Time Exposures using *istD

I have a *istD with which I am pleased. Firmware has been upgraded to v1.12.
Can anybody advise on a strange problem wrt time exposures?
After doing a long exposure (including from 30 seconds to 5 minutes), the camera then seems to spend almost exactly the same time doing 'internal processing' (shows hourglass icon and image counter flashes) before downloading to the CF chip. Downloading takes only a matter of seconds, as indicated by the LED on the back of the camera.
I get the same effect irrespective of whether I'm running on battery or AC power.
Firstly, is this normal? If, so why?
Secondly, can I do anything to elliminate these impractically long processing times?

Maybe someone with more experience might be in a position to either contradict what I'm about to say, or indeed elaborate upon it.

I get the same with my *ist DL on long exposure shots and by my understanding it's the camera working on noise reduction (something along the lines of 'taking' a blank/black still immediately after completing the exposure for the actual image then checking the black frame pixel by pixel to look for any sensor heat induced noise, then processing corresponding pixels on the actual image to compensate).

You can turn noise reduction for long exposures off in one of the menus and perform noise reduction in a PC based software package at a later stage should you feel it necessary. This would speed up the in camera process.

Here's the definitive answer

When you take a long exposure, the camera takes another image of the same duration and performs black-frame-subtraction. This greatly reduces the amount of noise visible in the image. The camera is taking a black image as a reference for the noise floor of the CCD (which varies according to temperature and conditions too!).

With the *ist-D you can turn this off - look in the custom menu, for the "Noise reduction" option (can't remember the exact term right now). But, if you disable noise reduction the noise level will shoot up for long exposures.

It's required because all semiconductor devices have inherent noise. As you increase the ISO sensitivity you are amplifying that background noise as well as genuine light levels. And when you use longer exposures, the noise also becomes much more significant.

The dark frame subtraction is generally carried out in astronomy. In fact, you can also try taking several images of shorter duration and stacking them together in PS (or the Gimp, or whatever ) Try doing a Google search for astronomical photography techniques

Hope that helps!
Matt

Many thanks Guys. I did wonder about the possibility of a noise reduction process. I'll experiment with turning off in-camera noise reduction.

I was using the 200 ASA setting, so perhaps noise might not be too much of a problem.

My interest in long exposures does relate to astrophotography. I'm taking my first steps in this direction, having just built and commisioned my own computer driven version of a "Barn Door" equatorial mount. As well as a collection of short focal length lenses, I've got a Sigma 100 to 300mm F4 APO EX lens (a cracker!) which with a X2 converter has already provided some fair detail of the surface of the moon.

Photo wise, I'm hedging my bets by trying out the *istD and an old film based MX.

Having used the Linux operating system for some years, I'm becoming familiar with GIMP for image processing. Under Linux, plugins are available to enable GIMP to handle RAW files.

Thanks again guys. I'll report back after trying out the "without noise reduction" option.

I've done some tests re 'noise reduction' versus 'no noise reduction' on the *istD.
Results are as follows:

'no noise reduction' eliminates the 'processing time' that I queried in my earlier posting.

ASA/TIME/noise reduction/Result
200/30sec/off/OK
200/5min/off/very noisy
200/15min/off/bad-useless

200/30sec/on/OK
200/5min/on/slightly noisy
200/15min/on/very noisy

800/30sec/on/OK
800/15min/on/bad/useless

(Sorry about layout here, but site reorganised my previous neat? layout)

During these tests, I noticed that the back of the *istD became warm which must have a significant adverse effect on noise level.

These results suggest a realistic limit of 5mins at 200 ASA for time exposures using the *istD.

It seems that 35mm film still has advantages over digital SLR for astro photography. The main problem with film is reciprocity failure on long exposures. Maybe one day, I shall invest in a cooled CCD astro camera. In the meantime, I'll dust off the trusty old mechanical MX and use something like Kodak ElliteChrome 200.

Again, thanks for your help

Ken Hough

I would leave the NR on, if you're taking time exposures yo're not exactly in a hurry, are you?

I've already come to the conclusion that I'd best leave NR on.

While not being in any particlar hurry, it would be a bit of a nuisance to have to wait around between successive astro shots, especially as weather and possibly other conditions are often changable.

Reluctantly, I now believe that for most long exposures, the optimum solution for the present is to use film. The old Pentax MX is ideally suited to this.

Ken Hough

Ken,
Try this:
* Take shot without NR
* Take second shot without NR
* ...take shot n without NR
* Take same length shot with the lens cap on.

You should be able to combine them in PS... can't remember the exact layer recipe... but you should be able to find it on the web

I reckon if you take several shorter images you will be better off than just a single long one Seem to remember (from somewhere) they do this for film astro-photography and stick all the layers together.

I'll have a look around in Google...
Matt

Interesting article here:
http://www.luminous-landscape.com/tutorials/luminous-heavens.shtml

They recommend taking many <5min shots and combining to create a 3 hour image! Says there's special software for this.
Think the astronomical term is a "stack".

Try these for software:
http://www.astrosurf.org/buil/us/iris/iris.htm - looks very feature rich
http://registax.astronomy.net/
http://www.astrosnap.com/

Hope that helps
Matt

Sorry for delay in responding.

Thanks for info'. I'll follow this up

I do wonder about the possible advantages of taking repeated long exposures. As with all semiconductors, noise is a function of temperature. It is unlikely that the temperature of the camera and hence of the CCD will remain constant and so noise levels in succesive exposures will differ.

Do PENTAX provide a solid state cooler for the *istD as used in astro CCDs?

Carrying on the DIY theme that's been going on today...

Why not wrap the camera body in a modified freezer wine-cooler? That should do the trick

Matt

What? And let the wine get warm? Not sure this idea will be popular.

G

I agree about the wine!

Seriously, I did have a thought about possibly strapping some sort of solid state cooling device to the rear of the camera. That's the part which seems to get warm.

I read a report on the Internet somewhere that someone had arranged to blow air through the internals of their DSLR to cool it and had found it to be of benefit. I'll pass on this one for now.

Having had time to ponder over the suggestion of taking a number of relatively short time exposures, this seems to make sense. The blank exposures for noise reduction are then effectively taken in betweeen the real exposures, so that temperature variations should be less of a problem than would be the case for a single long exposure.

I need to get out and take some pictures and try out all the theory. Just need a couple of clear nights preferably without moonlight.

At last, I've been able to find a bit of clear sky to take some astro time exposures. After the rather disapointing time exposure tests on my *istD, I wasn't too hopeful.

In fact, I was AMAZED. 3 min exposures at around f4 and 200 ASA give fantastic results (used a Pentax K 28mm f2.8 lens). Noise was not a problem. It was a cold night which probably helped. The 3 minute wait for the blank exposures were annoying.

I might get hooked on this.