How much Sony lenses are software corrected?

verybiglebowski · Sep 25, 2014

Tuloom said:
johnsmith404 said:

You're just messing around with the flange distance where even fractions of millimeters matter a lot. With current software technology, this level of blur is impossible to fix even when you know every last detail of the optical system you're dealing with.

Click to expand...

I thought this was what he was showing originally, too, but he is just referring to the color cast and vignetting, not the blur.

++++++++++

VBL, I just tried with the FE35.

With no power the lens falls out of focus past infinity. I did this:

Focused at MFD, decoupled the lens, lens falls out to what appears to be infinity focus.

Focused at infinity, decoupled the lens, lens falls out even further past infinity.

You can clearly see the FOV in the LCD change to a wider FOV you cannot replicate with power to the lens. Without power it is some resting distance past infinity. It's a position the lens was obviously not designed to shoot at, as you cannot get the same FOV with manual focus.

From what I can see, anything shot at the powerless position (past infinity) is invalid as far as a test for baked RAW with the FE35.

Tuloom thanks for trying.

I guess you now understand what makes it difficult to perform. For the FE 35/2.8 I posted here few shots (in a reply to older thread but I can't find it recently) which shows amount of vignetting and color cast correction in RAW.

In the same post I couldn't see traces of "baking" distortion control in RAW. However, my findings were related only to the in camera correction settings and lens correction external application. I tested few different RAW convertors that time, with RAW digger showing what RAW contains apart of the embeded data.

Roger Cicala some time ago tried FE 35/2.8 on the optical bench and results in terms of field curvature (sharpness in the edges) expressed in MTF were different (slightly worse) than with similar test performed with camera. He didn't mention in that article other corrections, and I don't know how he managed to focus lens. He also mentioned his doubt about test relevance because of the adapters for the optical bench if I remember well.

My latest thoughts how to improve relevance is to try to figure out at which position AF parks after decoupling. If I can find that position (assuming it's the same always), I might repeat test with focus exactly there.

Or, I would try to cover only some contacts.

Regards,

Viktor

verybiglebowski · Sep 25, 2014

jpr2 said:
verybiglebowski said:

You shouldn't look at absolute comparison from left to right, decoupled lens is simply OOF. More indicating is vignetting and slight increase in CA.

Comparing center with corners on decoupled lens image, indicates more pronounced field curvature, but I agree that it is hard to believe that any software can cope with that, so it is probably rather result of defocus. What thay might apply however, is something like unlinear unsharp mask, that will slightly improve perceived sharpness across the frame.

If anyone is able to make more controlled test, it would be great to see it. I will keep thinking how to improve my method...

Click to expand...

...you're exaggerating the amount of torsion the E-mount lens needs to be electrically decoupled - it is not 3-4 millimeters, IMO fraction of that is more than enough to get the lens decoupled;

so, from the purely mechanical stability/being-secure-in-their-mount your way is perfectly valid

as to a possibly better way to run your scenario - what happens with the AF (assuming all the time the OSS is securely OFF) when you switch the power off?

does this affect the AF'ing criticality?

if not, then perhaps you can have it AF'ed on tripod,

switch power OFF, decouple the lens ever so slightly,

turn power back on - and it should stay focused by decupled;

just an idea to try !!

btw. this is one of the reasons I do hate focus by wire (with EF lenses their mechanical friction coupling is enough to focus them even when there is no power [except of course the two f/1.2 ones]);

best, and congrats on well devised testing,

cheers, Quercy

stoppingdown · Sep 25, 2014

First, thanks for the test - it's excellent and definitely the best test about the SEL70200G I've read so far. Comparative tests are the most effective IMO. Keep them coming.

Back to the subject - I share the doubts that other casted about the amount of correction that Sony can do "before" the RAW. I think it's reasonable that they do something pixel-per-pixel (vignetting and most of CA correction), while for me it's hard to think of sharpening: not only that deconvolution would be a miracle one, but it would imply demosaicing, apply a matrix, and then remosaicing again, because the RAW file still has a Bayer layout. Then Lightroom would demosaic again... I can imagine separated deconvolution for each channel... but it would be a double miracle. Of course, it's only my opinion, I know basics of image processing, but I'm not an expert.

verybiglebowski · Sep 25, 2014

stoppingdown said:
First, thanks for the test - it's excellent and definitely the best test about the SEL70200G I've read so far. Comparative tests are the most effective IMO. Keep them coming.

