Lines on DSS stacked image

[jambouk]

DSS seems to be creating an artefact when I'm stacking some data. Vertical dark lines (subtle) appear on the stacked image which I can't see on the individual subs. I've tried stacking with darks and flats, with just darks, and with just flats, and stacking subs on their own with no darks nor flats, but the lines always appear. I've taken a screen shot of a small area and attempted to highlight where the lines are with some yellow lines on a duplicated image above. Also my DSS settings attached.

Any ideas?

Thanks.

James

Presentation1.pdf

[D4N]

Are you using bias frames? Read noise looks like this.

Sent from my iPad using Tapatalk

[jambouk]

Captured with a Canon 6D; I'm not taking bias frames but thought that the camera was doing some of this automatically... Would that read noise be evident on individual subs; or very subtle on individual subs but more prominent after stacking?

james

[Steve789]

I've recently encountered the same issue however I am using bias frames.

I assumed the lines appeared from over stretching the image. The histogram from the stacked image showed very narrow peaks which I assumed resulted from imaging during bright moonlight.

[D4N]

Better check in the camera manual if it is taking a bias frame, a dark, both or neither. If you double correct or fail to correct your noise you will get artefacts.

Personally I would turn off the cameras auto noise reduction and do it manually with master calibration frames.

Sent from my iPad using Tapatalk

[jambouk]

I think the artefact is on my individual subs; dss just amplifies the issue. Adding 79, 240, 800 bias frames made no difference, though there is "noise" in individual bias frames if the curves are pushed hard, there is no noise that i can detect after stacking bias frames.

I don't use noise reduction.

Thanks for the replies though.

I am yet to explore if:

The artefact is related to just long subs on a warmer night (240 seconds at 14 degrees ambient)

Or if related to the eos cls filter clip.

James

[jambouk]

Here is one bias frame (1/4000th second, same ISO as subs, with lens cap on) which I stretched in levels and curve and this noise emerges:

[jambouk]

As I say, if I stack bias frames, I can no longer find the noise when I fiddle with levels and curves.

James

[D4N]

My raw bias frames look like random noise, when I create the master bias the random noise goes away and I'm left with vertical lines that is my read noise.

Sent from my iPad using Tapatalk

[jambouk]

Is it possible to see your individual bias frames and a stacked one?

James

[D4N]

Sure, these are really heavily stretched to make the noise visible, they are both stretched the same amount... just try and ignore the JPEG noise

Single bias frame ISO800 1/8000s

Stack of 105 bias frames ISO800 1/8000s

Sent from my iPad using Tapatalk

[jambouk]

Someone on another forum forwarded me an article from Sky&Telescope on this banding, "horizontal and vertical banding noise" (HVBN), which describes some software the author has written, but which costs $240. Reading the article it seems the image is split into two layers, one with stars and bright bits on, and the dark stuff and banding is left on another; the banding is reduce and then the layers merged again. I'm not very good with processing so I will ask my mate if we can have a go at doing something similar in bog standard photoshop.

The other thing I am thinking about, and which I can't mind an answer to, is if my flats weren't as bright, might they help to reduce the banding. Do you think HVBN is related in some way to exposure length, or it is a largely fixed phenonomen for every exposure irrespective of the exposure length. Currently I try and get my histogram about 50% for flats, but the bright light will make the HVBN artefact difficult to detect; if I dropped the histogram down to 25% or so might the banding be more discernible? I'm not even sure the banding is fixed so stacking "better" flats may just cancel them all out and not help at all....

Tricky.

James

[D4N]

I think super bias in PI is meant to deal with banding but since it hasn't been a problem for me so long as I have sufficient frames I haven't tried it.

TSED70Q, iOptron Smart EQ pro, ASI-120MM, Finepix S5 pro.