Long shutter speeds, aperture priority

[helged]

Advertisement (gone after registration)

Can someone explain why aperture priority limits long exposures on SLx to 60 sec (*), despite the longest possible shutter speed can much longer, e.g. 30 min for SL2-S? I ask since this limits the use of e.g. 'slow' time-lapse-type of photography In low light; say taking one photo every 30-60 min or so for days or more.

Yes, bulb is available in manual mode, but this is not useful for 'slow' time-lapse-type of photography lasting from days to weeks or more, since in this case the camera have to run in some sort of  automatic mode. Alternatively, one is dependent on a sensor measuring the light as seen by the camera through the actual lens, or at least guesstimating that light, followed by operating the shutter speed in bulb mode, implying a time-lapse-timer with a light sensor.

Please feel free to chime in if there is something that I have missed... To illustrate sort of what I mean, have a look at the post below.

Let me also add that e.g. Leica S3 allows shutter speeds up to 8 min in aperture priority mode.

(*) The 60 sec limit on shutter speeds in aperture mode holds for all SLx- and Lumix S-bodies, as far as I can tell. 

 

 

 

[Photoworks]

I like to point out that the time changes at different ISO. if you have6400 I think you only get 4sec.

On longer exposures you need to consider noise reduction. so if you have a 30min exposure you need to add 30min noise reduction.

and for the days on end option power will bayou bigger dilemma. I suggest full internal battery, possible hand grip with battery and a USB-C battery that can be changes without moving the camera.

[Advertisement]

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

[farnz]

2 hours ago, helged said:

Can someone explain why aperture priority limits long exposures on SLx to 60 sec (*), despite the longest possible shutter speed can much longer, e.g. 30 min for SL2-S?

I suspect it's to do with the heat produced by the CMOS sensor that could damage the sensor if allowed to 'runaway' over a long period. 

The difference in long exposure that's allowed between the SLx and the SL2-S is likely to be because the SL2-S's sensor is Back Side Illuminated (BSI) and heat produced in the sensor is able to dissipate more easily/effectively.  Just a guess.

Pete.

[helged]

10 minutes ago, farnz said:

I suspect it's to do with the heat produced by the CMOS sensor that could damage the sensor if allowed to 'runaway' over a long period. 

The difference in long exposure that's allowed between the SLx and the SL2-S is likely to be because the SL2-S's sensor is Back Side Illuminated (BSI) and heat produced in the sensor is able to dissipate more easily/effectively.  Just a guess.

Pete.

I follow you - at least partly... Heat production is/can be an issue, but I don't understand why SL2-S, as an example, allows 30 min shutter speed in Manual mode, but ends at 60 sec in Aperture mode. The heat problem should be independent on which mode you are in. Yes, I understand that metering low light can be/is a challenge. But it is still hard to understand why shutter speeds and at 60 sec in Aperture mode.

[farnz]

Perhaps it's a software quirk that needs sorting out?  Beyond that I don't know.

Pete.

[helged]

26 minutes ago, Photoworks said:

I like to point out that the time changes at different ISO. if you have6400 I think you only get 4sec.

On longer exposures you need to consider noise reduction. so if you have a 30min exposure you need to add 30min noise reduction.

and for the days on end option power will bayou bigger dilemma. I suggest full internal battery, possible hand grip with battery and a USB-C battery that can be changes without moving the camera.

Yes, 30 min shutter speed is only possible at base ISO (ok for me). And yes, battery consumption is a problem (but the battery/camera can be powered by an USB battery packs etc.). And yes, mandatory NR double the shutter speed. But neither of these objections explain why 30 min works in Manual mode (at base ISO), but not in Aperture mode (max 60 sec shutter speed)... 

[helged]

Advertisement (gone after registration)

2 minutes ago, farnz said:

Perhaps it's a software quirk that needs sorting out?  Beyond that I don't know.

Pete.

Likely; I may ask a Leica representative...

[beewee]

7 hours ago, helged said:

Yes, I understand that metering low light can be/is a challenge. But it is still hard to understand why shutter speeds and at 60 sec in Aperture mode.

I suspect it’s because when you are in manual mode, you’re not using the camera to meter the brightness. When its dark enough that a 60 second exposure doesn’t cut it, it’s probably too dark for the camera to meter with a degree of accuracy to decide what’s actually a reasonable shutter speed.