Back to the subject - I share the doubts that other casted about the amount of correction that Sony can do "before" the RAW. I think it's reasonable that they do something pixel-per-pixel (vignetting and most of CA correction), while for me it's hard to think of sharpening: not only that deconvolution would be a miracle one, but it would imply demosaicing, apply a matrix, and then remosaicing again, because the RAW file still has a Bayer layout. Then Lightroom would demosaic again... I can imagine separated deconvolution for each channel... but it would be a double miracle. Of course, it's only my opinion, I know basics of image processing, but I'm not an expert.

stoppingdown · Sep 25, 2014

PS About OSS and power on... Placing the lens elements in the correct position might be not only a matter of applying power: we don't know whether there's a "default" correct position for the lens elements when power is on, the camera might need to set an explicit value. That's why I think any sharpness analysis with a disconnected camera is problematic.

Chris Malcolm · Sep 25, 2014

verybiglebowski said:
Tuloom said:

verybiglebowski said:

But you can see how corners looks without correction in comparison

Click to expand...

Unless Sony has some sort or alien tech, hellfire deconvolution going on, I have a hard time believing that. Something else is happening. If algorithms were that good, I think Sony would be proud to advertise it. I would.

Click to expand...

You shouldn't look at absolute comparison from left to right, decoupled lens is simply OOF. More indicating is vignetting and slight increase in CA.

Comparing center with corners on decoupled lens image, indicates more pronounced field curvature, but I agree that it is hard to believe that any software can cope with that, so it is probably rather result of defocus. What thay might apply however, is something like unlinear unsharp mask, that will slightly improve perceived sharpness across the frame.

Sony certainly corrects field curvatures in alpha models, and CA as well as vignetting. Lens corrections in-camera are a switchable option, plus if you set taken image review as a display option you see the before and after corrections in camera just after you press the shutter. Regardless of whether you have the in-camera jpeg lens corrections set or not, you can select whether or not to apply them to the RAW files in Sony's IDC.

Given how well and easily all that works in the alpha cameras, I see no reason why they wouldn't do the same in the NEX cameras.

verybiglebowski said:
If anyone is able to make more controlled test, it would be great to see it. I will keep thinking how to improve my method...

--
http:// www.photoworkshop.eu
http://verybiglobo.com/
http://www.flickr.com/photos/viktor_viktor/

--
Chris Malcolm

antoineb · Sep 25, 2014

Hi,

RAW should be un-corrected - or else it's not RAW.

JPEGs will normally be corrected.

Just take the same shot once in RAW and once in JPEG - or even both at once.

You'll have your answer.

verybiglebowski said:
Not a big deal, but during my FE 70-200 f/4 G OSS lens test, I tried to compare image with a lens properly mounted and slightly rotated just to loose connection.

On the left it's properly mounted lens and on the right - without electronic communication.

Lens starts to move focus immediately after being decoupled, but I managed to snap the shot pretty fast, losing minimum resolution in the center

But you can see how corners looks without correction in comparison

Whatever Sony did, it did it well, but I probably don't want to know

If you want to read my rolling review of the lens, here are the links:

Sony FE 70-200 f/4 G OSS Part 1

Sony FE 70-200 f/4 G OSS Part 2

--
http:// www.photoworkshop.eu
http://verybiglobo.com/
http://www.flickr.com/photos/viktor_viktor/

verybiglebowski · Sep 25, 2014

Chris Malcolm said:
verybiglebowski said:

Tuloom said:

verybiglebowski said:

But you can see how corners looks without correction in comparison

Click to expand...

Unless Sony has some sort or alien tech, hellfire deconvolution going on, I have a hard time believing that. Something else is happening. If algorithms were that good, I think Sony would be proud to advertise it. I would.

Click to expand...

You shouldn't look at absolute comparison from left to right, decoupled lens is simply OOF. More indicating is vignetting and slight increase in CA.

Comparing center with corners on decoupled lens image, indicates more pronounced field curvature, but I agree that it is hard to believe that any software can cope with that, so it is probably rather result of defocus. What thay might apply however, is something like unlinear unsharp mask, that will slightly improve perceived sharpness across the frame.

Click to expand...

Sony certainly corrects field curvatures in alpha models, and CA as well as vignetting. Lens corrections in-camera are a switchable option, plus if you set taken image review as a display option you see the before and after corrections in camera just after you press the shutter. Regardless of whether you have the in-camera jpeg lens corrections set or not, you can select whether or not to apply them to the RAW files in Sony's IDC.

Given how well and easily all that works in the alpha cameras, I see no reason why they wouldn't do the same in the NEX cameras.

