#1  
Old 04-11-2013, 09:14 AM
nebulosity.'s Avatar
nebulosity. (Jo)
Registered User

nebulosity. is offline
 
Join Date: Jul 2013
Location: Cecil Plains QLD
Posts: 1,228
What is making my image like this???

G'day everyone,

With my new wide field setup I seem to be having a lot of trouble with processing, I've posted an image as and example of how they've been turning out. In this particular image I've got just over 9 hours of two and a half minute exposures at iso 1600 with the lens at F8. I also got 50 darks, 60 flats and 60 biase frames. The image was stacked and streatched in Nebulosity and comes out looking like this.

What am I doing wrong? Any help appreciated.

Cheers
Jo
Attached Thumbnails
Click for full-size image (mhmm.jpg)
184.7 KB98 views
Reply With Quote
  #2  
Old 04-11-2013, 10:23 AM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
What does the image look like before you apply the flats etc?

I can get odd results with my large telescope with flats. It can be quite fussy.

Here's what I do for what its worth:

1. Darks. I use a temperature regulated CCD but still applies to DSLR. I take about 6-10 darks (my camera is very clean anyway but still is needed). I combine these in CCDStack using clip max/min combine.

2. Bias. I take only a few. Perhaps there is a gain to taking many but perhaps that applies more to DSLRs - I am not sure. Bias though are good because you want to use adaptive darks with DSLRs as the noise level will vary quite a bit as the temp you took the shot at will vary in different time periods. The bias is used to scale the darks to match your shot and it works quite well. I even use this occassionally with my astro CCD and I get good results. Best practice is darks matching lights for temperature. So for a DSLR try to take your darks under similar temperatures as your lights to minimise the difference. To get this in perspective if you look at the specs of many astro CCDs they double in noise with every 6C increase in temp. A DSLR would be similar.

3. Flats.

I have different approaches to taking flats. Both ways work. I take flats with a cotton sheet cover over the end of the scope. The scope is in focus and the camera is cooled and stable. I keep everything exactly the same as when taking the lights (you can't remove the camera, rotate it, change filters -anything).

I take 3 flats per filter. In your case there are no filters unless you are using a light pollution filter so take several. I am not sure what the gain is from taking 60 flats but perhaps for DSLRs they need more.

I do not dark subtract from my flats now. I used to take a dark for the flat duration and subtract that from the flat. But I am getting better results by only doing max/min clipping combine in CCDstack no subtractions. Then when I apply the flat I use the bias to subtract from the flat. Not sure if this is best practice but this is simply what works best on my setup.

I go for 20,000 to 40,000 ADU brightness in the flats. On a DSLR I think that equates to 1/3rd of a histogram exposure. For me I think I will implement Richard Crisps advice on flats (he has written a paper on it) where he recommends 40,000 ADU. Sometimes I find dust donuts are stubborn to disappear. I either take flats at dusk (you have to be fast) or in my observatory the next day as there is enough random light that filters in that flats work quite well.

Its best to keep everything spotless than rely on flats only as a result.

Dust donuts are more of an issue with long focal length scopes.

I find 2 mirror plus a corrector/reducer the toughest format for flats. Refractors are not fussy and often hardly vignette much at all.The CDK17 is tough work and the most difficult part of image processing with it. I see some weird effects like that if the darks do not match well or there is no bias for the flats or some combination is slighly mismatched.

So my advice is fresh darks, fresh bias and flats all at the closest you can get to the temperature when you imaged, don't change anything when you do flats.

Cameras do change slowly over time and a DSLR will be quite different between seasons. In Sydney we have high temps at night already now so that would throw things off if your darks were taken in winter.

Greg.
Reply With Quote
  #3  
Old 04-11-2013, 10:31 AM
allan gould's Avatar
allan gould
Registered User

allan gould is offline
 
Join Date: May 2005
Location: Brisbane
Posts: 4,485
Great and valuable advice there Greg. First rate and extremely helpful.
Reply With Quote
  #4  
Old 04-11-2013, 11:17 AM
Octane's Avatar
Octane (Humayun)
IIS Member #671

Octane is offline
 
Join Date: Dec 2005
Location: Canberra
Posts: 11,159
How long were your flat frame exposures, and, at what ISO?

