Flat frames - what to do when they are not perfect?

[jlodge]

Hello everyone, I am new to the forum and have just started a trial of PixInsight and am processing my first NB DSO data from a new ZWO ASI1600MM-Pro camera having made the move from DSLR imaging (processed with Nebulosity for calibration and Photoshop/Lightroom thereafter).

I use a society observatory that has a white section painted on the inside of the dome for taking flats (although I have recently ordered an EL panel, but it has not arrived yet). It is lit from the white lights around the dome and I point the OTA at a section that looks the most even. I use the flats calibration wizard in SGP to get an ADU close to 30,000 which gives a nice histogram where everything is in the middle.

I am at the stage of calibrating my lights and came across this post https://pixinsight.com/forum/index.php?topic=10577.0, reply #2 by Niall Saunders where he mentions: 'using a large piece (A3-size) of white foam-board inned to the observatory wall about 2m (6 feet) from the aperture of the OTA, and I 'point' the OTA (very roughly) 'at' the board. This is illuminated by a 60W (eq.) CFL lamp about 4m (12 feet) away (and slightly to one side) - is the illuminated field 'even'? Again, I haven't found any means within PI to tell that it isn't. Is the illumination 'perfectly white' - no, it isn't, it's anything but white - nut I have a very simple routine in PI that sorts all that out before the MasterFlat actually gets used to calibrate the raw Lights.'

This is similar to my situation and I wonder what the simple routine is in PixInsight that sorts out the flats? As I'm sure that despite my best efforts the field will not be evenly illuminated.

Also I have read in a few places that Bias frames are no good for this camera, so I have taken dark frames for my flats to match, temperature, gain/offset and exposure. These darks will be subtracted from the flats to create the master flat as opposed to using the darks which I took for my light frames, does this sound like the correct approach?

[Niall Saunders]

Hi,

You wrote

This is similar to my situation and I wonder what the simple routine is in PixInsight that sorts out the flats?

Well, yes the routine is reasonably 'simple - but it is really a series of pre-existing PixInsight Processes that I apply to get the reults that I am looking for. Unfortunately, these processes are not perhaps the easiest to follow if you are a complete PI novice.

You then asked

Also I have read in a few places that Bias frames are no good for this camera, so I have taken dark frames for my flats to match, temperature, gain/offset and exposure. These darks will be subtracted from the flats to create the master flat as opposed to using the darks which I took for my light frames, does this sound like the correct approach?

What you are trying to achieve is the solution to the following pseudo-equation
MasterLight = RegisteredAverage[ (Lights - MasterDark) / (MasterFlat - MasterFlatDark) ]

Note that there is no reference anywhere in the above equation to Biases or even to MasterBias frames. This is only valid if Bias ADU readings are so small as to be insignificant, or if the temperature is identical for Lights and LightDarks, and Flats and FlatDarks also have a similar, identical, temperature - but where the two temperatures themselves do not need to be identical.

I hope this makes sense.

[jlodge]

Well, yes the routine is reasonably 'simple - but it is really a series of pre-existing PixInsight Processes that I apply to get the reults that I am looking for. Unfortunately, these processes are not perhaps the easiest to follow if you are a complete PI novice.

Hi Niall, thank you for the quick reply, that makes sense regarding the Bias frames, I am interested to hear more about your routine.

[Niall Saunders]

Hi again,

You might want to have a look at the following link
https://pixinsight.com/forum/index.php?topic=12865.msg78210#msg78210

It is a very brief synopsis - one of these days I may get round to making the PixelMath explanation more detailed.

[jlodge]

Thanks Niall, yes that does look quite in-depth, correct me if I am wrong, but it looks as though it applies more to colour issues with OSC cameras? Mine is mono so hopefully that won't apply for my situation, my concern is over my field not being evenly illuminated. That being said i used the same technique for DSLR images and didn't notice any problems, I'll see how I get on.

[Niall Saunders]

Hi again (? ? ?),

If you are using a light-box (or even an EL panel) you can take a stack of, say, 30-odd FlatFrames, and then repeat the exercise again, acquing another batch of FlatFrames - but this time you would 'rotate' the orientation of the light-box by around 90°. You would not need to acquire BiasFrames, FlatDarks, or any kind of lights.

Perform two very simple,  unregistered stacks of each data set, and then subtract one MasterFlat from the other (thus eliminaating all Bias and Dark noise). Examine the final result under a fairly strong STF.

What you should see is a uniform, extremely low-level ADU across the whole image. If you don't have a uniform source of illumination, then you are going to see a spread of ADU values.

Of course, each of the stacks themselves will have variations in illumination - it is, after all, this very variation that you are trying to capture to correct anomalies in your optical train.

Now, that was easy enough for a light-box setup, where the alignment ot the light-box with respect to the X and Y axes of the imaging sensor was relatively easy to control. But, things are totally different if you are trying to work with WallFlats.

It isn't just as simple as rotating the target, and rotating to source of illumination is a complete non-starter. Perhaps (and I am free-thinking here - I have never tried this for myself) you could get a crafter's wooden 'needle-point' frame, large enough to cover your OTA, and stretch a clean cotton tee-shirt (or similar) over the frame. The whole frame could then be rotated as the OTA points directly overhead, gathering enough data to be able to demonstrate that your tee-shirt flat is at least 'uniformly illuminated'.

Could you then use one of the tee-shirt MasterFlats to then be subtracted from a stacked MasterWallFlat? Would that give you the confidence that you are looking for when using WallFlats?

All that aside, I just take plenty of Flats, and use these to calibrate an optical train that is pretty much 'perfect' (259mm / 10" f/8 Ritchey-Chrétien OTA, with a 75mm / 3" aperrture focuser, illuminating an APS-C sized CCD imager). Gone are the days when my 250mm / 10" f/4 SN tube left me fighting to illuminate image corners that looked as if they had been exposed down the centre of a toilet-roll tube

Hope this helps.