All
I seem to have a problem with flats. I use a Lepus reducer on my LX200R and is shows some pretty good vignetting. So flats are very necessary. Normally I dont seem to have issues but on the Sombrero I seem to be getting a sort of bas relief effect when I combine
I've uploaded examples of:
Flat master - sky flats (20 subs)
Sombrero Raw - uncalibrated
Somebrero - Calibrated
Sombrero - Mean combined
I've tried several with ADU ranges from 15K to 50K - the best I can seem to do is around 45K, though I normally use 24K with the 8300 cameras
I think there are two issues here. The flat is obviously not scaled properly. The other is that it doesn't seem to be taken at the same focus point or camera orientation. The latter is a bigger issue, hard to fix to match the dust motes. The first one is easy though. Just use a proper bias.
I get this problem occassionally as my CDK is very fussy with flats and they often only work so well.
I'd say you need to take some fresh darks. Take say 16 and sigma reject combine.
You also need to take flat darks and subtract them from the flats.
The scope needs to be focused when you take flats so in case you took flats and then later focused the scope to image then they will be off to some degree (not sure how much).
As far as a bias goes you could try that. I usually use flat darks on the basis the bias is included in them plus the noise at the time of the flat.
The flat also needs to be taken at the same temp as the lights or at least the dark flats do.
Camera needs to be in exactly the same position.
I usually use dusk flats with a white t-shirt over the end of the scope and take them quite quickly.
Perhaps also take more flats. I was using 3 but have started taking 6 each of the filter flat + the flat dark (I share the flat dark between the filters and binnings so I usually only need 2 flat darks - the 2x2 one for the rgb flats and the 1x1 for the luminance).
Ideally you take a flat for each rgb and Luminance + any Ha or other filters. They need to be fairly fresh as dust can shift and change.
You have to have a flat for each filter if you want to get rid of dust motes which may be specific to each filter.
Flats are not a good substitute for cleanliness as I rediscovered recently. Its far easier to clean your camera and filters often so flats are less needed.
In fact with some recent cleaning and my TEC110 and ML8300 I did not need any flats as the images were so evenly illuminated and there were no dust motes. That is better than needing flats due to dust motes.
But if you stretch your images hard to try to make up for lack of exposure time then dust motes will raise their ugly head.
You can also use the healing or clone tool in photoshop set to luminosity and clone them out later without too much trouble so if the flats are too hard work then cleaning everything and photoshopping can be a plan B.
How are you taking flats?
I have found variable results on a tricky setup (reducers/correctors often create a hot spot in the centre of the image that is hard to flat out). Different brightnesses between late dusk (too late to get a decent flat) and then the next day with the roof closed and a white cover over the end of the scope worked better. I think a certain amount of trial and error and get your procedures worked out.
Richard Crisp has done a paper on flats and that would be worth studying. He also does daytime flats using aluminium foil. Worth trying. It gets rid of the rush to get it done quickly but in my experience and in his recent posts on the subject dusk flats are hard to beat. You have to be fast and efficient as the light fades rapidly and fast camera downloads becomes more important. Some use electro panels and have no problem.
Perhaps Mike Sidonio can chim in here as he seems to be having no trouble with flats and he uses a Newt with a corrector and 16803 camera so flats would be vital. He uses a light panel from Astronomiks I am pretty sure. Spectrum of the light from these can have an effect as well. Too bright, too dim. With some cameras like FLI flats have to be at least a 5 second exposure otherwise you see the shutter causing shadows in the flat which are not of the scope/camera system and are useless.
I wouldn't mind hearing more about flats as they have been a source of problem for me at times with my CDK. I think part of that is I need to flock the scope and its carbon fibre trusses as they are reflective and throw the flats off sometimes.
Marc - thanks.
