Electronic Shutter Readout & Hand Shake (dumb question no. 1)

[Overgaard]

Advertisement (gone after registration)

Yet another, this one I actually like the most warped one with the motorbike warping away into the right of the frame in the 2nd last shot of the series. Looks like an old 21mm lens even it is a 50mm APO. A burst at 1/1000 electronic shutter speed.

If you notice the green street sign above, you can sense how I'm about to move the camera away in the last frame, and that creates an additional warp of the entire frame. 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited May 9, 2022 by Overgaard

[Overgaard]

And a last one, at 1/640 second burst of a Rolls Royce passing by, warped into a brand-new futuristic model. 50mm APO. You can clearly sense here how the sensor reads from top row and down, and there's a delay of the readout. 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[graphlex]

Reminds me of Lartigue’s race cars, except that with his cars the top of each tire was further along than the bottom. 

[AndrewLT]

@Overgaard's first two snaps are interesting in two respects: they illustrate the effect we've been discussing, but they also demonstrate that, for me at least (YMMV), in many instances, it really won't matter too much.  The last one is a perfectly good example of what I believe is referred to as "rolling shutter".  Surely, this is precisely the effect that one would expect from a sensor with a relatively slow readout (in this case, 1/10s)?

At the risk of falling into geekdom, a couple of things strike me about this image (with apologies for what is probably completely irrelevant musing, so please don't waste your time reading this):

1. This is a long vehicle, about 5.2m.  Judging by the relative position of the wooden pole behind, I estimate that, between the two frames, the Rolls has moved c.3m.  If (big assumption no. 1) these were shot in continuous mode at 4.5fps, that implies a speed of c.13.5m/s, or c.30mph, which seems plausible.
2. Now, the rear of the boot (= trunk in other dialects) actually slopes back slightly so, based on the size of the straight line immediately behind the tail light, I estimate that the bottom of that line is displaced by approx. 20cm, relative to the top of the boot / trunk.
3. Vertically, that section covers just under a third of the image, so the difference in time between the readout of the top and the bottom of that section will be approx. 0.03s.  
4. In 0.03s, if the car is moving forwards at 30mph, it would cover c.40cm.

Perhaps that implies that the image has been cropped by roughly 50% in each dimension (which would halve the readout time gap)?  There are other possibilities, of course (there are too many unknowns to be sure).  For example, roughly half a second between the images (i.e. not 4.5fps continuous shooting) would imply a speed of around 15mph and a horizontal displacement of more like 20cm, which is what I estimate the displacement to be.

Incidentally, if the images are not cropped, the field of view of a 50mm lens implies that it was taken from c.3.5m, or c.7m if each dimension was halved, either of which seem plausible.  Or something else.

But the most likely thing is that I have messed up the arithmetic.  It is rare for me not to get at least one calculation wrong each day, so it's nice to get today's boo-boo out of the way before lunch.

At any rate, the Jelly Roller illustrates exactly what I think we'd expect to see.  None of this makes me less tempted by the M11 ....😉

Andrew

Edited May 9, 2022 by AndrewLT

[LocalHero1953]

1 hour ago, AndrewLT said:

@Overgaard's first two snaps are interesting in two respects: they illustrate the effect we've been discussing, but they also demonstrate that, for me at least (YMMV), in many instances, it really won't matter too much.  The last one is a perfectly good example of what I believe is referred to as "rolling shutter".  Surely, this is precisely the effect that one would expect from a sensor with a relatively slow readout (in this case, 1/10s)?

At the risk of falling into geekdom, a couple of things strike me about this image (with apologies for what is probably completely irrelevant musing, so please don't waste your time reading this):

1. This is a long vehicle, about 5.2m.  Judging by the relative position of the wooden pole behind, I estimate that, between the two frames, the Rolls has moved c.3m.  If (big assumption no. 1) these were shot in continuous mode at 4.5fps, that implies a speed of c.13.5m/s, or c.30mph, which seems plausible.
2. Now, the rear of the boot (= trunk in other dialects) actually slopes back slightly so, based on the size of the straight line immediately behind the tail light, I estimate that the bottom of that line is displaced by approx. 20cm, relative to the top of the boot / trunk.
3. Vertically, that section covers just under a third of the image, so the difference in time between the readout of the top and the bottom of that section will be approx. 0.03s.  
4. In 0.03s, if the car is moving forwards at 30mph, it would cover c.40cm.

