On Sharpness, ISO and Shutter Speed

Started Nov 29, 2013 | Discussions
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: On Sharpness, ISO and Shutter Speed
In reply to Steen Bay, Dec 1, 2013

Steen Bay wrote:

Anders W wrote:

Ken Strain wrote:

These data are puzzling, two features disturb me most:

The spread between H and V gets to about 20% in several of the plots, including the points for the EM5 at 1/4000s. I am reluctant to accept that particular case as being due to shutter shock. If those data are bad, how can we tell which are good?

The EM1 plot is the only one that looks somewhat like I would expect if shutter shock were the main cause of the spread, but even here it appears that something else breaks the H-V symmetry with the opposite sign (such that there is roughly 6% discrepancy between the averaged H and V scopes of the 3 fastest speeds).

Hi Ken,

To me, the graphs for the E-M5, the A7R, and the D610 are pretty much in line with what would be expected in the presence of shutter shock, not just that of the E-M1. Of course, there may be other reasons for a difference between vertical and horizontal resolution too, as strongly suggested by the graphs for the D610 and A7.

A bit strange that the blur is mostly vertical with E-M1 and A7R, and mostly horizontal with D610.

You are right about that. I overlooked that it was the horizontal rather than vertical resolution that dipped in the critical range in this case. And I guess the D610 shutter (which is what distinguishes it from the D600 and its problems with dirt on the sensor) is moving vertically just as the other ones.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Steen Bay
Veteran MemberPosts: 6,424
Like?
Re: On Sharpness, ISO and Shutter Speed
In reply to Anders W, Dec 1, 2013

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Ken Strain wrote:

These data are puzzling, two features disturb me most:

The spread between H and V gets to about 20% in several of the plots, including the points for the EM5 at 1/4000s. I am reluctant to accept that particular case as being due to shutter shock. If those data are bad, how can we tell which are good?

The EM1 plot is the only one that looks somewhat like I would expect if shutter shock were the main cause of the spread, but even here it appears that something else breaks the H-V symmetry with the opposite sign (such that there is roughly 6% discrepancy between the averaged H and V scopes of the 3 fastest speeds).

Hi Ken,

To me, the graphs for the E-M5, the A7R, and the D610 are pretty much in line with what would be expected in the presence of shutter shock, not just that of the E-M1. Of course, there may be other reasons for a difference between vertical and horizontal resolution too, as strongly suggested by the graphs for the D610 and A7.

A bit strange that the blur is mostly vertical with E-M1 and A7R, and mostly horizontal with D610.

You are right about that. I overlooked that it was the horizontal rather than vertical resolution that dipped in the critical range in this case. And I guess the D610 shutter (which is what distinguishes it from the D600 and its problems with dirt on the sensor) is moving vertically just as the other ones.

Seems that a good, heavy tripod isn't enough to prevent blur at certain shutter speeds, so could be interesting to know which shutter speeds are used by Photozone, Lenstip, DxO, SLRgear, etc. when testing lenses. Can we trust the results from cameras known to be affected?

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: On Sharpness, ISO and Shutter Speed
In reply to Steen Bay, Dec 1, 2013

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Ken Strain wrote:

These data are puzzling, two features disturb me most:

The spread between H and V gets to about 20% in several of the plots, including the points for the EM5 at 1/4000s. I am reluctant to accept that particular case as being due to shutter shock. If those data are bad, how can we tell which are good?

The EM1 plot is the only one that looks somewhat like I would expect if shutter shock were the main cause of the spread, but even here it appears that something else breaks the H-V symmetry with the opposite sign (such that there is roughly 6% discrepancy between the averaged H and V scopes of the 3 fastest speeds).

Hi Ken,

To me, the graphs for the E-M5, the A7R, and the D610 are pretty much in line with what would be expected in the presence of shutter shock, not just that of the E-M1. Of course, there may be other reasons for a difference between vertical and horizontal resolution too, as strongly suggested by the graphs for the D610 and A7.

A bit strange that the blur is mostly vertical with E-M1 and A7R, and mostly horizontal with D610.

You are right about that. I overlooked that it was the horizontal rather than vertical resolution that dipped in the critical range in this case. And I guess the D610 shutter (which is what distinguishes it from the D600 and its problems with dirt on the sensor) is moving vertically just as the other ones.

