Pentax K5 - Dark frame subtraction - possible to turn it off?

Nass said:

Jose, it looks strongly from one person who has tried k-5 at Photokina with B at > 30s that the answer is yes, it can be switched off. He was able to take another within 1 sec. Hopefully Falk will confirm on Sunday. Happy days.

Click to expand...

• Sir Winston Churchill

• Albert Einstein

BobORama said:

Hi,

I write my own stacking software to process much dirtier source material from cheapo CCD based cameras. I have had good results using a "brightest pixel" stacking method and applying a single ( or averaged ) DFS reference frames to remove dead / stuck pixels. This method works well because you can stack hundreds of shorter exposures - each with vastly less noise - and get continuous star trails. SLRs, and other cameras with short shot to shot cycle times are much better candidates because the gap between frames is very short, and you can get a > 99.6% duty cycle using 30 second exposures, rather than the 50% duty cycle of in-camera DFS. So the very short gaps between shots where the trails dovetail is not noticeable, and the trails appear continuous in the stacked image.

The only other "secret" to doing DFS in post processing is to use continuous shooting, and let the camera shoot for about 20 minutes prior to your actual shoot. This lets the sensor get up to temperature. Then you run a series of dark frames ( pop on lens cap ). Then your main series. Then more dark frames at the end. Since the sensor will have achieved a steady state temperature - the dark frames will properly correct the exposures in your series.

Do not underestimate the importance of letting the sensor get to a steady state temperature. This is a test run I did against a consumer P&S which shows almost 2 stop increase in noise as the camera heats up. If your dark frames were sampled when cold, they would under correct - severely. http://www.vimeo.com/817772 So if you sample the dark frame after it warms up, it will provide better correction. Having the DFS reference frame shot when the sensor is a different temperature is one of the reasons people are unhappy with DFS done in post processing.

Click to expand...

José Ramos said:

I know that DFS is unnecessary because I can turn it off in my K10d, do 30-240 second exposures, and then just as I open the RAW files in Adobe Camera Raw, it automatically eliminates all the hot pixels.

Click to expand...

BobORama said:

Hi,

I write my own stacking software to process much dirtier source material from cheapo CCD based cameras. I have had good results using a "brightest pixel" stacking method and applying a single ( or averaged ) DFS reference frames to remove dead / stuck pixels. This method works well because you can stack hundreds of shorter exposures - each with vastly less noise - and get continuous star trails. SLRs, and other cameras with short shot to shot cycle times are much better candidates because the gap between frames is very short, and you can get a > 99.6% duty cycle using 30 second exposures, rather than the 50% duty cycle of in-camera DFS. So the very short gaps between shots where the trails dovetail is not noticeable, and the trails appear continuous in the stacked image.

The only other "secret" to doing DFS in post processing is to use continuous shooting, and let the camera shoot for about 20 minutes prior to your actual shoot. This lets the sensor get up to temperature. Then you run a series of dark frames ( pop on lens cap ). Then your main series. Then more dark frames at the end. Since the sensor will have achieved a steady state temperature - the dark frames will properly correct the exposures in your series.

Do not underestimate the importance of letting the sensor get to a steady state temperature. This is a test run I did against a consumer P&S which shows almost 2 stop increase in noise as the camera heats up. If your dark frames were sampled when cold, they would under correct - severely. http://www.vimeo.com/817772 So if you sample the dark frame after it warms up, it will provide better correction. Having the DFS reference frame shot when the sensor is a different temperature is one of the reasons people are unhappy with DFS done in post processing.

-- Bob
http://www.vimeo.com/boborama/videos
http://www.youtube.com/profile_videos?user=bob0rama
http://public.fotki.com/boborama/

Click to expand...

BobORama said:

Just keep in mind that DFS is not the same thing as hot pixel removal. DFS attempts to remove the baseline noise from all photosites ( pixels if using JPG ) on the sensor. Whereas hot / dead pixel repair focuses on, what should be, a very small number of badly functioning photosites.

Click to expand...

José Ramos said:

I did some side-by-side comparisons between 30-180 second long real world long exposures with DFS on and off, and couldn't find any difference between them, besides the hot pixel issues.

Click to expand...