It works the same, with one note.

Any RAW converter that you use will apply some of the preset corrections. Use RAW digger if you want to be sure that no additional corrections were applied.

Corrections that you can switch on or off in the IDC are embeded in the RAW metadada. They are set of instructions how to apply software corrections in the apropriate program.

What I was trying to find out, (I am repeating this all the time) are "hidden" corrections coded in the RAW during compression. They are part of an image processing pipe line and they could be hardware realted as well as software related. Don't mix in canera selectable corrections such as vignetting, CA and distortion, with the ones that are applied as a part of the lens design.

However, my results posted here are certainly compromised, because as another poster said, it is hard to get relevant data once the lens is decoupled.

Chris Malcolm said:
Chris Malcolm said:

If anyone is able to make more controlled test, it would be great to see it. I will keep thinking how to improve my method...

--
http:// www.photoworkshop.eu
http://verybiglobo.com/
http://www.flickr.com/photos/viktor_viktor/

Click to expand...

--
Chris Malcolm

--
http:// www.photoworkshop.eu

Verybiglobo photo blog

Everything about photography incl. some field tests of cameras and lenses

verybiglobo.com

viktor_viktor

Explore viktor_viktor’s 1,217 photos on Flickr!

www.flickr.com

verybiglebowski · Sep 25, 2014

antoineb said:
Hi,

RAW should be un-corrected - or else it's not RAW.

JPEGs will normally be corrected.

Just take the same shot once in RAW and once in JPEG - or even both at once.

You'll have your answer.

verybiglebowski said:

Not a big deal, but during my FE 70-200 f/4 G OSS lens test, I tried to compare image with a lens properly mounted and slightly rotated just to loose connection.

On the left it's properly mounted lens and on the right - without electronic communication.

Lens starts to move focus immediately after being decoupled, but I managed to snap the shot pretty fast, losing minimum resolution in the center

But you can see how corners looks without correction in comparison

Whatever Sony did, it did it well, but I probably don't want to know

If you want to read my rolling review of the lens, here are the links:

Sony FE 70-200 f/4 G OSS Part 1

Sony FE 70-200 f/4 G OSS Part 2

--
http:// www.photoworkshop.eu
http://verybiglobo.com/
http://www.flickr.com/photos/viktor_viktor/

Click to expand...

Thanks sir, but it's much more complicated than that....

--
http:// www.photoworkshop.eu

Verybiglobo photo blog

Everything about photography incl. some field tests of cameras and lenses

verybiglobo.com

viktor_viktor

Explore viktor_viktor’s 1,217 photos on Flickr!

www.flickr.com

verybiglebowski · Sep 25, 2014

stoppingdown said:
PS About OSS and power on... Placing the lens elements in the correct position might be not only a matter of applying power: we don't know whether there's a "default" correct position for the lens elements when power is on, the camera might need to set an explicit value. That's why I think any sharpness analysis with a disconnected camera is problematic.

--
Fabrizio Giudici
http://stoppingdown.net

I think that you have it right. Stabilization is messing with this, that's the only rational reason that I can see. It is interseting thiugh, that despite being off, it keeps "parking" or simply moving elements, once the connection is lost. Just to make it clear:

Camera is on the tripod, using lens tripod colar. All lens corrections are switched off. A mode, ISO 100. WB daylight.

1. OSS on the lens is switched off.

2. Mount lens on the camera.

3. Acquire focus (using AF or MF)

3a. If AF was used, once the focus was acquired, switch the AF to MF on the lens.

4a. Slightly rotate the camera to break connection.

4b. Switch the camera off before slightly rotating to break connection.

5a. AF will most probably move out off the focus

5b. It might happen that after switching canera on, lens will close to f/18 or f/22 or around. Focus however seems to stay, but you need to wait a while between switching camera off and on.

Whatever method I tried, as soon as camera lost connection, image in the viewfinder (LCD/EVF) "jumps a bit upward. That would indicate that OSS is making some activity and as I said, it is the only thing that I can see as the cause of such a difference in corner performance.

--
http:// www.photoworkshop.eu
http://verybiglobo.com/
http://www.flickr.com/photos/viktor_viktor/

verybiglebowski · Sep 25, 2014

verybiglebowski said:
stoppingdown said:

PS About OSS and power on... Placing the lens elements in the correct position might be not only a matter of applying power: we don't know whether there's a "default" correct position for the lens elements when power is on, the camera might need to set an explicit value. That's why I think any sharpness analysis with a disconnected camera is problematic.

