Flat frames

[Astronix]

Hi All,

 I hope everybody is doing well.

Could someone give me an idea of how you they collect their flat frames. I am really struggling at the moment. I have tried the T shirt over the scope with a light source from my iPad method with very mixed results. I have also taken the frames in the morning sky with the white T shirt and had better results. The problem with this method is the weather and clouds 😩

does anyone have a better method? I am using an Asiair to gather the frames and I am getting some good data, but I think that my flats are letting the final image down. Darks and bias frames seem fine.

many Thanks 

[barbulo]

IMO, the most important thing is to maintain the optical train used with the light frames unaltered (don´t change focus, fliters...). Also an even source of light is important, hence the t-shirt. Some cameras are specially sensible to short exposure times. With the ASI294MM I use between 5s and 35s exposures depending on the filter (to reach around 30.000ADU - IRRC).

What camera are you using?

[ONIKKINEN]

How are your current flat frames not doing the job? Also, what scope and camera do you have?

Some cameras care about exposure times with flats, others do not. Some cameras also have issues with certain gain values (like the 294) not having a linear response so flats calibration is tricky. I am assuming that you are also taking other calibration frames and using them: Matching darks, darkflats (bias is ok for some models). Also this depends on the camera, with some DSLRs you shouldn't use darks at all, so hard to advice at this point.

[Astronix]

15 minutes ago, barbulo said:

IMO, the most important thing is to maintain the optical train used with the light frames unaltered (don´t change focus, fliters...). Also an even source of light is important, hence the t-shirt. Some cameras are specially sensible to short exposure times. With the ASI294MM I use between 5s and 35s exposures depending on the filter (to reach around 30.000ADU - IRRC).

What camera are you using?

Asi533 mcpro 

 

[Astronix]

14 minutes ago, ONIKKINEN said:

How are your current flat frames not doing the job? Also, what scope and camera do you have?

Some cameras care about exposure times with flats, others do not. Some cameras also have issues with certain gain values (like the 294) not having a linear response so flats calibration is tricky. I am assuming that you are also taking other calibration frames and using them: Matching darks, darkflats (bias is ok for some models). Also this depends on the camera, with some DSLRs you shouldn't use darks at all, so hard to advice at this point.

The scope is a Skywatcher 72ED the camera is an Asi533 pro colour, I also have an EAF auto focused.

Edited November 11, 2022 by Astronix

[PeterCPC]

I use a light pad (box) with mine.

[Astronix]

I think the main thing causing the problem is the light source, what do you use?

[Astronix]

4 minutes ago, PeterCPC said:

I use a light pad (box) with mine.

Do you have a link for where you bought it?

[PeterCPC]

I can't be sure but I think that it was on Amazon. It was about £20.

Something like this https://www.amazon.co.uk/Portable-Adjustable-3-Levels-Brightness-Animation/dp/B07GSVB7RR/ref=sr_1_9?crid=116RBNRUN5ZE1&keywords=light+box&qid=1668176081&sprefix=light+box%2Caps%2C161&sr=8-9

Edited November 11, 2022 by PeterCPC

[Anthonyexmouth]

A cheap EL panel from ebay and a bit of ply with a hole the same size of the dew shield so it slips over and cant fall off. Gives a more even light than an LED panel. 

[Elp]

I still sometimes get odd flats. I've found at the moment using the led tracer panel suggested above, dim to lowest setting, and use a couple of opal and dark grey transparent perspex sheets dims it enough to get around 5-10s exposure flats at around 25-30k white level. It's too bright otherwise.

[PeterCPC]

I use sheets of A4 paper on mine to dim the intensity.

[ONIKKINEN]

1 hour ago, Astronix said:

The scope is a Skywatcher 72ED the camera is an Asi533 pro colour, I also have an EAF auto focused.

With the 533 you dont have to worry about exposure time since its a very linear response sensor, even flats in the milliseconds of exposure range will work provided that you take an enormous pile of them.