Seems that a good, heavy tripod isn't enough to prevent blur at certain shutter speeds, so could be interesting to know which shutter speeds are used by Photozone, Lenstip, DxO, SLRgear, etc. when testing lenses. Can we trust the results from cameras known to be affected?

Yes. The same thought has occurred to me. What I hope is that they prefer not to have more light in the studio where they do this than they really need, and that they therefore shoot at shutter speeds where shutter shock ceases to be much of a problem (although its effect may not be completely gone), like at 1/20 or below.

At least SLR Gear are aware of the problem as you can see from their recent review of the new Pany 14-140 here:

http://slrgear.com/reviews/showproduct.php?product=1611

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Steen Bay
Veteran MemberPosts: 6,424
Like?
Re: On Sharpness, ISO and Shutter Speed
In reply to Anders W, Dec 1, 2013

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Ken Strain wrote:

These data are puzzling, two features disturb me most:

The spread between H and V gets to about 20% in several of the plots, including the points for the EM5 at 1/4000s. I am reluctant to accept that particular case as being due to shutter shock. If those data are bad, how can we tell which are good?

The EM1 plot is the only one that looks somewhat like I would expect if shutter shock were the main cause of the spread, but even here it appears that something else breaks the H-V symmetry with the opposite sign (such that there is roughly 6% discrepancy between the averaged H and V scopes of the 3 fastest speeds).

Hi Ken,

To me, the graphs for the E-M5, the A7R, and the D610 are pretty much in line with what would be expected in the presence of shutter shock, not just that of the E-M1. Of course, there may be other reasons for a difference between vertical and horizontal resolution too, as strongly suggested by the graphs for the D610 and A7.

A bit strange that the blur is mostly vertical with E-M1 and A7R, and mostly horizontal with D610.

You are right about that. I overlooked that it was the horizontal rather than vertical resolution that dipped in the critical range in this case. And I guess the D610 shutter (which is what distinguishes it from the D600 and its problems with dirt on the sensor) is moving vertically just as the other ones.

Seems that a good, heavy tripod isn't enough to prevent blur at certain shutter speeds, so could be interesting to know which shutter speeds are used by Photozone, Lenstip, DxO, SLRgear, etc. when testing lenses. Can we trust the results from cameras known to be affected?

Yes. The same thought has occurred to me. What I hope is that they prefer not to have more light in the studio where they do this than they really need, and that they therefore shoot at shutter speeds where shutter shock ceases to be much of a problem (although its effect may not be completely gone), like at 1/20 or below.

Guess it's most likely that the lighting is constant and that the shutter speed changes with the f-stop, like it changes with the ISO in DPR's studio scene.

At least SLR Gear are aware of the problem as you can see from their recent review of the new Pany 14-140 here:

http://slrgear.com/reviews/showproduct.php?product=1611

Yes, and that's with a 250 pounds tripod!

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: On Sharpness, ISO and Shutter Speed
In reply to Steen Bay, Dec 1, 2013

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Ken Strain wrote:

These data are puzzling, two features disturb me most:

The spread between H and V gets to about 20% in several of the plots, including the points for the EM5 at 1/4000s. I am reluctant to accept that particular case as being due to shutter shock. If those data are bad, how can we tell which are good?

The EM1 plot is the only one that looks somewhat like I would expect if shutter shock were the main cause of the spread, but even here it appears that something else breaks the H-V symmetry with the opposite sign (such that there is roughly 6% discrepancy between the averaged H and V scopes of the 3 fastest speeds).

Hi Ken,

To me, the graphs for the E-M5, the A7R, and the D610 are pretty much in line with what would be expected in the presence of shutter shock, not just that of the E-M1. Of course, there may be other reasons for a difference between vertical and horizontal resolution too, as strongly suggested by the graphs for the D610 and A7.

A bit strange that the blur is mostly vertical with E-M1 and A7R, and mostly horizontal with D610.

You are right about that. I overlooked that it was the horizontal rather than vertical resolution that dipped in the critical range in this case. And I guess the D610 shutter (which is what distinguishes it from the D600 and its problems with dirt on the sensor) is moving vertically just as the other ones.

Seems that a good, heavy tripod isn't enough to prevent blur at certain shutter speeds, so could be interesting to know which shutter speeds are used by Photozone, Lenstip, DxO, SLRgear, etc. when testing lenses. Can we trust the results from cameras known to be affected?

