M 240 long exposure Bulb modus

[mjh]

Advertisement (gone after registration)

With incrasing exposure time the photon noise decreases, or to be correct, the signal to noise ratio is getting better, that's why we need long exposure time.

Ahem, no. The shot noise depends on the number of photons. You need a certain number of photons for an optimal exposure and you choose an exposure time so the required number of photons hit the sensor. In the end the number of photons captured is the same for any optimal exposure. The number of electrons that fit within a pixel is limited, so there are only so many photons a pixel can capture before it overflows. Long exposure times don’t remove this upper limit.

 

Further, the noise does not increase with ISO setting, it just becomes more visible.

Increasing the ISO setting requires an increase in amplification, which amplifies noise. Or put differently: If you increase the ISO setting the camera captures less photons so the signal is weaker. But the noise stays the same and the signal-to-noise ratio suffers.

 

Then coming back to my original question is there a work around, for example by using the exposure correction?

Sure there is. As I had mentioned when this issue came up first, a few weeks ago, one could underexpose and correct the exposure in raw conversion.

 

How likely is that Leica will offer a new firmware giving the customer more freedom?

I have no idea.

[tgm]

Ahem, no. The shot noise depends on the number of photons. You need a certain number of photons for an optimal exposure and you choose an exposure time so the required number of photons hit the sensor. In the end the number of photons captured is the same for any optimal exposure. The number of electrons that fit within a pixel is limited, so there are only so many photons a pixel can capture before it overflows. Long exposure times don’t remove this upper limit.

 

 

Increasing the ISO setting requires an increase in amplification, which amplifies noise. Or put differently: If you increase the ISO setting the camera captures less photons so the signal is weaker. But the noise stays the same and the signal-to-noise ratio suffers.

 

 

Sure there is. As I had mentioned when this issue came up first, a few weeks ago, one could underexpose and correct the exposure in raw conversion.

 

 

I have no idea.

 

Sorry, I can not agree with your first two satements, they are irrelevant for all most all long exposures. The number of photons per second is usually fixed, for example if you take a picture of the night sky, you can not change the photon flux, usually you work with the aperture wide open. That's why you need a long exposure time, to get a acceptable signal to noise.

 

So the question remains, is there another technical reason why the M 240 can not offer a real bulb mode?

 

Anyhow, the work around using low ISO settings and correcting in the raw converter, I tried that last week (not with M, I don't own one), but it does not look promising.

 

I would very much appreciate if Leica offers a real bulb modus with an updated firmware.

 

Let me put it this way, one of the strength of full frame sensor is their performance at low light levels offering low noise. The new M sensors seems a real improvement in terms of noise, so why restricting the bulb modus?

 

Thomas

[mjh]

Sorry, I can not agree with your first two satements, they are irrelevant for all most all long exposures. The number of photons per second is usually fixed, for example if you take a picture of the night sky, you can not change the photon flux, usually you work with the aperture wide open. That's why you need a long exposure time, to get a acceptable signal to noise.

Photon flux doesn’t matter. When the rate of photons hitting the sensor decreases, you increase the exposure time to compensate. Whenever the picture is correctly exposed, the total number of photons captured is the same. That is what ‘correct exposure’ means.

 

Only when you choose a higher ISO setting, you decrease the number of photons required for a correct exposure. When you switch from ISO 200 to 400 you must take care that the camera captures only half as many photons – otherwise the increased amplification would cause an overflow, i.e. clipping. The result is a weakened signal with a mostly unchanged level of noise, resulting in a compromised signal-to-noise ratio. While the weak signal gets amplified, so does the noise.

 

Let me put it this way, one of the strength of full frame sensor is their performance at low light levels offering low noise. The new M sensors seems a real improvement in terms of noise, so why restricting the bulb modus?

The Max sensor has a very low read-out noise, that much is true. But read-out noise (and shot noise, for that matter) doesn’t depend on the exposure time anyway. It is dark current noise that is the issue here, and dark current noise does increase with the exposure time (and temperature).

[xrogers]

This limitation disturbs me. In the cases I've encountered, digital cameras that limit long exposure do it because the noise gets too high to reasonably correct, even with a dark frame. Leica engineers aren't stupid, this isn't a mistake. Long exposures must have some real reason for being limited.

 

While it is possible the limit will be increased, the M9 was limited to a rather short 240 seconds, and hasn't been increased since it was introduced. I was hoping that a camera capable of video would have heat sinking that would make indefinite long exposure possible. Unfortunately, I've determined that if I buy an M, I'd do it with no expectation of ever using it for serious long exposure work.

 

With any camera, you really should only bank on the features that are already there.

[tgm]

Photon flux doesn’t matter. When the rate of photons hitting the sensor decreases, you increase the exposure time to compensate. Whenever the picture is correctly exposed, the total number of photons captured is the same. That is what ‘correct exposure’ means.

 

Only when you choose a higher ISO setting, you decrease the number of photons required for a correct exposure. When you switch from ISO 200 to 400 you must take care that the camera captures only half as many photons – otherwise the increased amplification would cause an overflow, i.e. clipping. The result is a weakened signal with a mostly unchanged level of noise, resulting in a compromised signal-to-noise ratio. While the weak signal gets amplified, so does the noise.