Perhaps that implies that the image has been cropped by roughly 50% in each dimension (which would halve the readout time gap)?  There are other possibilities, of course (there are too many unknowns to be sure).  For example, roughly half a second between the images (i.e. not 4.5fps continuous shooting) would imply a speed of around 15mph and a horizontal displacement of more like 20cm, which is what I estimate the displacement to be.

Incidentally, if the images are not cropped, the field of view of a 50mm lens implies that it was taken from c.3.5m, or c.7m if each dimension was halved, either of which seem plausible.  Or something else.

But the most likely thing is that I have messed up the arithmetic.  It is rare for me not to get at least one calculation wrong each day, so it's nice to get today's boo-boo out of the way before lunch.

At any rate, the Jelly Roller illustrates exactly what I think we'd expect to see.  None of this makes me less tempted by the M11 ....😉

Andrew

This is an easy test anyone can do.
Q2 panned across a window frame. I aimed to take a second to pan across the visible scene (I listened for the 1 second tick of a long case clock), and pressed the shutter as I faced the middle. The shot was taken at 1/50s, hence the motion blur. Measure the amount of  distortion and divide by the width of the frame - that is the fraction of a second that the sensor takes to be read: 55-60ms in this case, or about 1/20sec. This was why I was surprised that the M11 was slower, but I accept that higher res sensors take more time.

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Edited May 9, 2022 by LocalHero1953

[Stuart Richardson]

I believe the conclusion is: "don't use the electronic shutter handheld or for moving objects". Good thing it has a mechanical shutter. So I guess the only reason to not just do that is the impression that the M11 mechanical shutter is disturbing? I have not seen it or heard it, so I cannot comment on that. Personally, I have only really used the e-shutter to minimize shutter shock when using a tripod and photographing things like artwork or macro work. Occasionally I have used it for long telephotos on a tripod, as the weight, balance and focal length tend to amplify the shutter shock substantially. But I will be more careful with moving objects in the frame.

With respect to the M10R vs M11, is there is a big difference here other than if one thinks the M11 mechanical shutter is undesirable, as Thorsten has indicated in his articles? It seems like the choice is using a mechanical shutter on the M10R, or using a mechanical shutter on the M11, with the possibility of having an e shutter for the times when it useful (such as on a tripod when doing delicate shots that are sensitive to vibration). 

[jplomley]

Advertisement (gone after registration)

I have an insane and appalling amount of shutter shock on my SL2 at 1/125 -1/500th (my AutoISO shutter speed is 1/250th) that notably degrades the image when hand-held using either the 35 or 75 APO L lenses. I was truly shocked at the hit on image quality. I have since switched completely to eshutter and the problem has disappeared and I am getting the performance out of those fantastic APO L primes. Have not noticed any issue with implementing the eshutter for street. Is the read out speed different on the SL2? 

[FrankX]

Checking a flickering fluorescent lamp yields:

M11: about 1/10 s

SL2: about 1/25 s 

Fuji X-T3: about 1/30 s

[Overgaard]

8 hours ago, AndrewLT said:

@Overgaard's first two snaps are interesting in two respects: they illustrate the effect we've been discussing, but they also demonstrate that, for me at least (YMMV), in many instances, it really won't matter too much.  The last one is a perfectly good example of what I believe is referred to as "rolling shutter".  Surely, this is precisely the effect that one would expect from a sensor with a relatively slow readout (in this case, 1/10s)?

