zwo 1600 narrowband flats

[sonnymoon]

Hello,

I see that there have been other posts about zwo 1600mm narrowband flats in the past but I'm still struggling to understand whats going on with mine.

First one is Ha, then Sii and O3. The S2  looks like it might be okay - lighter in the middle and darker around the edges with dust spots etc but the other two have a strange banding.....

Flats were taken using an LED panel bought from Ebay put in a home made cardboard box. These ones are at 25000 adu, although the problem with the O3 one in particular shows up at 15000adu in a stack of lights, darks and flats and kind of ruins it.

Any thoughts greatly appreciated

[david_taurus83]

Hi. What scope are you using?

[vlaiv]

All is well with those flats.

Banding that you see is just artifact of scaled down image.

ASI1600 has very small "variance" between pixel quantum efficiency - in checkerboard pattern. This has been noted in various threads and is consequence of manufacturing process.

Many people think of pixels as being perfect squares, but they in fact are semi conductor devices on silicon. They have photo sensitive area, but also some circuitry around them. There is a difference between front and back illuminated sensors.

This is schematics of construction - here it's important to note that electronics sits on top of photo diode layer, and can also mask some of light gathering area.

And this is what a sensor looks like under microscope:

You will notice that there are many manufacturing artifacts, and pixels are not perfect. This affects quantum efficiency of pixels - one of the reasons we take flats, to even things out.

Back to your concern - This is what you will find if you zoom in on your flats:

There is checkerboard pattern of varying intensity (less than 1% difference, but it shows with increased contrast). If you scale down your image this pattern produces above lines.

Your flats will work as they should, so you don't need to concern your self with it.

Olly gave good recommendation for checking if flats will work as they should without need for actual light frames to calibrate. Take two sets of flats with same filter, but at different exposure lengths - like 15000 ADU and 25000 ADU - and divide them (do flat calibration on flats) - you should get perfectly flat frame (bar noise).

[sonnymoon]

Thanks for your replies. Vlaiv, I will enjoy trying to get my head around yours..:)

I'm using an ED80.

As I'm laid up with a bad case of man flu I've been looking into this further and have noticed that the Ha and O3 banding extends to the lights as well. What initially drew my attention to it was correspondence with Ivo Jaeger, Startools creator, who noticed the O3 stack I sent him showed the same banding. He didn't think it looked right and was interfering with processing.

If the same banding turns up on an individual O3 sub, but not an S2 sub, and on lights and flats, would that not mean there is something amiss? I hasten to add I'm not an expert here………..

Thanks again. I've attached s2, Ha and O3 subs. I am imaging in a white zone - 4 minute subs with camera gain 200 and offset 50.

 

[sonnymoon]

On 24/02/2019 at 14:07, vlaiv said:

All is well with those flats.

Banding that you see is just artifact of scaled down image.

ASI1600 has very small "variance" between pixel quantum efficiency - in checkerboard pattern. This has been noted in various threads and is consequence of manufacturing process.

Many people think of pixels as being perfect squares, but they in fact are semi conductor devices on silicon. They have photo sensitive area, but also some circuitry around them. There is a difference between front and back illuminated sensors.

This is schematics of construction - here it's important to note that electronics sits on top of photo diode layer, and can also mask some of light gathering area.

And this is what a sensor looks like under microscope:

You will notice that there are many manufacturing artifacts, and pixels are not perfect. This affects quantum efficiency of pixels - one of the reasons we take flats, to even things out.

Back to your concern - This is what you will find if you zoom in on your flats:

There is checkerboard pattern of varying intensity (less than 1% difference, but it shows with increased contrast). If you scale down your image this pattern produces above lines.

Your flats will work as they should, so you don't need to concern your self with it.

Olly gave good recommendation for checking if flats will work as they should without need for actual light frames to calibrate. Take two sets of flats with same filter, but at different exposure lengths - like 15000 ADU and 25000 ADU - and divide them (do flat calibration on flats) - you should get perfectly flat frame (bar noise).

Just to follow up on this -  I’ve been reassured by zwo as well as Vlaiv that the flats are normal. Zwo said that the banding ‘can be calibrated out by flats’.

My question is then, that if the banding appears on the flat, the single sub and the stacked images (with flats) then that means what? That the flats are not doing their job properly? How do I calibrate it out?

