Reciprocity failure.

[Dusty]

I was curious to know if digital sensors suffer the equivalent of film's reciprocity failure? If I take a long time exposure on digital, will the exposure time be linear, or will holding the shutter open for long periods result in weird effects?

Dusty

[Javelin]

I doubt this affects digital. but noise might make you shoot multiple shorter exposures and stack them instead of going for one very long exposure anyway

[David Kilpatrick]

It's linear, but the sensor may heat up (depending on design and heat sink efficiency, ambient temperature) and produce more noise than the exposure itself produces detail. Phase One sensors are rated for up to one hour exposure without degradation. Canon CMOS (even the original 300D) can generally be exposed for up to 20 minutes with good results. Sony original CCD (D7D etc) is OK up to 30 seconds, go longer than a minute and special measures need to be taken. I don't know how long the EXMOR CMOS can take but I would guess it's somewhere between Canon's excellent long exposure performance, and Sony CCD performance. I would try exposures like 2, 5 even 10 minutes without feeling I was wasting time. And no exposure compensation factor should be needed, just a straight multiplier, reciprocity intact.

David

[01af]

In the Sony DSLR Talk forum at dpreview.com, user 'Nordstjernen' from Norway keeps showing spectacular deep-sky astro shots taken with the Sony A900 on an 100-mm refractor telescope (approx. equiv. to an 800 mm 1:8 telephoto lens). He typically shoots series of exposures approx. 1 - 10 minutes each at ISO 1600/33°, then stacking them.

-- Olaf

[01af]

Another example of Nordstjernen's work.

Another here, taken with SAL 85 mm 1:1.4 ZA at f/4 (20 exposures 2 mins each, ISO setting not specified).

-- Olaf

[David Kilpatrick]

Can someone please find and send me Magnar's (Nordstjernen's) email. I have one for him, but it is never answered. I want to send him a large infra-red optimised lens, as a gift, for him to experiment with and I've emailed twice about it with no response.

David

[PhotoTraveler]

As mentioned it's linear.

I really wish there was a reprocity failure mode you could select. Where you could adjust it to either select one of a few pre-defined curves, or just select a rate.

While linear can be nice, it's also a pain. If you want that 4 hour star trail shot, you need to close the heck down the aperture and be at Base ISO and even then you will be over exposing quickly. Plus the trails will be hard stop.

If I could have it start at ISO 400, and degrade down reducing the amplification down to ISO 100 and then go lower (just as the camera makes Lo 80 and such) and keep going lower, it would be very nice at times.

The linear works well for some stuff, and you can calc the exposure with a high ISO wide open aperture test shots and then do some math, so that is nice. But you are time limited still. ISO 100 at F11 will over expose anything before too long.

[Dusty]

Well, if you want longer exposures, there's always neutral density filters. I cornered the market on them in '83 when we were having a full solar eclipse. Slow film, lots of NDs stacked up and you still end up with a couple of shots of a black disk over a white disk!

Dusty

[jcoffin]

Well, if you want longer exposures, there's always neutral density filters. I cornered the market on them in '83 when we were having a full solar eclipse. Slow film, lots of NDs stacked up and you still end up with a couple of shots of a black disk over a white disk!

Dusty

Another way to cut out a lot of light is to cross a pair of polarizers. In theory, with perfect polarizers, these would cut out absolutely all light. If memory serves, with real polarizers I've gotten a light decrease of around 15 or 16 stops. That gave something like a 5 minute exposure in broad daylight at ISO 100. This is long enough that you can take pictures of quite a few popular tourist attractions without the other tourists showing up at all -- very few people stay still for enough of a five minute exposure to produce an image in any one part of the picture. Of course, it only works in places that allow tripods. Oh, one other warning: when I first tried this with some fairly cheap polarizers, I got some strong color casts that varied with the angle at which the filters were crossed. Better polarizers seemed to eliminate that problem though...