The problem I have with orientation in this case is that the combination of Camera, reducer and spacer means that I cant swing the arrangement through the fork to do a one star align after the drift align. So in this case the sky flats were taken before drift alignment, the camera removed and then replaced (carefully I had thought) after the one star. So that would likely account for part of it
Greg
Thanks for the very detailed response. So in addition to above, I take 10 flats for each of LRGB, no T shirt and do this at dusk. I've tried calibration a couple of ways now on this set of images
1. Though CCDSTACk - making flat masters with simple mean, calibrating etc
2. Though Maxim DL - using the stack processing option
--> Both give me the same result so far
I also see what you mean on the hot spot. There is definitely one present without much latitude for error - 50 K and its there, 45K and its gone.
I've been trying some daytime flats inside this afternoon on a white surface with differing ADU counts. So i will see how I go there
I would also like to hear more from the group on flats. Its my new bugbear!!
Dan
Ths point on focus is a good one and not something I had thought of.
Interesting - no tshirt. Don't you end up with an image of something? Perhaps not. I've always used a t shirt otherwise I might end up with an image with stars in it or a cloud or something. Perhaps not needed.
The key point of the last message was dark flats -are you using them?
If not then you will probably hit trouble. Also I find accurate darks for the image to be key as well.
That's interesting about the ADU level being so sensitive. I'll experiment more with my own with that.
I have noticed that bias frames are incredibly important for good flat calibration, especially with narrowband imaging.
After bias subtraction :
If your image has an adu reading of say 4000 at the center and at the corner it is 1000 because of vignetting (4:1). Then you take the flats and your adu reading is 40000 at the center and 10000 at the corner (4:1).
This will correct well.
Before bias subtraction Bias of 500adu
If your image has an adu reading of say 4500 at the center and at the corner it is 1500 because of vignetting (3:1). Then you take the flats and your adu reading is 40500 at the center and 10500 at the corner (3.85:1).
This will over correct the corners and they will end up brighter than the center.
Daniel, flats are simply an image of a featureless object that you use to show up, then remove the effects of: dust, optical and ccd imperfections of your entire imaging system.
Twilight flats are best, as you are shooting in the same conditions as you are in the dark (i.e. identical instrument illumination and scattering ) , use a median combine ( with at least three exposures) with small field changes create a master flat to ensure you don't get any stars or cosmic rays in the data.
With SBIG cameras you *do not need* bias frames. Bias information ( which is simply a zero second dark frame) is contained within every dark frame, so all you need to calibrate your data well, is a master flat and a matched exposure dark frame.
Back to flats. For the flat field data to be valid, nothing should be changed between taking flat field and light field data. By nothing, I mean focus, filter position & camera orientation. Unless you have a very elaborate registration pin arrangement, there is no way you can remove, then remount a camera and expect the above to be identical. Change any of these and you will get the results you see now.
Flats also need to be dark frame calibrated before you can use them. Using auto dark darks when taking flats makes this a no-brainer. Calibration order is important. Your light frames need to be dark subtracted before you apply flat field data. Most software does this for you, but If you do this in the reverse order you end up with weird looking images.
Greg's comments regarding keeping everything as clean as possible should be heeded, no dust = no dust doughnuts, but vignetting etc will still need fixing via your flats. Hope that helps.
Based on the images supplied I would say you are trying to use flats that are not in exact alignment with the lights. If you remove the camera before taking flats then you will get the problem you have experienced. Peter (w) is correct using a pin alignment is the only way you can use flats from previous sessions or sessions afterwards.
I use a light box made up by EXFSO. If you can get the camera down to the same temp as the night of taking the flats then you can do flats at anytime with a light box so long as you don't have light leaks in you imaging system.
I always take my flats, dark flats and bias frames as I'm tidying up for the night. Flats and bias frames taken under the same conditions as the lights are very important for me to do. Darks are from a library that are at most a few months old but can be done at any time.