Thanks

[vlaiv]

Banding should appear on flats and on single subs. It should not appear calibrated single sub nor stack of such calibrated single subs.

If you still have banding on single calibrated sub - that means that your calibration did not go well. This can happen due to number of reasons, first being actual calibration procedure.

With ASI1600 you should do following to calibrate your sub - basic version, later you can add some enhancements:

- take number of dark frames - do as much as you can. Each dark frame should match light sub exactly - that means temperature, exposure length, gain and offset. You should take care that there is no light leak at all in your darks. Sometimes IR can leak in. My method of doing dark frames is to take camera of the scope (I do this regularly since I don't have permanent setup), place camera on wooden table, covered with cap, sensor down. Then I run my capture software and do anywhere between 64 and 256 - depending on exposure length.

- Take flats, and take flat darks. For flat darks you should do the same as above - match your flats in exposure, temperature (this is not critical if your lights are very short), gain and offset. I do this while camera is on my scope by just covering my scope with lens cap.

Calibration should be as follows:

- stack darks with average method to produce master dark

- stack flat darks with average method to produce master flat dark

- stack flats with average method. Subtract master flat dark from resulting stack and save it as master flat

- calibrate light frame by subtracting master dark and dividng by master flat

If you are using software to do calibration for you, just include above files and select average method for stacking of calibration files. Important bit is avoiding light leak and using exactly the same settings for matching darks. Don't use bias.

Enhancements are usage of sigma clip for darks for example - this can help with cosmic rays (or other radiation sources) in your darks but you don't have to do it now until you get properly calibrated sub and main stack free of banding.

 

[sonnymoon]

22 minutes ago, vlaiv said:

Banding should appear on flats and on single subs. It should not appear calibrated single sub nor stack of such calibrated single subs.

If you still have banding on single calibrated sub - that means that your calibration did not go well. This can happen due to number of reasons, first being actual calibration procedure.

With ASI1600 you should do following to calibrate your sub - basic version, later you can add some enhancements:

- take number of dark frames - do as much as you can. Each dark frame should match light sub exactly - that means temperature, exposure length, gain and offset. You should take care that there is no light leak at all in your darks. Sometimes IR can leak in. My method of doing dark frames is to take camera of the scope (I do this regularly since I don't have permanent setup), place camera on wooden table, covered with cap, sensor down. Then I run my capture software and do anywhere between 64 and 256 - depending on exposure length.

- Take flats, and take flat darks. For flat darks you should do the same as above - match your flats in exposure, temperature (this is not critical if your lights are very short), gain and offset. I do this while camera is on my scope by just covering my scope with lens cap.

Calibration should be as follows:

- stack darks with average method to produce master dark

- stack flat darks with average method to produce master flat dark

- stack flats with average method. Subtract master flat dark from resulting stack and save it as master flat

- calibrate light frame by subtracting master dark and dividng by master flat

If you are using software to do calibration for you, just include above files and select average method for stacking of calibration files. Important bit is avoiding light leak and using exactly the same settings for matching darks. Don't use bias.

Enhancements are usage of sigma clip for darks for example - this can help with cosmic rays (or other radiation sources) in your darks but you don't have to do it now until you get properly calibrated sub and main stack free of banding.

 

Thanks very much, that’s very helpful.

One question; what’s a ‘calibrated single sub’? I 

[vlaiv]

7 minutes ago, sonnymoon said:

Thanks very much, that’s very helpful.

One question; what’s a ‘calibrated single sub’? I 

You take a single light frame and calibrate it you get single calibrated light frame - > single calibrated sub

This is what stacking programs are doing they take your light frames / subs / sub frames, single exposures, call them what you like and they first calibrate them with some method (you should use above for ASI1600) and they get bunch of calibrated subs that they align / register and then they get stacked to get final image.

[sonnymoon]

25 minutes ago, vlaiv said:

You take a single light frame and calibrate it you get single calibrated light frame - > single calibrated sub

This is what stacking programs are doing they take your light frames / subs / sub frames, single exposures, call them what you like and they first calibrate them with some method (you should use above for ASI1600) and they get bunch of calibrated subs that they align / register and then they get stacked to get final image.

Thanks.

Is it possible to access a calibrated single sub? I thought that the calibration process in Dss puts them all together in a file. I didn’t know they existed as individual calibrated entities ?