Click to expand...

jpr2 · Sep 25, 2014

verybiglebowski said:
jpr2 said:

verybiglebowski said:

You shouldn't look at absolute comparison from left to right, decoupled lens is simply OOF. More indicating is vignetting and slight increase in CA.

Comparing center with corners on decoupled lens image, indicates more pronounced field curvature, but I agree that it is hard to believe that any software can cope with that, so it is probably rather result of defocus. What thay might apply however, is something like unlinear unsharp mask, that will slightly improve perceived sharpness across the frame.

If anyone is able to make more controlled test, it would be great to see it. I will keep thinking how to improve my method...

Click to expand...

...you're exaggerating the amount of torsion the E-mount lens needs to be electrically decoupled - it is not 3-4 millimeters, IMO fraction of that is more than enough to get the lens decoupled;

so, from the purely mechanical stability/being-secure-in-their-mount your way is perfectly valid

as to a possibly better way to run your scenario - what happens with the AF (assuming all the time the OSS is securely OFF) when you switch the power off?

does this affect the AF'ing criticality?

if not, then perhaps you can have it AF'ed on tripod,

switch power OFF, decouple the lens ever so slightly,

turn power back on - and it should stay focused by decupled;

just an idea to try !!

btw. this is one of the reasons I do hate focus by wire (with EF lenses their mechanical friction coupling is enough to focus them even when there is no power [except of course the two f/1.2 ones]);

best, and congrats on well devised testing,

cheers, Quercy

Click to expand...

Thanks Quercy,

You perfectly described what I actually did. Unfortunately focus kept moving away. My lens actually shown several anomalies during my effort, switching to f/22 e.g. on its own when decoupled but just sometimes.

Regards,

Viktor

Viktor,

is this whole bizarre behavior lens specific? Can you repeat/confirm by using any E-mount lens with a propensity to vignette BUT without OSS built-in? Because I think the OSS is the most likely culprit.

Also (another thought track, perhaps applicable in concert), there might be some large capacitor residing inside the lens itself, just an odd chance, but possible). Which means that between switching power OFF and the actual decoupling one would need quite a bit - say something between 30 sec. to several minutes. However, the use of any less WITHOUT built-in stabilization is much safer a bet IMO

,

cheers, Quercy

GaryW · Sep 25, 2014

jpr2 said:
jpr2 wrote:...

as to a possibly better way to run your scenario - what happens with the AF (assuming all the time the OSS is securely OFF) when you switch the power off?

The problem I see is that even with OSS turned off, there's a moving element in there that, when the power is off for the lens, could allow the element to move and throw the image out. All of this could be explained just by this, at least potentially. I'm sure the response will be that the operation was so precise that there's no way that such a floating element could budge. ;-) BTW, there was a recent thread on someone concerned about their 55-210 while looking at the rear of the lens, and the conclusion was that it was the OSS element.

If you want to test for corner vignetting correction, I think you could do this with a blank wall test and severe underexposure. Pull up the shadow information, and you should see a pattern (circular banding). There have been past threads on this pattern for other cameras/lenses, but the effect is too subtle to see in normal photos/exposures.

--
Gary W.

verybiglebowski · Sep 25, 2014

GaryW said:
GaryW said:

jpr2 wrote:...

as to a possibly better way to run your scenario - what happens with the AF (assuming all the time the OSS is securely OFF) when you switch the power off?

Click to expand...

The problem I see is that even with OSS turned off, there's a moving element in there that, when the power is off for the lens, could allow the element to move and throw the image out. All of this could be explained just by this, at least potentially. I'm sure the response will be that the operation was so precise that there's no way that such a floating element could budge. ;-) BTW, there was a recent thread on someone concerned about their 55-210 while looking at the rear of the lens, and the conclusion was that it was the OSS element.

If you want to test for corner vignetting correction, I think you could do this with a blank wall test and severe underexposure. Pull up the shadow information, and you should see a pattern (circular banding). There have been past threads on this pattern for other cameras/lenses, but the effect is too subtle to see in normal photos/exposures.

GaryW · Sep 25, 2014

EinsteinsGhost said:
EinsteinsGhost said:

verybiglebowski wrote:..

Click to expand...

It will take very little to create an image that will take a lot to fix, which is all that I see here. A lens is not secure, if it is not locked in place. In many cases, may not even focus at infinity.

FE 70-200 is not a lens built to serious size constraints. Those don't even apply to E 35. All that E 35 really needs is corrections against vignetting (clearly, Sony could have done that with auto fix but didn't).

EinsteinsGhost said:

EinsteinsGhost said:

You could also try it on an older E-mount body (they will be crop, but still should show issues around the edges), as they won't have the firmware to apply corrections. You would then look at JPEGs.

Click to expand...

Most noticable corrections are vignetting and color casting and they both relate to the corners. You might try it with native APS-C lenses though. BTW why should I look to JPEG?

Click to expand...

The image has a lot to fix than just vignetting and color cast, and across the frame. In other words, a lens not completely set on the mount is not the appropriate way to test anything. As I said, if you can't simply tape the contacts, then your best bet is to use a body that pre-dates the lens/has older firmware.

This makes the most sense -- find a Nex-3/5 which wouldn't have firmware to do corner or even vignetting correction. You can use RAW in that case -- shouldn't matter, as long as your RAW converter doesn't sneak in corrections automatically. Might be worth using some generic converter like Raw Therapee for this.

EinsteinsGhost said:
Why JPEG? So, you can eliminate any updated RAW converter as well, since you think Sony is writing all the corrections to RAW.

Using a 3 or 5 without corrections, and JPEG to eliminate RAW converters, at least we'd see a photo with actual view through the lens. If the corners were so bad, it should be visible in JPEG. Just grab your old camera!

EinsteinsGhost said:
Trust me, if this kind of corrections were easy, pretty much every lens will be very good, across the frame.

In the future, lenses will only need 1 or 2 elements, and software will fix everything up!

antoineb · Sep 25, 2014

verybiglebowski said:
antoineb said:

Hi,

RAW should be un-corrected - or else it's not RAW.

JPEGs will normally be corrected.

Just take the same shot once in RAW and once in JPEG - or even both at once.

You'll have your answer.

verybiglebowski said:

Not a big deal, but during my FE 70-200 f/4 G OSS lens test, I tried to compare image with a lens properly mounted and slightly rotated just to loose connection.

On the left it's properly mounted lens and on the right - without electronic communication.

Lens starts to move focus immediately after being decoupled, but I managed to snap the shot pretty fast, losing minimum resolution in the center

But you can see how corners looks without correction in comparison

Whatever Sony did, it did it well, but I probably don't want to know

If you want to read my rolling review of the lens, here are the links:

Sony FE 70-200 f/4 G OSS Part 1

Sony FE 70-200 f/4 G OSS Part 2

--
http:// www.photoworkshop.eu
http://verybiglobo.com/
http://www.flickr.com/photos/viktor_viktor/

Click to expand...

Click to expand...

Thanks sir, but it's much more complicated than that....

--
http:// www.photoworkshop.eu
http://verybiglobo.com/
http://www.flickr.com/photos/viktor_viktor/

Tuloom · Sep 25, 2014

verybiglebowski said:
... it certainly has to do someting with OSS. I will post shortly on my blog a similar wall test as you propose, with a focus on place (certainly within DOF) and huge difference in the corners. The only logical explanation is OSS moving some corrective element no matter if its on or off. (When the camera is on, OSS probably gets to the stand by mode, ready for the action if switched on, when the power is off it parks the compensation elements in a rest position. Because the lens was never meant to work without power, what we see is a result of the changed optical path, causing such significant diffenernce in a field curvature.

At keast we have I believe, one answer more...

What you describe there is exactly what I see with the FE35, which doesn't have OSS. When power is lost, the lens defaults to some "home" or "parked" state. For the FE35, that off state looks to be something past infinity.

I saw that when I set MF to infinity and watched a corner in the LCD. As soon as contact was lost, the lens "relaxes" and the FOV expands. It expands out further than what is possible with powered focus at infinity. That is, the FOV is one that you can't get with the lens powered. It's outside of the range between MFD and infinity. That's what make me leery of saying anything about RAW with no power to a lens, which I have no problem believing there are baked in corrections, it's just, as you state, very difficult to test for.

What's hard is not knowing exactly how the lenses work and what the elements are doing with the when powered. I don't have a clue.

verybiglebowski · Sep 25, 2014

Tuloom said:
verybiglebowski said:

... it certainly has to do someting with OSS. I will post shortly on my blog a similar wall test as you propose, with a focus on place (certainly within DOF) and huge difference in the corners. The only logical explanation is OSS moving some corrective element no matter if its on or off. (When the camera is on, OSS probably gets to the stand by mode, ready for the action if switched on, when the power is off it parks the compensation elements in a rest position. Because the lens was never meant to work without power, what we see is a result of the changed optical path, causing such significant diffenernce in a field curvature.

At keast we have I believe, one answer more...

Click to expand...

What you describe there is exactly what I see with the FE35, which doesn't have OSS. When power is lost, the lens defaults to some "home" or "parked" state. For the FE35, that off state looks to be something past infinity.

I saw that when I set MF to infinity and watched a corner in the LCD. As soon as contact was lost, the lens "relaxes" and the FOV expands. It expands out further than what is possible with powered focus at infinity. That is, the FOV is one that you can't get with the lens powered. It's outside of the range between MFD and infinity. That's what make me leery of saying anything about RAW with no power to a lens, which I have no problem believing there are baked in corrections, it's just, as you state, very difficult to test for.

What's hard is not knowing exactly how the lenses work and what the elements are doing with the when powered. I don't have a clue.

I had to move for few days fron the office, but when I get back, I will try with remaining FE lenses. The amount of corner smearing in my examples should certainly be result of optical performance and sudden change of optical path makes most sense. Is it because of OSS or focus by wire, or combination of both I don't know though. FE 35 due to its size is probably corrected more and what you suggest makes sense.

Thanks for trying.

jpr2 · Sep 25, 2014

verybiglebowski said:
Tuloom said:

verybiglebowski said:

... it certainly has to do someting with OSS. I will post shortly on my blog a similar wall test as you propose, with a focus on place (certainly within DOF) and huge difference in the corners. The only logical explanation is OSS moving some corrective element no matter if its on or off. (When the camera is on, OSS probably gets to the stand by mode, ready for the action if switched on, when the power is off it parks the compensation elements in a rest position. Because the lens was never meant to work without power, what we see is a result of the changed optical path, causing such significant diffenernce in a field curvature.

At keast we have I believe, one answer more...

Click to expand...

What you describe there is exactly what I see with the FE35, which doesn't have OSS. When power is lost, the lens defaults to some "home" or "parked" state. For the FE35, that off state looks to be something past infinity.

I saw that when I set MF to infinity and watched a corner in the LCD. As soon as contact was lost, the lens "relaxes" and the FOV expands. It expands out further than what is possible with powered focus at infinity. That is, the FOV is one that you can't get with the lens powered. It's outside of the range between MFD and infinity. That's what make me leery of saying anything about RAW with no power to a lens, which I have no problem believing there are baked in corrections, it's just, as you state, very difficult to test for.

What's hard is not knowing exactly how the lenses work and what the elements are doing with the when powered. I don't have a clue.

Click to expand...

I had to move for few days fron the office, but when I get back, I will try with remaining FE lenses. The amount of corner smearing in my examples should certainly be result of optical performance and sudden change of optical path makes most sense. Is it because of OSS or focus by wire, or combination of both I don't know though. FE 35 due to its size is probably corrected more and what you suggest makes sense.

Thanks for trying.

...a pretty strange dance if you ask me. Esp. the results Tuloom describes with the FE35 "parking" way beyond infinity are really mind boggling (btw. it is possible to circumvent any camera's tomfoolery when powered OFF just by pressing battery release lever - an instant cut of power, should prevent any "behind the scenes" lens movements).

Unfortunately, a week ago I have parted with my last remaining E-mount lens so I can't test them here, but... as I'm still having many Canon's with IS, I've just tried some of them mounted through the Speed Booster. So, to summarize:

both with IS switched ON and OFF, and the EF 100/2.8L IS lens,
AF'ing on a target,
and then switched to MF (on the lens - easy to do with EF glass);
switched power OFF, and waited few seconds before decoupling the contacts and powering the camera ON again;
and... lo and behold = focus was back as it was before the lens decoupling, exactly as expected.

What we need to know also is that the EF 100/2.8L IS, while being equipped with an excellent stabilization, is rather notoriously known for a "rattling" internal IS element (when the lens is powered OFF and moved around something moves inside pretty audibly). Yet no surprises from this department

!!

Quercy

Paul1974 · Sep 27, 2014

All corrections are optional. If they aren't, what's the big deal trying to disable them anyway?

You should bite the bullet and accept correction because it simply improves image quality.

How much Sony lenses are software corrected?

Senior Member

Senior Member

Forum Enthusiast

Senior Member

Forum Enthusiast

Senior Member

Veteran Member

Senior Member

Senior Member

Senior Member

Senior Member

Forum Pro

Forum Pro

Senior Member

Forum Pro

Veteran Member

Senior Member

Senior Member

Forum Pro

Member

Keyboard shortcuts