Banding

Bugs, glitches and unexpected behaviour.
Post Reply
JLP
Posts: 19
Joined: Thu Apr 14, 2016 11:46 am

Banding

Post by JLP »

I am not sure where to post this so Have put it under the heading of Bug Reports I do not think it is a bug in Startools but it is unexpected behavior.

I have been using my Olympus OMD-EM1MKII camera reasonably successfully for a couple of years but since early November 2019 I am getting a colour band appearing in my developed images using Startools. I have had the DSS stacked fits file processed in PI and it shows the same band. I have tried all manner of camera settings and have tried to take long exposure images not of the sky and I cannot reproduce the fault in the image. It only appears when I take images of the sky at night. It is most evident when processing in the colour module of ST and the RGB setting.

I think there is a problem with my camera sensor but I cannot reproduce it taking normal pictures. Anyone come across a similar problem or know how I explain this to Olympus. I am concerned that they will blame the processing or that I am using the camera to take images which it is not designed for. It worked before does not happen with my other Olympus EM5 MKII camera.

I use Olympus Capture to take the RAW images on an iOptron CEM25EC mount using a USB extender link which saves the images to my laptop computer (intel CORE i7). These are then stacked in DSS and processed in ST.

Last nights image was the worst I have seen it was a full moon and poor seeing but I was trying to eliminate potential problems areas.
The band first appeared when I took images of NGC7000.

I have been trying to solve this issue since Nov 2019 and do not think I can see clearly anymore. Any help or explanation would most appreciated.
Attachments
First time I noticed the band on this image of NGC7000
First time I noticed the band on this image of NGC7000
NGC 7000 small EM1MKII.jpg (260.08 KiB) Viewed 4068 times
This was an image of M82 under very poor seeing and processed in ST colour module RGB
This was an image of M82 under very poor seeing and processed in ST colour module RGB
Startools colour RGBs.jpg (327.46 KiB) Viewed 4068 times
happy-kat
Posts: 372
Joined: Sun Feb 01, 2015 11:31 am

Re: Banding

Post by happy-kat »

Hi
As the same can be seen on a processed image in PI it is not a bug with software.
Are you dithering (when you take the images)?

Working backwards has any kit changed since the last good images taken?
Has the mount changed?
Have you started guiding?
Have you changed power?
JLP
Posts: 19
Joined: Thu Apr 14, 2016 11:46 am

Re: Banding

Post by JLP »

Hi
Thanks for reply.
I have not been dithering but did give it a try with no improvements. The dithering was very limited on the night I tried and I had other issues trying to dither. Looking at the width of the banding I assume that I would have to dither very aggressively for the stacker to randomise the fixed pattern noise if that what it is.

I have changed several things but I have had successful results with those changes. I do not guide as I have an iOptron CEM25EC mount. I am only taking 60secs exposures with a 300mm - 420mm lens.

One of the things I have struggled with is seeing improvements when using calibration files so I have tended not to do any and have had some acceptable results, for me. I know that the wisdom is you should take them but when you do and can little to no improvement why would you bother.

The one thing that is different is that the outside has been significantly colder and I am wondering if that cold has some negative affect on my images. I had thought a colder sensor would have been better.

Today I have played around with flat files and notice that if I open a master flat in ST Wipe it I can see the banding. Although I did not take flat files at the time of the M82 I have since done some and just stacked them with the light files and it is very much improved (see below). The original M82 images showed some significant finger prints on the lens which I had clean off before taking a set of 20 flats so those smudges are still there.

I did the same exercise with my EM5MKII which uses a different sensor and there is much finer banding but in the horizontal direction. Maybe they are there in my images but I just cannot see them.

So at the moment it seems that Flat frames are more important than Dark frames for my setup.

The next time I get a clear night I will take images with Noise reduction on a set of flats and dither and see how that turns out.

I guess that what this shows is just how good ST is at getting every bit of a change in photons out of an image.
Attachments
This is the same Lights but stacked with 20 flats taken today.
This is the same Lights but stacked with 20 flats taken today.
Startools colour RGB flats.jpg (357.87 KiB) Viewed 4053 times
User avatar
admin
Site Admin
Posts: 3364
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: Banding

Post by admin »

Let us know if you ever figure out what is causing that strange 1/4th horizontal division-like pattern.
I have never seen anything quite like it... :think:
Ivo Jager
StarTools creator and astronomy enthusiast
JLP
Posts: 19
Joined: Thu Apr 14, 2016 11:46 am

Re: Banding

Post by JLP »

Hi Ivo
I think I have found the cause and it seems to be a "feature" of the sensor. If I take 20 odd flat frames and create a master frame then process that master frame in ST I get the 1/4th horizontal division-like pattern. so by adding (subtracting?) flat frames during stacking the bands do not appear. I think it must be something to do with the camera reading (fixed pattern read noise) of photosites and as the bayer matrix is RGGB it may be realted to that but I am not knowledgeable on things like this.

I did not see the effect on my other Olympus camera an OMD-EM5MKII but I carried out the same test with flats but this time I got a much finer horizontal bands of very faint lines. The sensor are thought to be from different a Manufacturer.

I had a mega imaging session last night (over 6hours). I dithered using NINA direct guider (I do not guide as have an encoder mount) used Noise reduction 60sec (so 2mins per frame) exposures ISO 800 with the same lens (OMD-EM1MKII 300mm F4) and camera as the example above stacked the 180 images with 20 Flats and 20 Dark Flats in DSS 4.2.3 and processed in ST 1.6 and got a much better image with no signs of the pattern. Not fussed around with ST settings just a quick process using defaults and 25% binning.
Attachments
NGC 1893 & IC 405 Flaming Star Nebula
NGC 1893 & IC 405 Flaming Star Nebula
NGC 1893 & IC405.jpg (479.21 KiB) Viewed 4002 times
Post Reply