At the risk of falling into geekdom, a couple of things strike me about this image (with apologies for what is probably completely irrelevant musing, so please don't waste your time reading this):

1. This is a long vehicle, about 5.2m.  Judging by the relative position of the wooden pole behind, I estimate that, between the two frames, the Rolls has moved c.3m.  If (big assumption no. 1) these were shot in continuous mode at 4.5fps, that implies a speed of c.13.5m/s, or c.30mph, which seems plausible.
2. Now, the rear of the boot (= trunk in other dialects) actually slopes back slightly so, based on the size of the straight line immediately behind the tail light, I estimate that the bottom of that line is displaced by approx. 20cm, relative to the top of the boot / trunk.
3. Vertically, that section covers just under a third of the image, so the difference in time between the readout of the top and the bottom of that section will be approx. 0.03s.  
4. In 0.03s, if the car is moving forwards at 30mph, it would cover c.40cm.

Perhaps that implies that the image has been cropped by roughly 50% in each dimension (which would halve the readout time gap)?  There are other possibilities, of course (there are too many unknowns to be sure).  For example, roughly half a second between the images (i.e. not 4.5fps continuous shooting) would imply a speed of around 15mph and a horizontal displacement of more like 20cm, which is what I estimate the displacement to be.

Incidentally, if the images are not cropped, the field of view of a 50mm lens implies that it was taken from c.3.5m, or c.7m if each dimension was halved, either of which seem plausible.  Or something else.

But the most likely thing is that I have messed up the arithmetic.  It is rare for me not to get at least one calculation wrong each day, so it's nice to get today's boo-boo out of the way before lunch.

At any rate, the Jelly Roller illustrates exactly what I think we'd expect to see.  None of this makes me less tempted by the M11 ....😉

Andrew

50mm and no crop. It is what it is, if you move the camera during shooting, or the object moves, it's gonna "roll" or "warp".

Here's a scene where I let the camera drop down after I took the last shot. 1/10,000 and 1/12,500 of a sec. Could have held the camera there for another second, but I have the habit of taking the photo and let the camera drop back to the hip. 

To avoid warp, include a 'grace period' of 2 seconds after the last photo. Then the warp will be microscopic. 

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[jonoslack]

On 5/9/2022 at 6:54 AM, Overgaard said:

And here is another one of the warping effect of the Leica M11 electronic shutter. Camera handheld of course, and not much movement, a series taken within 2 seconds at 1/250 sec.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

Hi Thorsten

It is easy to demonstrate, and the same as most other cameras with electronic shutters, but I don’t think it’s really much of an issue as long as you understand what it’s good for and what it isn’t (like banding in artificial light)

Best

[lct]

22 hours ago, Stuart Richardson said:

[...] I believe the conclusion is: "don't use the electronic shutter handheld or for moving objects".[...]

OK for moving objects and artificial light but i mostly use e-shutter handheld on my Sony A7r2 and Leica CL and i've never noticed the least issue i must say. I just disable e-shutter on moving subject matters and artificial light that's all. YMMV. On the M11 the shutter is so discrete that i have not used the e-shutter yet but it is another story.

[hdmesa]

10 hours ago, jonoslack said:

Hi Thorsten

It is easy to demonstrate, and the same as most other cameras with electronic shutters, but I don’t think it’s really much of an issue as long as you understand what it’s good for and what it isn’t (like banding in artificial light)

Best

Well, we would all like for it to be good for handheld snaps when subject is not moving and the photographer is trying not to, but TO's examples show that casual movement on the part of the photographer demand use of the mechanical shutter.

I think the only handheld use for the readout speed of the M11 electronic shutter is when concentrating in order to hold still during the photo and without taking a series of shots – for example, holding still while taking a landscape shot handheld.

Edited May 10, 2022 by hdmesa

[Overgaard]

15 hours ago, jonoslack said:

Hi Thorsten

It is easy to demonstrate, and the same as most other cameras with electronic shutters, but I don’t think it’s really much of an issue as long as you understand what it’s good for and what it isn’t (like banding in artificial light)

Best

Yes, that is the drill on this, with the Leica M11. I never found warping to be a noticeable problem on other cameras, likely because the SL2 and Q2 have faster readout time.

The banding in artificial light. Yes, that is another story. I find that mechanical shutter is the only safe bet, which is easy to do as artificial light seldom demand higher than 1/4000 shutter speeds. 

[Tailwagger]

On 5/10/2022 at 8:06 PM, Overgaard said:

The banding in artificial light. Yes, that is another story. I find that mechanical shutter is the only safe bet, which is easy to do as artificial light seldom demand higher than 1/4000 shutter speeds. 