H
Reply With Quote
  #5  
Old 04-11-2013, 11:32 AM
DJT (David)
Registered User

DJT is offline
 
Join Date: Nov 2011
Location: Sydney
Posts: 1,452
Quote:
Originally Posted by allan gould View Post
Great and valuable advice there Greg. First rate and extremely helpful.
+1 for me as well, Cheers Greg
Reply With Quote
  #6  
Old 04-11-2013, 11:58 AM
nebulosity.'s Avatar
nebulosity. (Jo)
Registered User

nebulosity. is offline
 
Join Date: Jul 2013
Location: Cecil Plains QLD
Posts: 1,228
Thanks a lot for your reply Greg, extremely helpful

I'll give a more detailed description of what I did, so I got the camera going around 6 o'clock (my cooled, modded 350d) and after about half an hour (to let it cool down) I started capturing darks. Then around 9 o'clock framed and focused and started imaging. I imaged all night and in the morning when the sky had brightened and I couldn't see any more stars I took the flats. I didn't touch the camera (left it pointing at the sky) beyond changing the exposure to 1/200 and the iso to four hundred. The histogram was about 2/3 of the way up so that may have been a problem?

The camera would have sat on around -8 most of the night I think.

All the darks and bias frames were stacked with a pure average and the flats with a 2x2 median (what ever that is, it's what Nebulosity recommended)
I removed the bias from the flats, and the darks and flats from the lights.

I'm not sure if Nebulosity has a clip max/min combine but it can do a standard deviation stack which I think chucks out any pixel thats over a certain value, would that be similar you reckon?

The dust is mainly on my sensor thats why it's so big and dark.

Cheers and thanks again
Jo
Reply With Quote
  #7  
Old 04-11-2013, 01:04 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Quote:
Originally Posted by nebulosity. View Post
Thanks a lot for your reply Greg, extremely helpful

I'll give a more detailed description of what I did, so I got the camera going around 6 o'clock (my cooled, modded 350d) and after about half an hour (to let it cool down) I started capturing darks. Then around 9 o'clock framed and focused and started imaging. I imaged all night and in the morning when the sky had brightened and I couldn't see any more stars I took the flats. I didn't touch the camera (left it pointing at the sky) beyond changing the exposure to 1/200 and the iso to four hundred. The histogram was about 2/3 of the way up so that may have been a problem?

The camera would have sat on around -8 most of the night I think.

All the darks and bias frames were stacked with a pure average and the flats with a 2x2 median (what ever that is, it's what Nebulosity recommended)
I removed the bias from the flats, and the darks and flats from the lights.

I'm not sure if Nebulosity has a clip max/min combine but it can do a standard deviation stack which I think chucks out any pixel thats over a certain value, would that be similar you reckon?

The dust is mainly on my sensor thats why it's so big and dark.

Cheers and thanks again
Jo
Hi Jo,

Firstly I am not 100% sure but one thing that stands out there is you changed the ISO. DSLRs have different responses at different ISOs. For one thing dynamic range falls as you raise ISO. ISO is simply amplification of the signal, its amplifying the image you took much like stretching an astroimage. So if the lights are at ISO800 and your darks are at ISO400 or your flats are at ISo400 that may make a difference.

If you are taking flats using the sky there is an area called the null area I only found out about recently. Its on the opposite side of the sky to the sun and is considered to be the most neutral part of the sky. I can't say I have applied this myself but try pointing the scope to the null area of the sky.

Did you put something over the end of your scope for doing flats like a cotton sheet or a Tshirt or something? You may need that as well or you may pick up some fading stars.

Also when I use bias I don't subtract it from the flat when making the master flat. I subtract when applying it. Not sure if there is any difference there but just in case.