At some point, everything just look almost ‘black’ to the camera. At very low light, the metering accuracy might exceed 1 stop so for a 120s exposure, the camera could recommend anything from 60s up to 240s. It wouldn’t be pleasant if you took one shot and it recommended 60s and the next one took 240s in aperture priority. So probably for that reason, Leica just limits it to 60s to avoid frustrating users when the camera is not able to provide a reliable/accurate exposure estimate.

[farnz]

1 hour ago, beewee said:

At very low light, the metering accuracy might exceed 1 stop

A non-linear behaviour that's reminiscent of Reciprocity Error in its effect (but not in its cause)?

Pete.

[beewee]

36 minutes ago, farnz said:

A non-linear behaviour that's reminiscent of Reciprocity Error in its effect (but not in its cause)?

Pete.

I’m not suggesting any non-linearity. All I’m saying is that signal to noise ratio will drop to a point where metering no longer becomes reliable due to lack of signal. It’s  possible that there’s a hard limit where quantization error in the metering measurement becomes problematic. I’m not suggesting there is any one reason, just that there are likely technical limitations related to metering when it gets too dark to do so accurately.

That being said, I don’t think reciprocity error is the reason. Digital signals are inherently linear. Analog to digital converters generally saturate or clip at their maximum and is noise dominant when there is a lack of signal. Amplification circuits usually create problems when they approach saturation but don’t have linearity issues when the signal is low. They just amplify noise that’s already in the system.

Edited March 27, 2021 by beewee

[farnz]

8 minutes ago, beewee said:

I’m not suggesting any non-linearity. All I’m saying is that signal to noise ratio will drop to a point where metering no longer becomes reliable due to lack of signal. It’s  possible that there’s a hard limit where quantization error in the metering measurement becomes problematic. I’m not suggesting there is any one reason, just that there are likely technical limitations related to metering when it gets too dark to do so accurately.

That being said, I don’t think reciprocity error is the reason. Digital signals are inherently linear. Analog to digital converters generally saturate or clip at their maximum and is noise dominant when there is a lack of signal. Amplification circuits usually create problems when they approach saturation but don’t have linearity issues when the signal is low. They just amplify noise that’s already in the system.

I agree with the concept of quantizing errors being a possible factor particularly as the signal approaches the noise floor.  I wasn't suggesting Reciprocity Error per se, rather behaviour that is similar in nature.  I recognise that response in digital sensors is linear and Reciprocity is therefore not a factor. 

I recall reading a paper that reasoned that the activity of 'filling up' the photosite wells with radiant energy is an analogue function, which suggests the possibility of Reciprocity in digital sensors at quantum level but I remain unconvinced.

Pete.

[helged]

4 hours ago, beewee said:

I suspect it’s because when you are in manual mode, you’re not using the camera to meter the brightness. When its dark enough that a 60 second exposure doesn’t cut it, it’s probably too dark for the camera to meter with a degree of accuracy to decide what’s actually a reasonable shutter speed.

At some point, everything just look almost ‘black’ to the camera. At very low light, the metering accuracy might exceed 1 stop so for a 120s exposure, the camera could recommend anything from 60s up to 240s. It wouldn’t be pleasant if you took one shot and it recommended 60s and the next one took 240s in aperture priority. So probably for that reason, Leica just limits it to 60s to avoid frustrating users when the camera is not able to provide a reliable/accurate exposure estimate.

Thanks; yes it may very well be linked to the accuracy of light metering.

Interestingly though, the S3 allows 8 min exposure (at base ISO) in Aperture mode, and the metering is quite fine (have tested it several times running the camera in time-lapse-mode for several days in row). 

[beewee]

1 hour ago, helged said:

Thanks; yes it may very well be linked to the accuracy of light metering.

Interestingly though, the S3 allows 8 min exposure (at base ISO) in Aperture mode, and the metering is quite fine (have tested it several times running the camera in time-lapse-mode for several days in row). 

I’m guessing the main difference is that the SL2/SL2-S is metered using the main sensor whereas the S3 has a dedicated light meter like an SLR so it can achieve higher sensitivity. The SL2/SL2-S sensor metering is limited by the noise of the imaging sensor.

[helged]

21 minutes ago, beewee said:

I’m guessing the main difference is that the SL2/SL2-S is metered using the main sensor whereas the S3 has a dedicated light meter like an SLR so it can achieve higher sensitivity. The SL2/SL2-S sensor metering is limited by the noise of the imaging sensor.

Yes & good point, the metering could be a mirror vs mirrorless difference. 

[helged]

Coming back to this question: Why is the longest shutter speed on SL2-S, when used in aperture priority mode, limited to 60 sec?

