Back to forums

Bias noise not calibrating out

Processing Bias Frames Deep Sky Banding
11 replies•348 views
Christopher Gomez avatar
Christopher Gomez avatar
Hi all,
I am dealing with a vertical banding issue on my integrated images. I am getting vertical banding that seems to correspond with a noise pattern seen in the master bias frame. It was not present on one night of shoothing and then present in onother night's shooting. I have attached my master dark, bias, and an integration of about 30 subs for the data that does not have the bias noise as well as the data that has it. You can see it towards the left of the image. This is only present when shooting in 1x1 binning and I dont see it in 2x2 binning data despite the same banding pattern in the 2x2 master bias. I've tried calibrating my lights an all ways posible and even shot new darks and bias but I cant seem to get rid on the banding noise.

Master Bias 1x1: https://drive.google.com/file/d/1OrNMoLK2IeG9T5JLVJD8ZYQOpB_VJr37/view?usp=sharing
Master Dark 1X1: https://drive.google.com/file/d/10oTkcY2QS_4SghaG67YRnofwSQkp9RFk/view?usp=sharing
Master Bias 2x2: https://drive.google.com/file/d/1iTgzUUR7uya_1PywVHbRbxjG9VA_URD5/view?usp=sharing
M51L no noise: https://drive.google.com/file/d/1ciDQXtvREycWu7TMZ6vN5wpSB2uV2w5F/view?usp=sharing
M51L with banding noise: https://drive.google.com/file/d/1ZzrRFqnPiSUTMdbBHBcbENW3Y8BuRNsV/view?usp=sharing

Any help is gratly appreciated.

Thanks!

Chris Gomez
andrea tasselli avatar
andrea tasselli avatar
I'd suggest you integrate the master dark without bias subtraction, and do not subtract the master bias from the light frames during calibration. BTW, the banding is still visible even in the one supposed to be without it. What camera were you using?
Helpful Concise
shenmesaodongxia avatar
shenmesaodongxia avatar
If you dont care about the data accurate, you can use NXT (denoise 0.1,detail 0) to treat all of the sub frame first. It can solve the problem in my mind.
Christopher Gomez avatar
Christopher Gomez avatar
andrea tasselli:
I'd suggest you integrate the master dark without bias subtraction, and do not subtract the master bias from the light frames during calibration. BTW, the banding is still visible even in the one supposed to be without it. What camera were you using?

I'll try to do it that way and see if its any better. the one that I say doesn't have the banding drastically less notisable. I am using a QSI690
Christopher Gomez avatar
Christopher Gomez avatar
If you dont care about the data accurate, you can use NXT (denoise 0.1,detail 0) to treat all of the sub frame first. It can solve the problem in my mind.

I'm trying keep the raw data as original as possible but I may do that if all else fails.
Christopher Gomez avatar
Christopher Gomez avatar
andrea tasselli:
I'd suggest you integrate the master dark without bias subtraction, and do not subtract the master bias from the light frames during calibration. BTW, the banding is still visible even in the one supposed to be without it. What camera were you using?

Doing the calibration like yo suggested didn't do much of a difference. There is still the banding pattern. What I did find is that the banding pattern is not reliably constant as it changes from master to master. Here are two sets of master darks without bias subtraction shot on different days. If you overlay them the banding pattern is different and seems to shift over. 

https://drive.google.com/file/d/1Jr5hccb_FK4iz4X1eILND25IOCdI9PwS/view?usp=sharing

https://drive.google.com/file/d/1uHGWRr2JnB2VQwmtdBOcMw0JgZfAlsjp/view?usp=sharing

In the past, I would shoot my set of calibration files and be good. The noise was constant and repeatable. I don't know what has changed that would make the banding noise shift/change from session to session
Helpful Insightful Respectful
andrea tasselli avatar
andrea tasselli avatar
As you have a CCD camera my previous suggestion was unlikely to work (but not impossible) but I didn't have that info at the time. For a CCD the only time I had issue similar to yours was when the power line was modulating the response of the camera, due to a fault in the power cable. But then the banding was visible across the whole of the field.

One way to ameliorate the issue is to use the CanonBandingReduction script in PI, rotating the image 90deg, see below:

I didn't have time to flatten it again but it can be done.
Helpful Concise
andrea tasselli avatar
andrea tasselli avatar
This is the best I have managed so far:
dkamen avatar
dkamen avatar
If it doesn't show at bin2 and comes and goes at bin1, it could be cabling. Bin2 is 25% the data transfer so 4 times less opportunity for the issue to manifest.
Christopher Gomez avatar
Christopher Gomez avatar
If it doesn't show at bin2 and comes and goes at bin1, it could be cabling. Bin2 is 25% the data transfer so 4 times less opportunity for the issue to manifest.

Cabling in terms of power supply? USB connection? or both? I run everything off of a powered USB hub and the power is supplied to everything off of a RigRunner.  I have the image optimization setting checked on under the camera setting to help minimize noise.
dkamen avatar
dkamen avatar
Well, I think it could be just a faulty USB cable. Number one reason for aggressive banding is problems in readout/transmission and cables are the usual culprits. There are other reasons such as interference among the various electronics, insufficient grounding etc but those tend to manifest consistently.

Have you inspected the bias frames individually? Maybe you will see there are actually fewer bands, moving between the frames.
Well Written Helpful Insightful Concise
Christopher Gomez avatar
Christopher Gomez avatar
Well, I think it could be just a faulty USB cable. Number one reason for aggressive banding is problems in readout/transmission and cables are the usual culprits. There are other reasons such as interference among the various electronics, insufficient grounding etc but those tend to manifest consistently.

Have you inspected the bias frames individually? Maybe you will see there are actually fewer bands, moving between the frames.

I look through them and see if I see anything between frames. Does the length of the USB cord have have any affect? All the USB cords running to th power USB hub are short but then the main line comming off of it to the laptob is a 12 ft USB cord. All cords are relatively new and under a year old as I recently put this rig together