Yes. The same thought has occurred to me. What I hope is that they prefer not to have more light in the studio where they do this than they really need, and that they therefore shoot at shutter speeds where shutter shock ceases to be much of a problem (although its effect may not be completely gone), like at 1/20 or below.

Guess it's most likely that the lighting is constant and that the shutter speed changes with the f-stop, like it changes with the ISO in DPR's studio scene.

That would be my guess as well.

At least SLR Gear are aware of the problem as you can see from their recent review of the new Pany 14-140 here:

http://slrgear.com/reviews/showproduct.php?product=1611

Yes, and that's with a 250 pounds tripod!

Yep. Although I suspect that there is something special about this lens and other Power OIS lenses with, as it seems, similar problems (X 14-42, X 45-175).

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
RussellInCincinnati
Senior MemberPosts: 3,197
Like?
does your 100-300 work better at slower than 1/25th?
In reply to Anders W, Dec 1, 2013

Anders W wrote:It's a challenge to get tack-sharp results but it is certainly possible. Here's one available at full resolution (it is a crop) for your pixel-peeping pleasure:

Don't see how this could have been done better.

A tripod sometimes help but is no panacea. I have shutter-shock problems with my E-M5 and 100-300/4-5.6 at 300 mm in the critical range of shutter speeds (from about 1/25 to about 1/250) even in tripod-based shooting. In order to get sharp results with this combo, I simply have to move out of that range.

Are you saying that at 300mm, the results are sharper at 1/15th of a second than at 1/50th?

Reply   Reply with quote   Complain
ProfHankD
Senior MemberPosts: 2,160Gear list
Like?
Another possible experiment
In reply to Jack Hogan, Dec 1, 2013

Jack Hogan wrote:

2) Are the dips we see in MTF50 around 1/10th of a second to about 1/300th possibly due to shutter shock - or something else? What implications does this have for people choosing a high resolution camera today?

I'd guess that vibration is a factor here, and imprecision in electronic 1st curtain reset may be the cause of the other issue, but here's a possible experiment to try to resolve this....

In either a focal plane shutter or a not-really-simultaneous-global-electronic shutter, the sensor exposure interval is not really the same for all pixels. Some are exposed earlier and others later, with the overlap fraction dependent on shutter speed. Thus, it should be possible to profile the vibration/movement over time by using multiple targets positioned so that they are exposed in temporal sequence as the shutter travels across the frame. Shutter shock should be an impulse with periodic decay of the vibration over time, which I would expect to cause a measurable pattern in the MTF50 measurements along the shutter's travel direction.... The same test method could also be used to test external damping methods; for example, I'd actually expect human hands to do better damping than a tripod (us humans are after all a type of fluid damping system, right?   ).

A more precise test could be made using controlled light pulses, but that's not as easy to do....

 ProfHankD's gear list:ProfHankD's gear list
Canon PowerShot A640 Canon PowerShot A720 IS Canon PowerShot S70 Canon PowerShot G1 Canon PowerShot G5 +22 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: does your 100-300 work better at slower than 1/25th?
In reply to RussellInCincinnati, Dec 1, 2013

RussellInCincinnati wrote:

Anders W wrote:It's a challenge to get tack-sharp results but it is certainly possible. Here's one available at full resolution (it is a crop) for your pixel-peeping pleasure:

Don't see how this could have been done better.

A tripod sometimes help but is no panacea. I have shutter-shock problems with my E-M5 and 100-300/4-5.6 at 300 mm in the critical range of shutter speeds (from about 1/25 to about 1/250) even in tripod-based shooting. In order to get sharp results with this combo, I simply have to move out of that range.

Are you saying that at 300mm, the results are sharper at 1/15th of a second than at 1/50th?

If shooting from a tripod so that we can rule out ordinary camera shake, yes. The shutter-shock problem peaks at about 1/100 s or slightly above that point. As you go higher, it disappears quite quickly and rather abruptly at a speed roughly corresponding to the flash sync speed (1/250 on the E-M5). As you go lower it disappears more gradually and it is more difficult to specify any particular limit.

The reasons are in all likelihood as follows: As you go to higher speeds, an increasingly small part of the displacement of the camera caused by the shutter will have time to register in the image. Possibly, the different behavior of the shutter above the flash sync limit also plays a part here. As you go to lower speeds, an increasingly small proportion of the entire exposure will be affected by the short-lived displacement of the camera.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Steen Bay
Veteran MemberPosts: 6,424
Like?
Re: On Sharpness, ISO and Shutter Speed
In reply to Anders W, Dec 1, 2013

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Ken Strain wrote:

These data are puzzling, two features disturb me most:

The spread between H and V gets to about 20% in several of the plots, including the points for the EM5 at 1/4000s. I am reluctant to accept that particular case as being due to shutter shock. If those data are bad, how can we tell which are good?