I have an LED tracing panel from amazon, a really cheap trinket but it does the job just fine. I used to use printer paper between the panel and my scope, but that can actually cause extra gradients rather than solve them, since printer paper is hardly optical quality. White t-shirts, maybe not so much of an issue but in my opinion there is no reason to bother doing this at all, just put the panel in minimum brightness and use some automatic flats taking software to figure out the exposure (Asiair?). If the panel is bright the exposure could be like 0.1s but thats ok, just take 50 and call it a day.

But back to the original question i had, how are your current flats not working? Are they not removing vignetting? Are they overcorrecting so that edges are bright? Do you have a weird circle around the center of the frame, or is it something else?

[Astronix]

4 hours ago, ONIKKINEN said:

With the 533 you dont have to worry about exposure time since its a very linear response sensor, even flats in the milliseconds of exposure range will work provided that you take an enormous pile of them.

I have an LED tracing panel from amazon, a really cheap trinket but it does the job just fine. I used to use printer paper between the panel and my scope, but that can actually cause extra gradients rather than solve them, since printer paper is hardly optical quality. White t-shirts, maybe not so much of an issue but in my opinion there is no reason to bother doing this at all, just put the panel in minimum brightness and use some automatic flats taking software to figure out the exposure (Asiair?). If the panel is bright the exposure could be like 0.1s but thats ok, just take 50 and call it a day.

But back to the original question i had, how are your current flats not working? Are they not removing vignetting? Are they overcorrecting so that edges are bright? Do you have a weird circle around the center of the frame, or is it something else?

It looks like I am getting a lot of noise and I thought that flats were taken to reduce this

 

[ONIKKINEN]

18 minutes ago, Astronix said:

It looks like I am getting a lot of noise and I thought that flats were taken to reduce this

 

Flats have nothing to do with noise reduction of the image, flats are taken to flatten the image by removing vignetting and shadows from dust on/close to your camera (like on the field flattener lenses or filters).  The noise in your flats themselves reduces just by taking more of the flats. How many are enough depend on how short your exposure is, but with your already very low noise 533 you dont need to take all that many (maybe 30) to have a very good master flat.

You can decrease noise of your stacked image by building up a longer integration. How long is enough depends greatly on your sky conditions, if you have light pollution you will need to shoot much longer compared to dark skies where a few hours is enough.

[andrew s]

Two points.

Flats, properly applied, remove fixed pattern noise due to pixel response differences. The dominant noise a high signal levels. This then makes shot noise the dominant noise.

With some ASI CMOS cameras they have different response at very short exposures compared to 1s or more although in the example both regions were linear.  See C Buils spectroscopy Web site.

Regards Andrew

Edited November 11, 2022 by andrew s

[jacko61]

A word of warning on the £20 a4 led tablets on amazon. I've been using mine about a year and I'm finding the usb socket is getting very loose and sometimes needs the plug bending in one direction or another to get the power to the tablet which can be very frustrating at 3am when I'm wanting to close the roof and get to bed.   I'm thinking of getting it working then putting a decent blob of hot glue around it to hold it in the right place.  

Graeme

[tooth_dr]

2 hours ago, jacko61 said:

A word of warning on the £20 a4 led tablets on amazon. I've been using mine about a year and I'm finding the usb socket is getting very loose and sometimes needs the plug bending in one direction or another to get the power to the tablet which can be very frustrating at 3am when I'm wanting to close the roof and get to bed.   I'm thinking of getting it working then putting a decent blob of hot glue around it to hold it in the right place.  

Graeme

Agree with this. I’ve already scrapped one with a failed usb port. 

[bottletopburly]

I use a zwo533mc i use an old galaxy tab2 for flats on a 70mm refractor here is how i do it , save an image of a white photo , Dim the brightness down  try on its lowest setting place a sheet of white grease proof paper(L-Extreme filter )2 sheets for no filter ) in front of panel  aim for 25k ADU now you want an exposure between 3-5 secs  ,using the L Extreme my exposure is 4.375 sec and i use APT flats wizard to find the correct settings ,you may need to tinker with panel brightness to find the sweet spot to get between 3-5 sec  with a 25k adu , also have a look at Adam blocks videos there is a Three part video well worth a watch adam explains why you should use 3-5 sec exposures using a cmos sensor 

 

 

[ninjageezer]

12 hours ago, jacko61 said:

A word of warning on the £20 a4 led tablets on amazon. I've been using mine about a year and I'm finding the usb socket is getting very loose and sometimes needs the plug bending in one direction or another to get the power to the tablet which can be very frustrating at 3am when I'm wanting to close the roof and get to bed.   I'm thinking of getting it working then putting a decent blob of hot glue around it to hold it in the right place.  

