Hi, Important issues have delayed my planned photographical activities. As a newcomer, I start now with the intellectual preparations (i.e. mostly reading the sharpcap manual and the forum), so I cannot present any data files yet.
I came across the old rule in programming: do not divide by zero. So, what happens with the formula
(lights minus master dark) / (master flat minus master dark flat) ?
alternatively:
(lights minus master dark) / (master flat minus master bias flat) ?
Could a dead pixel cause a division by zero? Possibly noise would prevent a value exactly at zero after the subtraction operations both in the numerator and the denominator. But a value close to zero in the denominator is not desirable too, it could create a large final pixel value (a “division star“) which is not eliminated, since all corrections (e.g. those for hot pixels) have been done by previous subtractions of darks/bias according to the formulae.
Could the motionless “division star“ hamper the algorithm for star tracking during sharpcap’s live stacking (as for the case of hot pixles, if those have not been corrected)?
A difference in the commanded values for the offset=brightness=black level for the flats and the dark flats would remove the apparition of a “division star” in some cases (not all). However, it would change the maths of the „flat“-correction; therefore this is not desirable.
I would very much like to learn the opinion of the experts on the problem outlined above. And I would like to learn the arguments, if the problem does not exist in reality.
In the sharpcap forum I found only:
dead pixel suppression in
viewtopic.php?t=5678
which deals with another aspect, however.
With thanks in advance, Albrecht
Dead pixels, flats and dark flats
-
admin
- Site Admin
- Posts: 13420
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: Dead pixels, flats and dark flats
Hi Albrecht,
an interesting question
When SharpCap does flat frame correction, it will not apply *any* correction to pixels where the (flat - dark flat) value is zero. That would eliminate issues for any completely dead pixels (that produce the exact same value under all situations) providing the flat correction is being done in SharpCap. I suspect that other processing software may have similar safeguards, but cannot be sure of that.
If the pixel is instead just 'sick' and only produces a very small signal then it's quite possible that you will get a bright pixel in the image after flat frame correction as you describe. However, this is unlikely to have a serious effect on SharpCap's live stacking, since the star detection algorithms are designed to try to avoid detecting a single high value pixel as a star (unless you turn off the 'suppress hot pixels' option in the star detection settings). In any case, the alignment is based on many stars rather than just a few, and tends to use the ones with the highest total brightness, which would mean that hot pixels would not get used unless all the other stars were really faint.
In real life, I don't think this is a big issue - I have rarely seen truly dead pixels on a camera that I have worked with (and that's quite a lot!). Hot pixels which give a spuriously high reading are far, far more common.
Hope this helps,
Robin
an interesting question
When SharpCap does flat frame correction, it will not apply *any* correction to pixels where the (flat - dark flat) value is zero. That would eliminate issues for any completely dead pixels (that produce the exact same value under all situations) providing the flat correction is being done in SharpCap. I suspect that other processing software may have similar safeguards, but cannot be sure of that.
If the pixel is instead just 'sick' and only produces a very small signal then it's quite possible that you will get a bright pixel in the image after flat frame correction as you describe. However, this is unlikely to have a serious effect on SharpCap's live stacking, since the star detection algorithms are designed to try to avoid detecting a single high value pixel as a star (unless you turn off the 'suppress hot pixels' option in the star detection settings). In any case, the alignment is based on many stars rather than just a few, and tends to use the ones with the highest total brightness, which would mean that hot pixels would not get used unless all the other stars were really faint.
In real life, I don't think this is a big issue - I have rarely seen truly dead pixels on a camera that I have worked with (and that's quite a lot!). Hot pixels which give a spuriously high reading are far, far more common.
Hope this helps,
Robin