The EM1 plot is the only one that looks somewhat like I would expect if shutter shock were the main cause of the spread, but even here it appears that something else breaks the H-V symmetry with the opposite sign (such that there is roughly 6% discrepancy between the averaged H and V scopes of the 3 fastest speeds).

Hi Ken,

To me, the graphs for the E-M5, the A7R, and the D610 are pretty much in line with what would be expected in the presence of shutter shock, not just that of the E-M1. Of course, there may be other reasons for a difference between vertical and horizontal resolution too, as strongly suggested by the graphs for the D610 and A7.

A bit strange that the blur is mostly vertical with E-M1 and A7R, and mostly horizontal with D610.

You are right about that. I overlooked that it was the horizontal rather than vertical resolution that dipped in the critical range in this case. And I guess the D610 shutter (which is what distinguishes it from the D600 and its problems with dirt on the sensor) is moving vertically just as the other ones.

Seems that a good, heavy tripod isn't enough to prevent blur at certain shutter speeds, so could be interesting to know which shutter speeds are used by Photozone, Lenstip, DxO, SLRgear, etc. when testing lenses. Can we trust the results from cameras known to be affected?

Yes. The same thought has occurred to me. What I hope is that they prefer not to have more light in the studio where they do this than they really need, and that they therefore shoot at shutter speeds where shutter shock ceases to be much of a problem (although its effect may not be completely gone), like at 1/20 or below.

Guess it's most likely that the lighting is constant and that the shutter speed changes with the f-stop, like it changes with the ISO in DPR's studio scene.

That would be my guess as well.

At least SLR Gear are aware of the problem as you can see from their recent review of the new Pany 14-140 here:

http://slrgear.com/reviews/showproduct.php?product=1611

Yes, and that's with a 250 pounds tripod!

Yep. Although I suspect that there is something special about this lens and other Power OIS lenses with, as it seems, similar problems (X 14-42, X 45-175).

Yes, seems that the amount of blur will vary not only from camera to camera, but also from lens to lens, and also depends on the tripod, tripod head, hand helding technique, etc. It's rather complex.

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: Another possible experiment
In reply to ProfHankD, Dec 1, 2013

ProfHankD wrote:

Jack Hogan wrote:

2) Are the dips we see in MTF50 around 1/10th of a second to about 1/300th possibly due to shutter shock - or something else? What implications does this have for people choosing a high resolution camera today?

I'd guess that vibration is a factor here, and imprecision in electronic 1st curtain reset may be the cause of the other issue, but here's a possible experiment to try to resolve this....

In either a focal plane shutter or a not-really-simultaneous-global-electronic shutter, the sensor exposure interval is not really the same for all pixels. Some are exposed earlier and others later, with the overlap fraction dependent on shutter speed. Thus, it should be possible to profile the vibration/movement over time by using multiple targets positioned so that they are exposed in temporal sequence as the shutter travels across the frame. Shutter shock should be an impulse with periodic decay of the vibration over time, which I would expect to cause a measurable pattern in the MTF50 measurements along the shutter's travel direction.... The same test method could also be used to test external damping methods; for example, I'd actually expect human hands to do better damping than a tripod (us humans are after all a type of fluid damping system, right? ).

A more precise test could be made using controlled light pulses, but that's not as easy to do....

I have already conducted an experiment for the purpose of distinguishing between bell-type vibration and brute displacement of the camera as the mechanism whereby shutter action leads to blur. I did so by placing my E-M5 with my most shutter-shock-sensitive lens (Panasonic 100-300 at 300 mm, 600 mm EFL) on a very heavy and rigid surface (my kitchen floor with tiles on top of concrete). This prevents the camera from being displaced but not from vibrating. Result: No sign of blur. So at least in this case, brute displacement rather than vibration is the cause.

Note that I get clear blur due to shutter action in the critical range of shutter speeds when shooting the same combo from my reasonably sturdy tripod (Sirui T-1205X with G-10 head).