 

 

The Max sensor has a very low read-out noise, that much is true. But read-out noise (and shot noise, for that matter) doesn’t depend on the exposure time anyway. It is dark current noise that is the issue here, and dark current noise does increase with the exposure time (and temperature).

 

 

Actually, I am not interested in an artifical discussion with you. Do you think there is a need trying to teach me in light detection? Your arguments presented above should certainly hold also for analog detectors like a film. So why does the Leica M7 offers a bulb mode without restriction?

 

Just to finish that, photon flux off course matters, in order to compensate for low flux you need long exposure time!

 

But now you come up with some bad news, the dark current of the new M sensor is the real problem (by the way, I wrote above there are probably different sources of noise, your answer "Ahem, no") So how does the noise due to dark current compares with that other full frame CMOS sensors or that of the M9?

 

Thomas

[jaapv]

The M9 has no CMos sensor but a CCD and there the problem is heat.

[theno23]

Advertisement (gone after registration)

Yes, it really is that simple. At 60s you can see the shadows are starting to get noisy.

 

The longer the sensor is exposing for, the more heat it generates, and the noisier it gets.

 

 

Leaf had the same issue with thier MF backs until recently. The made a specific version that could do long exposures without being to noisy, but they had to put a lot of engineering effort into it. There was an article about it on LuLa.

[mjh]

I was hoping that a camera capable of video would have heat sinking that would make indefinite long exposure possible. Unfortunately, I've determined that if I buy an M, I'd do it with no expectation of ever using it for serious long exposure work.

Long exposure times and video (and live view) are completely different issues. With video and live view the sensor needs to be read out 24 to 30 times per second, generating heat. In a CCD it is the read out process itself that is turning electricity into heat whereas with a CMOS sensor with integrated ADCs it is probably the ADCs generating most of the heat. The Max sensor reduces the heat by combining the signals from several columns of pixels so only a fraction of the 6000 ADCs need to be used during video capture.

 

During a long exposure the sensor is basically doing nothing, other than letting incident photons create electron-hole-pairs and storing the electrons. There is just one read-out event at the end of the exposure, just as it is with a short exposure. However there is the dark current that is adding more noise as time passes, and the longer the exposure time the more dark current noise there is. This isn’t an issue for video and live view as the exposure times are relatively short.

[jaapv]

Maybe, but in astrophotography with exceedingly long exposure times CCD sensors need to be cooled by liquid Nitrogen.

[mjh]

Your arguments presented above should certainly hold also for analog detectors like a film. So why does the Leica M7 offers a bulb mode without restriction?

Silver-halide film and sensors are facing quite different issues. With a sensor there is dark current noise and hot pixels to be dealt with whereas silver-halide film suffers from the Schwarzschild effect that has to be taken into account. Film and sensors aren’t really comparable in this regard.

 

Just to finish that, photon flux off course matters, in order to compensate for low flux you need long exposure time!

I’ve tried to explain this and I don’t think I can explain it any better than I already did.

 

But now you come up with some bad news, the dark current of the new M sensor is the real problem (by the way, I wrote above there are probably different sources of noise, your answer "Ahem, no")

Bad news? I had told you that dark current noise is the issue about two months ago: http://www.l-camera-forum.com/leica-forum/leica-m9-forum/274473-new-m-questions.html#post2334740. It seems we have come full circle.

 

So how does the noise due to dark current compares with that other full frame CMOS sensors or that of the M9?

As far as I know, the dark current hasn’t been specified by CMOSIS.

[tgm]

Maybe, but in astrophotography with exceedingly long exposure times CCD sensors need to be cooled by liquid Nitrogen.

 

Actually, I wouldn't call 8 sec ( for ISO 1600) exceeingly long.

 

Coming back to my previous question, how does that compare to other cameras? Is Leica only extremely caucious or is the heat problem much worse than that of other CMOS sensors?

 

Thomas

[jaapv]

Ermm- CCDs in astrophotography are exposed for times in the order of 30 minutes and are quite large - I don't think anybody would be daft enough to use a small sensored rangefinder for that purpose.

[tgm]

Bad news? I had told you that dark current noise is the issue about two months ago: http://www.l-camera-forum.com/leica-forum/leica-m9-forum/274473-new-m-questions.html#post2334740. It seems we have come full circle.

 

 

As far as I know, the dark current hasn’t been specified by CMOSIS.

 

Yes, that's true, we are coming to a full cicle, and it the reason why I was surprised that you are trying to explain, that the noise related to photon statistics is the reason why Leica restricts the bulb mode. Please, read again what you wrote today, I think you will agree with me that your comments are at best very missleading

 

I started the new thead after I made some tests trying to check the work around with underexposure und corrections in the raw converter, the results are not so promising. May be the amplification is not linear, that's why I asked if there are other work arounds using for example with exposure correction.

 

The key question, how does the M compares to the other CMOS cameras, if CMOSIS have not given any numbers are there any practical comparisons?

 

Thomas

[PetersDreams]

I tried what jaapv suggested on my M. The shutter stays open as long as I want...

