27
\$\begingroup\$

For example, my Canon 550D will only go as far as 30 sec (and then of course the bulb mode).

For measuring exposure, I think all the camera needs is a timer, and I have no doubt they would have to do no extra work to remove this upper limit if it were only for the timer.

What am I missing here? Why do cameras have this upper limit?

\$\endgroup\$
1
  • 9
    \$\begingroup\$ .... There isn't a limit. The internal meter/timer only goes to 30 seconds, but you could take hours long exposures if you wanted, by putting the camera in bulb mode and holding the shutter button until your battery dies. \$\endgroup\$
    – Fake Name
    May 14, 2011 at 8:46

4 Answers 4

34
\$\begingroup\$

I think there a several reasons that together make sense to limit the shutter speed at about 30 seconds.

  • At exposures requiring more than 30 seconds, light is so weak your TTL meter will not be able to measure it correctly.
  • 30 seconds is already longer than you'd ever need for any "normal" night scene.
  • In a digital camera, sensor heat starts to build up and thermal noise becomes noticeable.
  • With film, calculated exposure and actually needed exposure are significantly different thanks to reciprocity failure.
  • Bulb mode helps to sell shutter cables and intervalometers to enthusiasts. Accessories are usually more profitable for manufacturers than cameras themselves - the photographer is already "locked" into the system.

30 seconds is a very common, but not an universal limit. For example, Panasonic Lumix DMC-GF2 limits at 60 seconds. Phase One 645DF limits at 60 minutes.

\$\endgroup\$
6
  • 2
    \$\begingroup\$ Point 1 is probably a strong reason for the limit actually. \$\endgroup\$
    – JamWheel
    May 14, 2011 at 8:01
  • 4
    \$\begingroup\$ Not just sensor heat — heat generated by other electronics starts to have a measurable impact (often color blobs on one side of the frame). And also, light leaking in through the viewfinder becomes a big deal. My camera comes with a cover for that, but I have no idea where it is now. (Since I don't usually do this kind of shot.) \$\endgroup\$
    – mattdm
    May 14, 2011 at 13:24
  • \$\begingroup\$ @Imre I think you mean "reciprocity failure", not "effect". \$\endgroup\$
    – whuber
    May 14, 2011 at 18:41
  • \$\begingroup\$ @Imre: Just read your very helpful answer to photo.stackexchange.com/questions/12045/… about calculating long exposure times. How does that technique correlate to your first point here? If the camera can correctly meter at a high ISO in low light, why can't it do the math for me at a lower ISO? \$\endgroup\$
    – Hank
    May 15, 2011 at 18:34
  • 1
    \$\begingroup\$ @Imre: I see; you were suggesting to use the high ISO only to shorten the test shot exposure time, not to use the exposure meter. \$\endgroup\$
    – Hank
    May 15, 2011 at 23:34
11
\$\begingroup\$

Noise and heat are the real issues. The longer the exposure, the more heat the sensor generates and the higher the image noise gets.

These things obviously kick in differently depending on ambient temperature shooting ISO, they have to cut it off at a reasonable point. Canon, Nikon, Pentax, Fuji stop at 30s. Panasonic and Olympus stop at 30s. Sony usually stops at 30s, sometimes 60s.

The cameras which tried to do something more clever are rare because it becomes confusing. Some models have a 30s limit for ISO 100-400 and then 15s for ISO 800, 8s, for ISO 1600, etc (there are plenty of variants on the theme).

One thing that is accepted is that when shooting in Aperture priority or Program mode, there can be a shorter limit (1-60s depending on the model) due to metering sensitivity. In Manual obviously one can set a time limit without this restrictions or go to Bulb mode.

Bulb mode also has limits. Only they are poorly documented. Most DSLRs allow between 4 and 30 minutes (some reach hours) of bulb exposure, presumably to avoid sensor overheating. Strangely, several Olympus models allow you to manually set the bulb limit between 4 and 30 minutes. Those models never allow more.

For digital cameras which do not forcibly use dark-frame subtraction, one can use exposure stacking to simulate longer bulb exposures.

\$\endgroup\$
7
\$\begingroup\$

Once upon a time, the reasons were quite practical. Back when shutters were timed mechanically, the long shutter speeds involved running a mechanical timer -- i.e., the first shutter curtain would open, then the time would run, then the second shutter curtain would close. To let that timer run for a longer interval, you needed a larger mainspring that took more turns to wind. As such, the maximum shutter speed affected the overall size, cost, and usability of the camera.

Modern shutters are timed electronically, of course, but the longest shutter speed supported hasn't really been an issue/selling point for most, so it's remained pretty much the same for quite a long time.

From a practical perspective, 30s is also long enough that exposing by hand is almost certainly plenty accurate. For most people, doing a 1 second exposure by hand would be pretty challenging. Reactions faster than 1/10th of a second are quite rare, and noticeably slower than that are fairly common.

When you're talking about an exposure over 30 seconds, however, being off by even a second or two no longer makes much real difference. Just for example, let's assume you wanted something that was 1/3rd of a stop longer than 30 seconds -- about the smallest increment most cameras directly support anyway. Since one more stop would be a minute, 1/3rd more would be about 40 seconds. As long as you hit somewhere between 37 and 43 seconds (or so) your accuracy is probably close to as good as the camera will do anyway (and possibly more accurate than its fastest shutter speeds).

If you were starting with a base exposure of, say, one second, you'd need reactions on the order of an Olympic athlete's to even have a hope of getting that kind of accuracy and/or repeatability. Granted, most people could probably do all right at least than 30 seconds. They could probably remove the 30 second and 15 second spots without any real loss, and 8 seconds probably wouldn't be a big problem either. Much below that, however, would probably be getting a lot more difficult to justify.

\$\endgroup\$
2
  • \$\begingroup\$ Love knowing the vintage mechanism, springs and gears etc! Not because it is extremely useful to know, but because I just find it fascinating how people "program" things with hardware before silicon chips and computers etc. I am a programmer ;) \$\endgroup\$
    – Gapton
    Dec 17, 2012 at 2:33
  • \$\begingroup\$ I suspected there was an historical reason, but my digitally oriented mind was pointing at a limitation of the binary representation of the number (shutter interval value) in the days of the early devices with smaller bit-words. But now that I read about the mechanical implementation, that makes a lot more sense. \$\endgroup\$
    – Jahaziel
    Oct 3, 2017 at 21:03
0
\$\begingroup\$

I think the answer probably is along the lines of "you have to draw the line somewhere". There will probably be a finite amount of memory in the hardware for settings to be programmed into and whilst you could add lots of different settings, where would you stop? Would you have a shutter speed increment until "forever"? :)

I suppose once you get to a point, it is probably easier to let the user determine how long they want to expose for and allows finer control than the increments allow once you get down to that end of the shutter speed scale.

\$\endgroup\$
3
  • \$\begingroup\$ 30s is too less! \$\endgroup\$
    – Lazer
    May 14, 2011 at 10:24
  • \$\begingroup\$ Yes but if the next increment was 60 seconds you would be on here complaining about 60 seconds being too long :) \$\endgroup\$
    – JamWheel
    May 14, 2011 at 10:29
  • 6
    \$\begingroup\$ I'm pretty sure the shutter speeds could be generated by a function, and really really big numbers only actually take a few bytes, giving a sequence as long as any human could reasonably select from in finite time. So I don't think that's really it. \$\endgroup\$
    – mattdm
    May 14, 2011 at 13:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.