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: On Sharpness, ISO and Shutter Speed
In reply to Steen Bay, Dec 1, 2013

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Steen Bay wrote:

Anders W wrote:

Ken Strain wrote:

These data are puzzling, two features disturb me most:

The spread between H and V gets to about 20% in several of the plots, including the points for the EM5 at 1/4000s. I am reluctant to accept that particular case as being due to shutter shock. If those data are bad, how can we tell which are good?

The EM1 plot is the only one that looks somewhat like I would expect if shutter shock were the main cause of the spread, but even here it appears that something else breaks the H-V symmetry with the opposite sign (such that there is roughly 6% discrepancy between the averaged H and V scopes of the 3 fastest speeds).

Hi Ken,

To me, the graphs for the E-M5, the A7R, and the D610 are pretty much in line with what would be expected in the presence of shutter shock, not just that of the E-M1. Of course, there may be other reasons for a difference between vertical and horizontal resolution too, as strongly suggested by the graphs for the D610 and A7.

A bit strange that the blur is mostly vertical with E-M1 and A7R, and mostly horizontal with D610.

You are right about that. I overlooked that it was the horizontal rather than vertical resolution that dipped in the critical range in this case. And I guess the D610 shutter (which is what distinguishes it from the D600 and its problems with dirt on the sensor) is moving vertically just as the other ones.

Seems that a good, heavy tripod isn't enough to prevent blur at certain shutter speeds, so could be interesting to know which shutter speeds are used by Photozone, Lenstip, DxO, SLRgear, etc. when testing lenses. Can we trust the results from cameras known to be affected?

Yes. The same thought has occurred to me. What I hope is that they prefer not to have more light in the studio where they do this than they really need, and that they therefore shoot at shutter speeds where shutter shock ceases to be much of a problem (although its effect may not be completely gone), like at 1/20 or below.

Guess it's most likely that the lighting is constant and that the shutter speed changes with the f-stop, like it changes with the ISO in DPR's studio scene.

That would be my guess as well.

At least SLR Gear are aware of the problem as you can see from their recent review of the new Pany 14-140 here:

http://slrgear.com/reviews/showproduct.php?product=1611

Yes, and that's with a 250 pounds tripod!

Yep. Although I suspect that there is something special about this lens and other Power OIS lenses with, as it seems, similar problems (X 14-42, X 45-175).

Yes, seems that the amount of blur will vary not only from camera to camera, but also from lens to lens, and also depends on the tripod, tripod head, hand helding technique, etc. It's rather complex.

It certainly is. secretworld just told me in a thread on the MFT forum here

http://www.dpreview.com/forums/post/52626611

that he gets clear signs of shutter-shock blur with his 14 mm lens on an E-PM2. That's of course unexpected for such a short FL and he suspects that a less than perfectly fixed focus element is the culprit in this case.

Possibly, something similar is true for at least some copies of the 45/1.8 used for the DPR studio samples. I don't have much shutter-shock trouble with my own copy of that lens (provided that I hold the camera right). But some other people, including some whose observations I am inclined to trust, apparently do.

Interestingly, this was the lens commonly used in the examples from about one and a half years ago that really placed shutter shock on the MFT agenda. See the threads I link to in this post from that time:

http://www.dpreview.com/forums/thread/3247661

As you can see, some people experienced a problem and some others couldn't reproduce it. In at least some cases, I suspect that there is a genuine difference in the behavior of different specimen of the same gear rather than something else, although "something else" is certainly a possibility.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
tko
tko
Forum ProPosts: 10,455
Like?
Re: isn't it just the IBIS?
In reply to Anders W, Dec 1, 2013

In addition, consider the results for the A7R and the D610 in the OP.

The A7R goes from about 2200 to 2500, a 13% change in sharpness. Nothing like the IBIS 50% change over the same shutter speed range.

If you widen your horizons beyond the particular sample images discussed in this thread, you will find that it is already well established that cameras without a stabilization system suffer from shutter shock problems too.

I can only consider what's been analyzed, and there were a lot of cameras analyzed in this thread. The links you posted lead to opinion and guesswork, not graphs, so I would discount them. Anyone can have an opinion, only in this thread did we have analysis.

BTW: Only some OIS systems mechanically lock the IS lens group when the system is turned off. Others keep the IS lens group floating even when off. It's just that the input from its sensing devices is no longer considered.

No. How on can you have a lens element "float" with no power? And if the lens floats, it must be stabilized, or else the images will have blur. Please document this claim.