2/3rds of a histogram on a DSLR may be too bright. Flats are divided into the image. If the flat is too bright I have also seen it through an image into a wobbly.

Do you have any light leaks if you are taking darks at 6pm, its still light then. Have you visually checked the darks to make sure they were in fact dark with no light leaks? Some DSLRs can leak light. There was an issue with the first batch of Canon 5D3's for this. It was a simple fix involving black electrical tape!

Another point on flats. It may not be an issue with DSLRs because of their curtain shutter but with my CCD with their aperture blade style shutter, I have to take flats at least 3 seconds long otherwise the shutter leaves an artifact in the image. You can see the shutter blades. At 3 seconds they disappear. It may not be an issue with DSLRs but I thought I would mention it in case you can see shutter artifacts.

Greg.
Reply With Quote
  #8  
Old 04-11-2013, 03:22 PM
nebulosity.'s Avatar
nebulosity. (Jo)
Registered User

nebulosity. is offline
 
Join Date: Jul 2013
Location: Cecil Plains QLD
Posts: 1,228
The lights and darks were taken at the same ISO(1600) and exposure, the flats and bias' were different, do you think I should keep the same ISO for everything?

The flats were taken pointing at the western sky and it looked pretty even but as I got a 10x15 degree FOV it would probably be a good idea to put something over the front to make sure.

Does it matter what colour the flats are? And should I go by the camera Histogram or what it says on the computer? The both seem to be different.

I got a pic of a single Light, Flat, Dark and the processed image in B&W, seems to be lots of dusty stuff, if you haven't already guessed its the Hyades cluster.

Cheers
Jo
Attached Thumbnails
Click for full-size image (IMG_7760.jpg)
93.7 KB35 views
Click for full-size image (IMG_7940.jpg)
23.4 KB36 views
Click for full-size image (IMG_7701.jpg)
20.4 KB25 views
Click for full-size image (done 2 (2).jpg)
195.8 KB37 views
Reply With Quote
  #9  
Old 04-11-2013, 03:27 PM
rmuhlack's Avatar
rmuhlack (Richard)
Professional Nerd

rmuhlack is offline
 
Join Date: Nov 2011
Location: Strathalbyn, SA
Posts: 916
Quote:
Originally Posted by nebulosity. View Post
...I'll give a more detailed description of what I did, so I got the camera going around 6 o'clock (my cooled, modded 350d) ...
The camera would have sat on around -8 most of the night I think....
given you're cooling your DSLR to those sorts of temps, are you sure that what you're seeing isn't due to condensation/icing...?
Reply With Quote
  #10  
Old 04-11-2013, 03:52 PM
nebulosity.'s Avatar
nebulosity. (Jo)
Registered User

nebulosity. is offline
 
Join Date: Jul 2013
Location: Cecil Plains QLD
Posts: 1,228
Quote:
Originally Posted by rmuhlack View Post
given you're cooling your DSLR to those sorts of temps, are you sure that what you're seeing isn't due to condensation/icing...?
Hmm, good question, although I don't think so because I've rehoused the camera in a fully airtight case and all the subs look quite clean.
Reply With Quote
  #11  
Old 04-11-2013, 04:24 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Quote:
Originally Posted by nebulosity. View Post
The lights and darks were taken at the same ISO(1600) and exposure, the flats and bias' were different, do you think I should keep the same ISO for everything?

The flats were taken pointing at the western sky and it looked pretty even but as I got a 10x15 degree FOV it would probably be a good idea to put something over the front to make sure.

Does it matter what colour the flats are? And should I go by the camera Histogram or what it says on the computer? The both seem to be different.

I got a pic of a single Light, Flat, Dark and the processed image in B&W, seems to be lots of dusty stuff, if you haven't already guessed its the Hyades cluster.

Cheers
Jo
I would keep the ISO the same. Others who do regular DSLR imaging may chime in here. But as a basic you want the flats/darks to be under the same conditions as the lights with regards to temperature and focus, not rotate the camera, leave any filters in place.

