Vertical and Horizontal banding - can't figure it out

[cmassa]

I have many hours of 600s exposures in a wide field around the Bubble nebula. I was having trouble with star cores so I decided to take some 60s exposures and try to do an HDR composition. The stack of 600s exposures looks good (except for the star cores), no lines or artifacts. But the 60s stack (60 x 60s, iso800) looks terrible. I've attached an example which has had an ABE and then a "super" STF to show it better but it is obvious after just the ABE and STF applied. The bias and flats were the same as used for the 600s stack. I used a previous 60 sec master dark. I even tried using the 600s master dark that I used for the 600s stack and it came out the same. If I look very closely at the RAW images I can faintly see it so it must be something in the camera. Camera used is a Canon T2i DSLR.

Any idea what causes it? Is it fixable? I was just using the BPP script with default settings for integration. If I manually do the integration is there a recommended setting which might get rid of this? The Canon debanding script helps a bit but doesn't get rid of it.  I was wondering if there is something else wrong that I need to fix or change.

Thanks,
Chris

[Alejandro Tombolini]

Hi Chris
For HDR purpose to improve stars cores do not worry too much. HDR will use from the low exposure image only the cores of the stars.
I saw those lines in my camera shooting to high iso and low time exposure. Any way in one minute you should be able to calibrate them. Try using a MasterDark made at de same temperature, time and ISO than your lights.
Saludos,
Alejandro.

[cmassa]

The master dark was made at the same iso, temp, duration as the lights. It was just made a few weeks ago but I can try to make a new one. But since I can see the lines in the raw frames, the darks shouldn't be the problem, right?

That's good that it should not affect the HDR process but what if was trying to image a bright object like a planetary nebula which are typically done with short exposures. Is there a way to prevent this?

Thanks,
Chris

[georg.viehoever]

I never found a way to get rid of the banding with calibration. CanonBandingScript just reduces the issue - but frankly it does not remove the root cause.
Observational techniques:
- Dithering your raws  (i.e. moving around your mount a bit while capturing) does help a bit, because it makes sure that values from different rows of the CCD image go into each final pixel, thus averaging out the banding
- Shooting at low temperatures or cooling your camera seem to help.
Georg

[cmassa]

Thanks, I was wondering if it was because of hot temps. Luckily our weather is starting to cool off. Most of my sensor temps through the summer have been 30-40C all night. Just starting to get sensor temps in the high 20's now. We'll see how it goes as winter comes on.

Chris

[MikeOates]

Chris,

You will go a long way to reducing those bands with a large set of bias frames, I suggest at least 150 to make a master bias. (I don't use a DSLR much now, but increasing the bias subs really made a huge difference). Then use the canon de-banding script to reduce the remainder. Remember to rotate the image 90 degrees before using the script and re-rotate back after. I forget which way it works, so if it does not remove them, try the 90 degree rotation.

Note I say "reduce", there will always be some trace of them especially if you do a lot of stretching.

Mike

[georg.viehoever]

Instead of using a huge number of bias frames, I get nice results with the SuperBias process. Unfortunately, it is not yet integrated into the BatchPreprocessing script.

It dont think that using many bias, darks or flats resolves the banding issue.
Georg

[cmassa]

Since bias frames are so easy to make, I've always made my master bias from 200 frames. I've made a superbias too, no change.

The Canon Banding script works for horizontal bands. I rotate the image to do the vertical bands.

Chris

[pfile]

what do your flat histograms look like? are all 3 channels equally strong?

rob

[cmassa]

Here is a flat opened in PS. The peaks are way to the right. I made them with an ipad on a white screen, t-shirt over the objective. Camera set a +2ev in AV mode, captured with BYEOS. I use 50 flats.

Chris

[gvanhau]

Hello Chris

Never seen such strong banding, certainly temp has influence on it.

Does your bias and dark frames or masters also show this banding?

If you are you using an AC adaptor, you could test imaging without it, yust to see if it has an influence....

Another test you can do is calibrate images without darks, bias or flats and see if one of them has a influence on the result or even you could try to align and stack the lights without calibration...

Geert

[cmassa]

I am using a cheap, off-brand AC adapter. I'll try with the battery to see if it makes a difference. If it does, I might have to go for the Canon power supply.

I'll try different stacking combinations.

Chris

[pfile]

the flat looks good, but i suppose it's better to open it with a tool that can show the linear histogram.

i had some banding that, as far as i can tell, was partially caused by weak signal in one channel of my flats.

rob