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: isn't it just the IBIS?
In reply to tko, Dec 1, 2013

tko wrote:

In addition, consider the results for the A7R and the D610 in the OP.

The A7R goes from about 2200 to 2500, a 13% change in sharpness. Nothing like the IBIS 50% change over the same shutter speed range.

So what? First, how do you explain that 13% decline for the A7R, which has no stabilization system? Second, there are lots of other variables that might explain why the effect is is more pronounced for the E-M1 and the A7R aside from IBIS, for example the lens used, the weight, and the shutter type.

If you widen your horizons beyond the particular sample images discussed in this thread, you will find that it is already well established that cameras without a stabilization system suffer from shutter shock problems too.

I can only consider what's been analyzed, and there were a lot of cameras analyzed in this thread. The links you posted lead to opinion and guesswork, not graphs, so I would discount them. Anyone can have an opinion, only in this thread did we have analysis.

Unlike you, I don't discount evidence on the basis of the form they take (e.g., graphs versus photos). That merely risks bias and/or incorrect conclusions about causality.

The data I point to have the advantage over the comparison between the A7R and the E-M1 that you prefer to make that they keep everything but the sensor and its being fixed or not constant or approximately constant: the lens used, the weight of the camera, and the type of shutter used (MFT cameras aside from the GM1 are reportedly using the same type of Copal shutter although its flash sync performance has increased a bit over time).

Those who have their mind open to evidence in the form of photos rather than graphs might also want to have a look at the tests reported here (see point 2)

http://cameraergonomics.blogspot.se/2012/05/micro-four-thirds-shutter-shake.html

as well as those reported here (where neither of the two cameras studied, NEX-5N and GF2 have IBIS):

http://www.microscopy-uk.org.uk/mag/artnov12/dw-SonyNEX5N.html

BTW: Only some OIS systems mechanically lock the IS lens group when the system is turned off. Others keep the IS lens group floating even when off. It's just that the input from its sensing devices is no longer considered.

No. How on can you have a lens element "float" with no power?

It has power all the time when the lens is mounted on the camera and the camera is on, regardless of whether OIS is on or off. When the lens is dismounted or when the camera is off, the IS lens group is free to move as it pleases.

And if the lens floats, it must be stabilized, or else the images will have blur. Please document this claim.

It is of course stabilized in the sense of being held in place even when the OIS system is turned off. The IS lens group is in this case "floating" (rather than locked mechanically) in a fixed position instead of moving around in response to the motion sensors.

An example of an OIS system that works like this is that used by Panasonic for their MFT lenses. If you shake the lens lightly when dismounted, you will hear the IS group rattle around. With the lens mounted and the camera on, you will hear a hissing sound from the OIS system regardless of whether it is on or off.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
ProfHankD
Senior MemberPosts: 2,160Gear list
Like?
Re: Another possible experiment
In reply to Anders W, Dec 1, 2013

Anders W wrote:

This prevents the camera from being displaced but not from vibrating. Result: No sign of blur. So at least in this case, brute displacement rather than vibration is the cause.

Well, I've never trusted tripods as much as bracing on part of a building anyway....  

For what it's worth, even with everything bolted to a concrete block, many cameras do suffer from what you're calling physical displacement. For example, bolting-down a camera and taking a series of non-blurry images, many camera+lens combos (even without mechanical image stabilization) will not quite have images align at the pixel level. The culprit there is apparently play in the lens system, which can be made much more visible if the system tries to focus for each shot, but can be induced by things as simple as the whack from the aperture stop-down lever on Sony A-mounts. Old (high-mass, tight-tolerance) manual lenses don't seem to have this problem at all.

 ProfHankD's gear list:ProfHankD's gear list
Canon PowerShot A640 Canon PowerShot A720 IS Canon PowerShot S70 Canon PowerShot G1 Canon PowerShot G5 +22 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: Another possible experiment
In reply to ProfHankD, Dec 1, 2013

ProfHankD wrote:

Anders W wrote:

This prevents the camera from being displaced but not from vibrating. Result: No sign of blur. So at least in this case, brute displacement rather than vibration is the cause.

Well, I've never trusted tripods as much as bracing on part of a building anyway....

For what it's worth, even with everything bolted to a concrete block, many cameras do suffer from what you're calling physical displacement.