Check what average ADU your flats are at. I find between 1/3 to 1/2 full well depth in the histogram is best. I now use Sequence Generator pro which calibrates the flats to give a user designated time of exposure with my light box to give an exact ADU for every filter. Solved all my flats problem in one step.
Allan
Daniel, flats are simply an image of a featureless object that you use to show up, then remove the effects of: dust, optical and ccd imperfections of your entire imaging system.
Twilight flats are best, as you are shooting in the same conditions as you are in the dark (i.e. identical instrument illumination and scattering ) , use a median combine ( with at least three exposures) with small field changes create a master flat to ensure you don't get any stars or cosmic rays in the data.
With SBIG cameras you *do not need* bias frames. Bias information ( which is simply a zero second dark frame) is contained within every dark frame, so all you need to calibrate your data well, is a master flat and a matched exposure dark frame.
Back to flats. For the flat field data to be valid, nothing should be changed between taking flat field and light field data. By nothing, I mean focus, filter position & camera orientation. Unless you have a very elaborate registration pin arrangement, there is no way you can remove, then remount a camera and expect the above to be identical. Change any of these and you will get the results you see now.
Flats also need to be dark frame calibrated before you can use them. Using auto dark darks when taking flats makes this a no-brainer. Calibration order is important. Your light frames need to be dark subtracted before you apply flat field data. Most software does this for you, but If you do this in the reverse order you end up with weird looking images.
Greg's comments regarding keeping everything as clean as possible should be heeded, no dust = no dust doughnuts, but vignetting etc will still need fixing via your flats. Hope that helps.
Some good tips there Peter. I'll have to try that sky flat approach. CCDsoft won't allow a dark subtract when doing a flat. I think I managed to trick it into doing it once or twice but not routinely
Some sky flats median combined might work well on my CDK which with its corrector generally has a hot spot centre and vignetting that does not correct fully with flats so far. I'll give it a go.
Peter
So two lessons - clean my optics but more importantly, dont disturb the optical train between taking flats and lights
Thanks
Dan
Quote:
Originally Posted by Peter Ward
Daniel, flats are simply an image of a featureless object that you use to show up, then remove the effects of: dust, optical and ccd imperfections of your entire imaging system.
Twilight flats are best, as you are shooting in the same conditions as you are in the dark (i.e. identical instrument illumination and scattering ) , use a median combine ( with at least three exposures) with small field changes create a master flat to ensure you don't get any stars or cosmic rays in the data.
With SBIG cameras you *do not need* bias frames. Bias information ( which is simply a zero second dark frame) is contained within every dark frame, so all you need to calibrate your data well, is a master flat and a matched exposure dark frame.
Back to flats. For the flat field data to be valid, nothing should be changed between taking flat field and light field data. By nothing, I mean focus, filter position & camera orientation. Unless you have a very elaborate registration pin arrangement, there is no way you can remove, then remount a camera and expect the above to be identical. Change any of these and you will get the results you see now.
Flats also need to be dark frame calibrated before you can use them. Using auto dark darks when taking flats makes this a no-brainer. Calibration order is important. Your light frames need to be dark subtracted before you apply flat field data. Most software does this for you, but If you do this in the reverse order you end up with weird looking images.
Greg's comments regarding keeping everything as clean as possible should be heeded, no dust = no dust doughnuts, but vignetting etc will still need fixing via your flats. Hope that helps.
Paul
Thanks - I think thats lesson learned on the alignment. The light box seems light a good idea and I've seen some good and simple designs in these forums
Dan
Quote:
Originally Posted by Paul Haese
Based on the images supplied I would say you are trying to use flats that are not in exact alignment with the lights. If you remove the camera before taking flats then you will get the problem you have experienced. Peter (w) is correct using a pin alignment is the only way you can use flats from previous sessions or sessions afterwards.
I use a light box made up by EXFSO. If you can get the camera down to the same temp as the night of taking the flats then you can do flats at anytime with a light box so long as you don't have light leaks in you imaging system.