A recap (SL2-S, FW 3.1, at ISO 50 or ISO 100):

• Aperture priority mode [A]: Max shutter speed = 60 sec
• Automatic program mode [P]: 60 sec
• Shutter speed priority mode [S]: 30 min
• Manual exposure setting mode [M]: 30 min

The explanation that accurate sensor metering fails in very low light seems to be in contradiction with the fact that the shutter speed extends to 30 min in [S] mode. Even using exposure bracketing in [A]-mode, max shutter speed is limited to 60 sec.

Strange. For me, at least...

The reason I would like to use SL2-S at long shutter speeds in [A]-mode is to run type of time lapse during night: You put the camera on a tripod, you dial in aperture, set the ISO to 100, and let the camera determine reasonable shutter speeds. This works brilliantly with eg S3 up to 8 min, although S3 is another type of animal. The SL2-S is a very fine camera in low light - actually the best Leica camera I have used - so it would be good to get beyond the 60 sec limitation in [A]-mode. I would think that exposure bracketing could extend the longest shutter speed beyond 60 sec. But no. I will contact Leica, but it would be good to know any reasonable reasons for this limitation...

[farnz]

On 3/26/2021 at 5:26 PM, helged said:

Likely; I may ask a Leica representative...

Did you manage to get a response from any Leica representative, Helged?

Pete.

[helged]

1 hour ago, farnz said:

Did you manage to get a response from any Leica representative, Helged?

Pete.

Hi Pete, 

Working on it - will report back! If anyone has any plausible explaination, please feel free to chime in...

Given that SL2-S allows shutter speeds up to 30 min in [S]-mode (involving sensor metering) and in [M]-mode, I tend to think that both sensor metering and sensor heating can be excluded as possible explainations. 

Edited August 8, 2022 by helged

[beewee]

I still think exposure metering is the main limitation. Yes, the sensor can theoretically run for 30 minutes but it would be completely impractical to do a 30min exposure just to estimate what shutter speed is suitable. In order to meter with the sensor itself, the camera would need to use the live view feed from the sensor and then estimate how much light is coming in. If it’s too dark for live view (especially if even enhanced live view is too dark), then the camera will not be able to meter accurately. The camera cannot use 1+s long exposure to meter because the refresh rate for the live view would become more or less useless.

Regarding the S3, since it’s an SLR, it can use an independent sensor for metering and it can run for much longer since it has no impact on live view. I would imagine, if you tried to use live view on the S3, it would probably run into the same limitation as on the SL and TL bodies.

Edited August 8, 2022 by beewee

[helged]

43 minutes ago, beewee said:

I still think exposure metering is the main limitation. Yes, the sensor can theoretically run for 30 minutes but it would be completely impractical to do a 30min exposure just to estimate what shutter speed is suitable. In order to meter with the sensor itself, the camera would need to use the live view feed from the sensor and then estimate how much light is coming in. If it’s too dark for live view (especially if even enhanced live view is too dark), then the camera will not be able to meter accurately. The camera cannot use 1+s long exposure to meter because the refresh rate for the live view would become more or less useless.

Regarding the S3, since it’s an SLR, it can use an independent sensor for metering and it can run for much longer since it has no impact on live view. I would imagine, if you tried to use live view on the S3, it would probably run into the same limitation as on the SL and TL bodies.

Thanks for the comment! A few responses (mostly guesswork, clearly):

• Sensor metering has to activated - in some way or another - in shutter speed priority mode [S]? In this case shutter speeds up to 30 min are ok. 
• If we assume that sensor readings are only possible, or deemed sufficiently accurate, for shutter speeds up to 60 sec (and thus ignoring the point above): Why can't exposure bracketing extend into, say, +1, +2 or +5 stops compared to a shutter speed of 60 sec? This is really strange IMHO.
• No guesswork, since I have the S3 in front of me 😉: 8 min exposure works fine with the camera in live view, and with the camera in aperture priority mode [A]. Actually, this is often how I start out night time time lapse type of photography with S3.

Hopefully the discussed limitation can be mediated in forthcoming FW-updates. At least as long as this doesn't damage the sensor (shouldn't be the case given that 2 out of the 4 modes accept shutter speeds of 30 min).

[Advertisement]

Hallo,

Look here: Leica L-Mount

[helged]

From Leica (& time will tell what future FW-updates will give):

"Thank you very much for your feedback.

I have forwarded your suggestion to our camera developers for consideration for future updates."