Graeme

I've just bought one B4 size so it fits my 8 inch edgehd as well .the good thing is its its got a short built in lead with a 5.5mm socket on so hope it lasts better ,very common fault on most led light pads the loose socket. I think it was someone on here who recommended the one I got.

the back of the panel is white as well and no where near as bright and this can be used without paper I believe. the front goes super bright !

https://www.amazon.co.uk/gp/product/B07WQXFZN5/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1

 

 

[ninjageezer]

On 11/11/2022 at 15:22, ONIKKINEN said:

With the 533 you dont have to worry about exposure time since its a very linear response sensor, even flats in the milliseconds of exposure range will work provided that you take an enormous pile of them.

I have an LED tracing panel from amazon, a really cheap trinket but it does the job just fine. I used to use printer paper between the panel and my scope, but that can actually cause extra gradients rather than solve them, since printer paper is hardly optical quality. White t-shirts, maybe not so much of an issue but in my opinion there is no reason to bother doing this at all, just put the panel in minimum brightness and use some automatic flats taking software to figure out the exposure (Asiair?). If the panel is bright the exposure could be like 0.1s but thats ok, just take 50 and call it a day.

But back to the original question i had, how are your current flats not working? Are they not removing vignetting? Are they overcorrecting so that edges are bright? Do you have a weird circle around the center of the frame, or is it something else?

when you say an enormous pile how many ?? i have a new 533mc pro also.😀

[ninjageezer]

On 11/11/2022 at 14:25, Elp said:

I still sometimes get odd flats. I've found at the moment using the led tracer panel suggested above, dim to lowest setting, and use a couple of opal and dark grey transparent perspex sheets dims it enough to get around 5-10s exposure flats at around 25-30k white level. It's too bright otherwise.

do you have a link at all for the perspex sheets please.

[Elp]

Something like this:

https://www.ebay.co.uk/itm/174042191592?hash=item2885b916e8:g:jxwAAOSwbl9dnxi8&amdata=enc%3AAQAHAAAA4MwpgssR%2FriubOTTJv%2FM1GCVx%2B3BafmNNyiyOkKX9MF%2FSASgWvji20NTOmBoafrCyDFVm3epHS4vDgMFQnKznTrvYXw6a4zpbAqpDA%2FNKdTXvX4FsNMxFTr0zXr5ZhI419%2BKoRKalIG2YUUtZ5%2F8EtXBtdZ5UdzBh%2FeE2hZW7I0v6J3Am7VW3k9Rjav2J0axkF1cGH03CHn9gLZQbDIvs9HREvn1aZv8jcIAPLzt4lkI9zv%2BkgLxozR2nrO0gY80u29SIjd2dlKCuExtdzYbfoTlJWers1arAdS%2FCZxrbfJZ|tkp%3ABk9SR8CW5qaOYQ

I have a plastic supplier nearby, I asked them if they had any offcuts and they kindly gave me some.

Note, only the white opal I've tried and the transparent grey allow light through. I've tried a black and grey gloss but they blocked too much light.

[ONIKKINEN]

42 minutes ago, ninjageezer said:

when you say an enormous pile how many ?? i have a new 533mc pro also.😀

Not sure if i could think of some concrete number, its something you see when inspecting the flat master. If the flat master looks noisy when stretched, it should have more exposures thrown in. I take 0.2s flats and usually take 50 or 60 of them. Result is quite clean and noise levels are acceptable. If you had a flat panel that necessitates lets say 50ms (0.05s) flats you might want to take a hundred, although when i imaged at a higher gain and needed 50ms flats, i still used just 50 flats. Up to you to decide if you care about that enough, most dont with these new cameras. So in short, take 50 and forget about it, if you have flats issues then maybe look back at the flats and see if something should be improved.

[Ouroboros]

I have wondered whether crossed polarisers might be a way to adjust down the light level of led panels.