Unfortunately banding due to LED lighting is something I encounter fairly frequently when shooting with the e-shutter as I do a fair bit of concert and theatre work where often silent operation is required to avoid a fist fight with audience members. The sad/good part, depending on your POV, is that it's really hit or miss as to whether or not you encounter the problem. At my last concert gig, I probably lost somewhere between one in five to one in ten frames to banding.  And of course, typical for me, the better shots in the series wind up being the ones I can't use unless I'm willing pretend that image was made by shooting a CRT.  Seems like some form of corrective software might be a nice feature to add to Topaz AI Denoise. SL2 in this case...

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

 

[jaapv]

On 5/10/2022 at 10:08 AM, jonoslack said:

Hi Thorsten

It is easy to demonstrate, and the same as most other cameras with electronic shutters, but I don’t think it’s really much of an issue as long as you understand what it’s good for and what it isn’t (like banding in artificial light)

Best

That is very true, but from observation I would say that people start to notice with cameras that have a higher resolution than 24 MP 

[hdmesa]

1 hour ago, jaapv said:

That is very true, but from observation I would say that people start to notice with cameras that have a higher resolution than 24 MP 

–> because with non-stacked sensors, sensor readout/scan speed gets slower with higher resolution.

With a stacked sensors like the ones used in the Z9 and A1, you can have the best of both worlds.

Edited May 15, 2022 by hdmesa

[SrMi]

5 hours ago, jaapv said:

That is very true, but from observation I would say that people start to notice with cameras that have a higher resolution than 24 MP 

I believe that the focal length and the presence/absence of stabilization (IBIS or OIS) also play a role.

[jaapv]

9 hours ago, SrMi said:

I believe that the focal length and the presence/absence of stabilization (IBIS or OIS) also play a role.

Of course. And the ability of the photographer to hold the camera steady  And the technical understanding that the camera must be held steady after the theoretical shutter time, as this  may be short, but the time that the shutter is exposing is longer. 

[hdmesa]

8 hours ago, jaapv said:

... the camera must be held steady after the theoretical shutter time, as this  may be short, but the time that the shutter is exposing is longer. 

Yes, this. The scan time of the electronic shutter on the GFX 50S/R cameras was so slow, it felt like I couldn't move for almost a full second. I could literally hear the electronics buzzing inside while it was scanning the sensor!

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[J12]

On 5/13/2022 at 8:19 PM, Tailwagger said:

Unfortunately banding due to LED lighting is something I encounter fairly frequently when shooting with the e-shutter as I do a fair bit of concert and theatre work where often silent operation is required to avoid a fist fight with audience members. The sad/good part, depending on your POV, is that it's really hit or miss as to whether or not you encounter the problem. At my last concert gig, I probably lost somewhere between one in five to one in ten frames to banding.  And of course, typical for me, the better shots in the series wind up being the ones I can't use unless I'm willing pretend that image was made by shooting a CRT.  Seems like some form of corrective software might be a nice feature to add to Topaz AI Denoise. SL2 in this case...

 

 

I find it possible to mitigate banding with the electronic shutter on cameras like the SL and TL2. Though sometimes this can require resorting to slower shutter speeds. The general rule is to set shutter speeds in correlation with the cycles of the power grid in your country or region. 50hz or 60hz.

 

There’s a good explanation on the workings of both mechanical and electronic shutters here:

https://johnplatt.com.au/electronic-shutters-banding-rolling-shutter-explained/

 

The unique benefit of an M camera with electronic shutter function would be the ability to use it with either the OVF (rangefinder) or Visoflex EVF attachment.

I often shoot in situations where absolute silence is required. Electronic shutter functionality is of great importance when considering a new digital camera.

Unfortunately the very slow readout speed of the sensor in the M11 makes the electronic shutter function impractical for many situations.

As good as EVFs can be, there are times when I’d prefer to use an OVF / rangefinder. Looking through an EVF for extended periods all day can also result in considerable vision fatigue. Use of the electronic shutter with OVF would provide welcome eye relief.