Well, I meant displacement of the entire camera rather than parts of it (as in your example below). In my example, we know that vibration is not the mechanism because there is no blur when displacement of the entire camera is prevented. In your example, vibration can be involved since you get blur even when displacement of the entire camera is prevented.

Obviously, we cannot rule out the possibility that the results of my experiment, using a particular lens on a particular camera, does not generalize to other cases. What it does show is that you can get blur due to shutter action that is not mediated by vibration but by displacement of the body as a whole.

One possibility that we are discussing with regard to the DPR studio scene samples at issue in this tread is that the problem might be due to something being slightly loose in the particular specimen of the lens used. See here:

http://www.dpreview.com/forums/post/52627215

One suprising thing about these particular samples is that they are presumably shot on a heavy studio tripod, which ordinarily should prevent the camera as a whole from being displaced to any significant degree due to shutter action. Furthermore, the FL used isn't very long, which should also limit the amount of blur due to any remaining displacement of the camera as a whole.

For example, bolting-down a camera and taking a series of non-blurry images, many camera+lens combos (even without mechanical image stabilization) will not quite have images align at the pixel level. The culprit there is apparently play in the lens system, which can be made much more visible if the system tries to focus for each shot, but can be induced by things as simple as the whack from the aperture stop-down lever on Sony A-mounts. Old (high-mass, tight-tolerance) manual lenses don't seem to have this problem at all.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
ProfHankD
Senior MemberPosts: 2,160Gear list
Like?
Re: Another possible experiment
In reply to Anders W, Dec 1, 2013

Anders W wrote: Well, I meant displacement of the entire camera rather than parts of it (as in your example below).

I understood you meant of the camera as a unit, but to me the difference between vibration and displacement is that one has a periodic component while the other doesn't. I haven't seen vibration due to play in the lens, just displacement -- which is why I mentioned it. Probably not the prime cause in the DPReview images -- more often a problem with cheap kit zooms -- but this could explain some of the anecdotal stories we're hearing.

One possibility that we are discussing with regard to the DPR studio scene samples at issue in this tread is that the problem might be due to something being slightly loose in the particular specimen of the lens used.

Agreed that a defective lens could do a world of harm. It could even be something as simple as an intermittent electrical contact or power drain that makes a lens focus/aperture drive motor twitch.

One suprising thing about these particular samples is that they are presumably shot on a heavy studio tripod

I don't trust tripods to be solid at all frequencies. They certainly have harmonic frequencies at which they will vibrate, so an impulse that would only cause minor displacement could easily drive vibration of a tripod that magnifies the problem.

 ProfHankD's gear list:ProfHankD's gear list
Canon PowerShot A640 Canon PowerShot A720 IS Canon PowerShot S70 Canon PowerShot G1 Canon PowerShot G5 +22 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: Another possible experiment
In reply to ProfHankD, Dec 1, 2013

ProfHankD wrote:

Anders W wrote: Well, I meant displacement of the entire camera rather than parts of it (as in your example below).

I understood you meant of the camera as a unit, but to me the difference between vibration and displacement is that one has a periodic component while the other doesn't. I haven't seen vibration due to play in the lens, just displacement -- which is why I mentioned it. Probably not the prime cause in the DPReview images -- more often a problem with cheap kit zooms -- but this could explain some of the anecdotal stories we're hearing.

How can you in practice, in the present context, distinguish between vibration inside the lens and displacement inside the lens as a cause of blur?

One possibility that we are discussing with regard to the DPR studio scene samples at issue in this tread is that the problem might be due to something being slightly loose in the particular specimen of the lens used.

Agreed that a defective lens could do a world of harm. It could even be something as simple as an intermittent electrical contact or power drain that makes a lens focus/aperture drive motor twitch.

One suprising thing about these particular samples is that they are presumably shot on a heavy studio tripod

I don't trust tripods to be solid at all frequencies. They certainly have harmonic frequencies at which they will vibrate, so an impulse that would only cause minor displacement could easily drive vibration of a tripod that magnifies the problem.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
The_Suede
Contributing MemberPosts: 603
Like?
Quite a lot of problem parameters...
In reply to Jack Hogan, Dec 1, 2013

A mechanical resonance in a complex system is quite complex.

The problems with most sensor-stabilized systems are well known, and there's even been some formal attempts at quantifying them. Falk Lumo, also present here on DPR, among others have some material in the public domain concerning mirror and shutter vibration effect on image sharpness.