 

On "S" it closes the shutter after 60 sec

On self timer (set to 2sec) ... it does not ... :-)

 

In both cases it then goes and applies the black frame subtraction (60sec). Where in the "hidden" "T" mode it does something before that which takes approx. the time difference between the time the shutter was open and 60 sec. The display stays black, but the red light is on.

 

Cheers

Peter

 

I've been thinking - could it be that the M has a hidden "T" setting just like the M8 and M9?

 

Could any M owner try the following -

 

Set self-timer to "0" (not really needed but easier)

Set the camera to Selftimer and "B"

Hit the shutter release.

The shutter will stay open until the release is operated a second time without time limit - at least not one I've been able to determine.

[thighslapper]

.... but what are the pictures like ..... ????

[stickan1]

Great. I have used this function on my M9 often. I got my M a couple of days ago so I will try. It is not really hidden as it is described in the manual, page 188, although somewhat cryptic!

[MarkP]

I do a reasonable amount of long-exposure night photography (see entires in Barnack's Monthly by myself and others). I used to think that 240sec on the M9 was limiting but it's been OK for almost everything.

 

For such photos I use my M9 on a tripod (obviously:rolleyes:) at ISO 160 (to optimise IQ), and usually set the lens to a middling aperture of 5.6 or so. Exposures are often 90 to 240 sec. However, the M9 files are very robust when underexposed (as mentioned above). I can always open up the lens 2-3 stops which would probably keep the exposure time under 60 sec but will it will limit creativity and DOF (unless I'm using a wide angle). For example, if I'm photographing water in low light and want to smooth it out then I would need more than 60 seconds! It will be hopeless for light painting!

 

I'm really looking forward to getting my M240 but I'm glad I'm also keeping the M9. A bit bizarre that I will have to go back to the M9 for my low-light photography . For whatever the technical reasons it is very disappointing that the maximum exposure time for the M is only 60 sec.

 

I'd be very interested to know whether on the M240 the 'hidden T setting' will bypass this but perhaps the power is cut to the sensor after 60 sec regardless.

Could someone please let us know?

 

Also, regarding the 'hidden T setting on the M9', does anyone know whether the sensor is active beyond 240 sec or does it shut down regardless.

[Delcredere]

I do a reasonable amount of long-exposure night photography (see entires in Barnack's Monthly by myself and others). I used to think that 240sec on the M9 was limiting but it's been OK for almost everything.

 

For such photos I use my M9 on a tripod (obviously:rolleyes:) at ISO 160 (to optimise IQ), and usually set the lens to a middling aperture of 5.6 or so. Exposures are often 90 to 240 sec. However, the M9 files are very robust when underexposed (as mentioned above). I can always open up the lens 2-3 stops which would probably keep the exposure time under 60 sec but will it will limit creativity and DOF (unless I'm using a wide angle). For example, if I'm photographing water in low light and want to smooth it out then I would need more than 60 seconds! It will be hopeless for light painting!

 

I'm really looking forward to getting my M240 but I'm glad I'm also keeping the M9. A bit bizarre that I will have to go back to the M9 for my low-light photography . For whatever the technical reasons it is very disappointing that the maximum exposure time for the M is only 60 sec.

 

I'd be very interested to know whether on the M240 the 'hidden T setting' will bypass this but perhaps the power is cut to the sensor after 60 sec regardless.

Could someone please let us know?

 

Also, regarding the 'hidden T setting on the M9', does anyone know whether the sensor is active beyond 240 sec or does it shut down regardless.

 

Please could you explain why you use a middling aperture rather than shooting as wide open as the lens (Leica I am assuming) permits? Many thanks.

[MarkP]

Primarily for depth of field, but also to get further improvement in IQ. Most fast lenses still don't offer optimal IQ at maximum aperture.

[Advertisement]

Hallo,

Look here: Leica M

[jaapv]

I do a reasonable amount of long-exposure night photography (see entires in Barnack's Monthly by myself and others). I used to think that 240sec on the M9 was limiting but it's been OK for almost everything.

 

For such photos I use my M9 on a tripod (obviously:rolleyes:) at ISO 160 (to optimise IQ), and usually set the lens to a middling aperture of 5.6 or so. Exposures are often 90 to 240 sec. However, the M9 files are very robust when underexposed (as mentioned above). I can always open up the lens 2-3 stops which would probably keep the exposure time under 60 sec but will it will limit creativity and DOF (unless I'm using a wide angle). For example, if I'm photographing water in low light and want to smooth it out then I would need more than 60 seconds! It will be hopeless for light painting!

 

I'm really looking forward to getting my M240 but I'm glad I'm also keeping the M9. A bit bizarre that I will have to go back to the M9 for my low-light photography . For whatever the technical reasons it is very disappointing that the maximum exposure time for the M is only 60 sec.

 

I'd be very interested to know whether on the M240 the 'hidden T setting' will bypass this but perhaps the power is cut to the sensor after 60 sec regardless.

Could someone please let us know?

 

Also, regarding the 'hidden T setting on the M9', does anyone know whether the sensor is active beyond 240 sec or does it shut down regardless.

There is no hidden T setting....