instead require a shaped reset across a significant number of pixel rows, to match the shading interaction of the lens and MSCS. And the faster the shutter speed, the more precise the EFCS reset shaping has to be to match the aperture, because a faster shutter speed implies a greater density of light and so the error skew of the EFCS increases as a percentage of total exposure time.
This is likely where the Sony warnings come from. To recount, here are the warnings from the NEX-5N manual for the EFCS:
"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"
I think we now know where the large diameter warning comes from - the larger the aperture, the more convex the shading and thus more difficult it will be for the EFCS timing to sync to that shading. I think this is also where the native-lens warning comes in (which is spelled out more clearly in the A7rII documentation) - without knowledge of the aperture from a non-native lens, the EFCS logic is flying blind with respect to what the convex shading will be. And this is where it's gets interesting - the camera would have to then guess what the shading will be for non-native lenses, and this guess will favor one range of aperture sizes to the detriment of another. For example, as depicted in my experiment the other day Mitakon f/0.95 vs f/5.6 (link).
For native lenses the camera will not only have the aperture communicated to it but will also know the ID/model of the lens and perhaps contain a database with more detailed shaping information for the various apertures the lens supports. This will depend on us figuring out exactly what causes the shading from aperture. Is it just a projection of light/shading from the aperture balde? The oblique angle of light? Or does its shape affect the shading too?
I haven't even discussed the EFCS speed observations from this test, which was the original point of the experiment. It's not clear those are even relevant at this point.
Experimentation sometimes yields accidental discoveries and the big one here is that different lens apertures shape the field of light in a convex projection during an exposure. The larger the aperture the deeper the convex projection.
Before further discussion, here are results from the experiment to demonstrate what's happening:
A7s Strobe Shutter Blade Expierment - 55mm FE and Mitakon 50mm
This convex projection likely complicates the implementation of an EFCS significantly because it implies that to properly sync the EFCS reset to the MSCS requires more than just simple row-based pixel resets. It would instead require a shaped reset across a significant number of pixel rows, to match the shading interaction of the lens and MSCS. And the faster the shutter speed, the more precise the EFCS reset shaping has to be to match the aperture, because a faster shutter speed implies a greater density of light and so the error skew of the EFCS increases as a percentage of total exposure time.
This is likely where the Sony warnings come from. To recount, here are the warnings from the NEX-5N manual for the EFCS:
"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"
I think we now know where the large diameter warning comes from - the larger the aperture, the more convex the shading and thus more difficult it will be for the EFCS timing to sync to that shading. I think this is also where the native-lens warning comes in (which is spelled out more clearly in the A7rII documentation) - without knowledge of the aperture from a non-native lens, the EFCS logic is flying blind with respect to what the convex shading will be. And this is where it's gets interesting - the camera would have to then guess what the shading will be for non-native lenses, and this guess will favor one range of aperture sizes to the detriment of another. For example, as depicted in my experiment the other day Mitakon f/0.95 vs f/5.6 (link).
For native lenses the camera will not only have the aperture communicated to it but will also know the ID/model of the lens and perhaps contain a database with more detailed shaping information for the various apertures the lens supports. This will depend on us figuring out exactly what causes the shading from aperture. Is it just a projection of light/shading from the aperture balde? The oblique angle of light? Or does its shape affect the shading too?
I haven't even discussed the EFCS speed observations from this test, which was the original point of the experiment. It's not clear those are even relevant at this point.
Last edited:
l_d_allan
Veteran Member
Thanks! to you (and Jim K and others) for tackling this.
If this is confirmed as the cause (or a cause) of the issue, I'm unclear on what the implications are for non-native lenses:
- Certain apertures are or aren't ok?
- Certain shutter speeds are or aren't ok?
- Certain combinations of shutter speeds and apertures are or aren't ok?
- Has the issue of whether the issue depends on the adapter been addressed? Commlite vs FotodioX vs various generations of Metabones vs others?
- Legacy (manual non-electronic from film era) vs Canon EF with electronics?
- Is there a way for a person with flash to check their particular combination
of camera model + adapter + lens?
[*](hope I'm not being an askhole )
Similar threads
Keyboard shortcuts
- f
- Forum