Recent cameras (modern constructions) have very low mirror-slap impact on sharpness, but stronger shutter-induced vibrations.

Since it's a damped resonance, you have a fast motion peak buildup and a slightly damped motion fade-out.

Some things that has an effect on this vibration displacement maximum in time is:

  • Lens>body system resonance (when hand-held)
  • Body flexibility between tripod plate and sensor-cage and lens mount (on tripod)
  • Sensor mount compliance/damping (for sensor-stabilized cameras)

Then different combinations of lens weight, mount-point (lens collar or camera), lens mount stability and so on further modulate the effect.

In general you have a vibration movement maxima coinciding with a point about 20-50ms after the first curtain hits the end stop. This means that the maximum vertical vibration effect is visible in exposures between 1/160s and up to 1/80s. As you get longer exposure times, the very short peak in vibration has less effect on the overall light throughput.

When the exposure is 2x longer than the time that fully encloses the full vibration peak (1/80-1/100s?) you have half of the exposure in almost stationary mode, also halving the shake PSF effect. At 1/20-1/25s you have pushed the shutter vibration two full stops down in effect, leaving just a slight lowering in MTF, and almost certainly no visible effect.

At 1/200s you're left with an effect that's equally strong, but only visible in the top half of the image..

On DSLR-type cameras with no sensor-stabilization, the highest amount of shutter-induced shake you get is typically when you shoot with longer lenses mounted to a tripod via the lens-collar foot.

.....................

Reply   Reply with quote   Complain
ProfHankD
Senior MemberPosts: 2,160Gear list
Like?
Re: Another possible experiment
In reply to Anders W, Dec 1, 2013

Anders W wrote:

ProfHankD wrote:

Anders W wrote: Well, I meant displacement of the entire camera rather than parts of it (as in your example below).

I understood you meant of the camera as a unit, but to me the difference between vibration and displacement is that one has a periodic component while the other doesn't. I haven't seen vibration due to play in the lens, just displacement -- which is why I mentioned it. Probably not the prime cause in the DPReview images -- more often a problem with cheap kit zooms -- but this could explain some of the anecdotal stories we're hearing.

How can you in practice, in the present context, distinguish between vibration inside the lens and displacement inside the lens as a cause of blur?

Two things:

  1. As I said, I personally have not seen blur from play in a lens, but displacement from frame to frame. (Theoretically, if it were just vibration, then there shouldn't be any displacement from frame to frame.)
  2. The "another possible experiment" I suggested, that started this thread, would resolve at precisely what points in the exposure cycle displacements happen.

Of course, I don't know if the lenses DPReview tests with have any such issues. I don't see any way to be sure of the cause of the blur just by examining the DPReview comparison images.

 ProfHankD's gear list:ProfHankD's gear list
Canon PowerShot A640 Canon PowerShot A720 IS Canon PowerShot S70 Canon PowerShot G1 Canon PowerShot G5 +22 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 19,066Gear list
Like?
Re: Another possible experiment
In reply to ProfHankD, Dec 1, 2013

ProfHankD wrote:

Anders W wrote:

ProfHankD wrote:

Anders W wrote: Well, I meant displacement of the entire camera rather than parts of it (as in your example below).

I understood you meant of the camera as a unit, but to me the difference between vibration and displacement is that one has a periodic component while the other doesn't. I haven't seen vibration due to play in the lens, just displacement -- which is why I mentioned it. Probably not the prime cause in the DPReview images -- more often a problem with cheap kit zooms -- but this could explain some of the anecdotal stories we're hearing.

How can you in practice, in the present context, distinguish between vibration inside the lens and displacement inside the lens as a cause of blur?

Two things:

  1. As I said, I personally have not seen blur from play in a lens, but displacement from frame to frame. (Theoretically, if it were just vibration, then there shouldn't be any displacement from frame to frame.)
  2. The "another possible experiment" I suggested, that started this thread, would resolve at precisely what points in the exposure cycle displacements happen.

But doesn't point one suggest that the displacement of whatever it is that is displaced happens between rather than during any given exposure, in which case it wouldn't really be relevant in the present context (that of specifying the reasons for blur)?

Of course, I don't know if the lenses DPReview tests with have any such issues. I don't see any way to be sure of the cause of the blur just by examining the DPReview comparison images.

No. I don't think we can be sure of the cause of the blur just by examining these images either. Further tests, which only DPR themselves are in a position to carry out, are most likely required.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads