Dusty Corona Australis

[Octane (Humayun)]

The more I look at it, the more I think it looks wonderful as is.

lol @ Ritchie Creation. That is GOLD.

H

[richardo (Rich)]

Originally Posted by Octane

The more I look at it, the more I think it looks wonderful as is.

lol @ Ritchie Creation. That is GOLD.

H

Hi H,
yeah pretty funny eh!
My late Friday afternoon work debriefing buddy lays claim to it,
he dropped this one on me and had me rolling on the floor.... classic I thought.. just classic

Cheers Humayun

Rich

[JD2439975 (Justin)]

Very nice indeed Rich, lovely fluffy dust.

[gregbradley]

I am not sure that is correct Rich. Doesn't a 16 bit CCD do 65535 ADU? Ins't that a measure of how many shades of grey it can
differentiate?

I'll have to check but I think well depth is a different thing. It measures how many electrons each well can hold. You could be right here though. Does that mean a camera with a well depth of less than 65535 cannot differentiate all shades of grey? (mind you your typical image only sits in a fraction of those 65535 levels of grey).

But what I was saying was go 80% of 65535 = 52,428.

I may well be the one here who has it wrong so let me check that question out first.

The other possibility with a light ring or other things is uneven illumination.

Also you didn't move the camera orientation after the flats did you? It has to be the same - probably was though if the dust donuts went.

Did you use median combine?

I have had odd flats at times that did not work. I put it down to the fact they seemed too bright. Also perhaps a wrong conclusion.

I get good results doing 3 flats median combined shooting for 20,000 adu or thereabouts with at least 2 or 3 second exposures (my cameras have physical shutters that show in less than 2 or 3 secs) using a white Tshirt at dusk and median combine with bias subtraction (not that important).

Can you post one of your flats?

Greg.

Originally Posted by richardo

Thanks for this Greg.
I've been using a light box.
I set my flats in all channels to around 29-30,000 ADU which is about 70-75% the full well depth of the KAI4021m chip.
While my Lum flats divided well after bias subtraction on flats and lights, I saw over correction in the colour channels to the corners where my vignetting is. Dust motes were corrected nicely though.
Has me baffled.
I've tried lower ADU levels with no success, I might give the 80% a go and see what happens.

Maybe I'll try the 'T' shirt flats with twilight and see how the colour gets on. Seems strange that the colour flats correct the motes but not correctly do the vignetting.
I'm a bit over it now as I've tried and tried..

As I've mentioned, once I get the 2" filters, vignetting will no longer be an issue. I still will be flat fielding the Lum however.
But hate it when things don't do as they should.

Cheers
Rich

[richardo (Rich)]

Originally Posted by JD2439975

Very nice indeed Rich, lovely fluffy dust.

Thanks very much Justin!
It's like where I live, dusty dusty dusty!

All the best
Rich

[Martin Pugh]

A fine image indeed Ricardo, I like the pastels in the way you have processed this.
cheers
Martin

[richardo (Rich)]

Originally Posted by gregbradley

I am not sure that is correct Rich. Doesn't a 16 bit CCD do 65535 ADU? Ins't that a measure of how many shades of grey it can
differentiate?

I'll have to check but I think well depth is a different thing. It measures how many electrons each well can hold. You could be right here though. Does that mean a camera with a well depth of less than 65535 cannot differentiate all shades of grey? (mind you your typical image only sits in a fraction of those 65535 levels of grey).

But what I was saying was go 80% of 65535 = 52,428.

I may well be the one here who has it wrong so let me check that question out first.

The other possibility with a light ring or other things is uneven illumination.

Also you didn't move the camera orientation after the flats did you? It has to be the same - probably was though if the dust donuts went.

Did you use median combine?

I have had odd flats at times that did not work. I put it down to the fact they seemed too bright. Also perhaps a wrong conclusion.

I get good results doing 3 flats median combined shooting for 20,000 adu or thereabouts with at least 2 or 3 second exposures (my cameras have physical shutters that show in less than 2 or 3 secs) using a white Tshirt at dusk and median combine with bias subtraction (not that important).

Can you post one of your flats?

Greg.

Greg, mmmm, there seems to be two camps amongst the FF experts on the net.
And as I'm pretty green to flat fielding, maybe I've gone off with the wrong camp
On one count, I've read, and as I've thought, that you need between 30-50% of your full well capacity: This I can recall was from a discussion on cloudy nights with a guy who was using a H16.
Here's a link
[http://stargazerslounge.com/imaging-...ibration.html]

And I've just had a quick look once again (hate to think how much time i've spent on trying to find definitive info on this matter )...
that as you've said... 16bit ccds aim for around 20,000 adu... or 1/3rd there of 65,535

OK, now I'm really confused

Here's another equation I found that is to give the correct target for flat fields, just to throw another spin on things, but this seems to indicate what you've been saying ..

full well/ADU gain*30-50%=target

So for my camera that equates as follows
40,000/ .6 * 30%=19,999.99rec so 20,000adu

I did have a read after you commented on Richard Crisps document pertaining to filter edges and things... interesting and I think this has solved a drama for a friend of mine...

No I didn't move the camera at all.. the donuts were gone with the rgb's.

I median combined my master bias subtracted flats to create a master flat for each individual channel.
I first wasn't bias subtracting from either the lights or the flats... things didn't work. Then I did bias subtract from both and that's when my lum flat fielded very well.
The rgb channels didn't however do that well with the vignetting.

Ok, so I'm going with you with aiming for 20,000adu....
I'll give that a try and see how things go.

Thanks for bring this to my attention..

Rich

[richardo (Rich)]

Originally Posted by Martin Pugh

A fine image indeed Ricardo, I like the pastels in the way you have processed this.
cheers
Martin

Thanks very much Martin!

Appreciate you taking a peek and commenting!

All the best
Rich

[gregbradley]

Hi Rich,

Here is the link to the paper on flat fields by Richard Crisp who is a CCD engineer and who started narrowfield imaging:

http://www.astromart.com/forums/view...ews_id=&page=3

You are right it is 80% of full well capacity not 80% of the 16bit levels (ie. 65535).

The bottomline here of his article is that the higher the % of full well used for flats the less are needed to get lower noise. The lower the % of full well used the more noise you will introduce and the more flats will be needed to achieve comparable noise levels.

So in your case of full well of 40,000 you would aim for flats of about 32,000 ADU. You would bias subtract (at the same temp).

Note the scope needs to be in focus for flats. So make sure your scope is in focus first. Important for me as I may do flats at dusk before I have imaged and thus before the scope is focused.

Illumination for the flats of course has to be even as flats are trying to even out the uneveness of illumination from both the optical path and the chip itself (it may not be uniformly sensitive across the whole chip).

Are you doing autodarks with your flats? The darks would be virtually identical to the bias being such a short exposure and isn't part of the
procedure.

With the colour channels are you doing flats 2x2 to match your colour channels 2x2? If you take a 1x1 colour flat and try to apply it you may get an error message but if not you will get weird results that don't work.
I have sometimes used a luminance 1x1 and reduced it to 2x2 in CCDstack and applied it to a colour channel when I didn't have a flat for that colour. Sometimes it worked but often it did not. Better to use another colour's flat if for some reason (like it got dark hehehe) you don't have the correct colour's flat.

The only other thing may be if your filters are loose and move around causing the lights to be different to the flats. Apart from that you would have to detail every minute step taken to see if there seemed anything odd.

It is easy to continue to shoot flats at 1x1 as you probably did your lumninance flat first and then the colours. An easy way to tell is to look at the file sizes of the colour flats versus the luminance. 2x2 is about 1/4 the file size of 1x1.


Greg.

[RobF (Rob)]

Wonderful work Rich. Inspiring stuff!

[richardo (Rich)]

Originally Posted by gregbradley

Hi Rich,

Here is the link to the paper on flat fields by Richard Crisp who is a CCD engineer and who started narrowfield imaging:

http://www.astromart.com/forums/view...ews_id=&page=3

You are right it is 80% of full well capacity not 80% of the 16bit levels (ie. 65535).

The bottomline here of his article is that the higher the % of full well used for flats the less are needed to get lower noise. The lower the % of full well used the more noise you will introduce and the more flats will be needed to achieve comparable noise levels.

So in your case of full well of 40,000 you would aim for flats of about 32,000 ADU. You would bias subtract (at the same temp).

Note the scope needs to be in focus for flats. So make sure your scope is in focus first. Important for me as I may do flats at dusk before I have imaged and thus before the scope is focused.

Illumination for the flats of course has to be even as flats are trying to even out the uneveness of illumination from both the optical path and the chip itself (it may not be uniformly sensitive across the whole chip).

Are you doing autodarks with your flats? The darks would be virtually identical to the bias being such a short exposure and isn't part of the
procedure.

With the colour channels are you doing flats 2x2 to match your colour channels 2x2? If you take a 1x1 colour flat and try to apply it you may get an error message but if not you will get weird results that don't work.
I have sometimes used a luminance 1x1 and reduced it to 2x2 in CCDstack and applied it to a colour channel when I didn't have a flat for that colour. Sometimes it worked but often it did not. Better to use another colour's flat if for some reason (like it got dark hehehe) you don't have the correct colour's flat.

The only other thing may be if your filters are loose and move around causing the lights to be different to the flats. Apart from that you would have to detail every minute step taken to see if there seemed anything odd.

It is easy to continue to shoot flats at 1x1 as you probably did your lumninance flat first and then the colours. An easy way to tell is to look at the file sizes of the colour flats versus the luminance. 2x2 is about 1/4 the file size of 1x1.


Greg.

Thanks for clearing this up concerning well depth.
Now I know where I am.

I only shoot my colour in 1xs binning. Back in 2002 I think it was when i first purchased my SX-H516, I liked the idea behind 2xs binning. Combining 4 pixels to make one and bringing the filtered light in quicker. Trying it I found I didn't like the extra work and end result so haven't binned since.

I haven't bothered with dark subtraction as I find the chip, at this time of the year, I can get away with sigma combining my subs with little or no residue of noise. Maybe in summer it might be a different story. Still generally, in summer, the temperature drops to the mid teens. The temp I use at the moment is -15C, this seems to be all that is needed to keep things fairly clean from noise.

I will attempt again to see how my flats go..... since I've had a break, I have a little more head space...

Rich

Originally Posted by RobF

Wonderful work Rich. Inspiring stuff!

Thanks very much Rob!
It's an inspiring area of the sky

Cheers guys
Rich