I have always put a white tshirt over my scopes for dusk flats so I don't take a photo of a cloud or star which obviously was not in the image. As far as which colour - I don't know there. You could try making the exposure black and white by desaturating it and see if it makes a difference. It might. A flat is correcting for unevenness of illumination not colour so something blue may throw it off. I'd experiment doing it both ways and see how it goes.

As far as which histogram I think I would work off the camera's one.
Just in case the computer is applying some adjusment in the conversion from RAW. Many latest cameras tend to apply hot pixel suppression and possibly some noise reduction smoothing.

Greg.
Reply With Quote
  #12  
Old 04-11-2013, 06:07 PM
nebulosity.'s Avatar
nebulosity. (Jo)
Registered User

nebulosity. is offline
 
Join Date: Jul 2013
Location: Cecil Plains QLD
Posts: 1,228
I aligned the RGB channels in my flat frame and did a comparison with the original flat, didn't change the colours of the light frame, but removed the dust a lot better.

1 Original flat.

2 RGB aligned flat.

3 Light with original flat removed.

Light with RGB aligned flat removed.

Cheers
Jo
Attached Thumbnails
Click for full-size image (IMG_7949 (1).jpg)
23.5 KB18 views
Click for full-size image (Master2.jpg)
21.8 KB19 views
Click for full-size image (1.jpg)
91.5 KB18 views
Click for full-size image (2.jpg)
93.7 KB24 views
Reply With Quote
  #13  
Old 04-11-2013, 07:31 PM
Octane's Avatar
Octane (Humayun)
IIS Member #671

Octane is offline
 
Join Date: Dec 2005
Location: Canberra
Posts: 11,159
You should be taking your flat frames at the lowest native ISO of your camera. In Canon's, that's ISO-100. The flats are only picking up the anomalies in your imaging train.

It's what worked for me when I was doing DSLR work.

H
Reply With Quote
  #14  
Old 04-11-2013, 07:33 PM
Octane's Avatar
Octane (Humayun)
IIS Member #671

Octane is offline
 
Join Date: Dec 2005
Location: Canberra
Posts: 11,159
Doesn't matter what colour the flats are. But, grey works best.

You want to watch the exposure on the back of your camera's histogram.

You want it to peak somewhere between 1/3rd- to 1/2-way across the X-axis.

Use LiveView Exposure Simulation (if your camera has it) for working out exact shutter speed required. Or, just view the histogram as you take test shots.

H
Reply With Quote
  #15  
Old 08-11-2013, 05:57 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Not sure what you mean by RGB aligned as this is a colour DSLR so the output is already colour. Do you split the image into colours and process separately? (Some DSLR imagers do that).

The dust is removing nicely but there still seems some vignetting present, at least in the small thumbnails.

Did you desaturate the flat to get a mono? I was wondering about that the other day from this post. If you use a flat that is mostly blue and flats are dividing would'n't that weaken the blue channel in the image?
Perhaps you can do a test and find out. Perhaps desaturating your colour flat may work. There are DSLR specific astro sites like digitalastro yahoo group or Images Plus discussion group where these sorts of things no doubt are debated.

Also you need to clean your camera - that's pretty dusty and cleaning your gear regularly is a lot easier than flats!

My current procedure for flats is to take several and then max/min clip combine in CCDstack then when I apply them to the image I select the bias subtract option.

There is a paper by Richard Crisp on flats. He is a semi conductor engineer and an expert in this field. Apart from the complex maths etc (best to skip over that unless you are into maths) it is quite informative.

He basically states that with CCDs (it should be similar for DSLRs) that they are very linear (meaning they respond in a predictable straight line type way at different light levels). He recommends taking a quite bright flat and a flat dark subtract (a dark of the same duration and temp as the flat) to give the minimum amount of noise the flat introduces into the image.

Flats can be a tricky area of processing so its worth a thread by itself as there are bound to be different approaches from various imagers here.

Greg.

Greg.
Reply With Quote
  #16  
Old 08-11-2013, 08:33 PM
nebulosity.'s Avatar
nebulosity. (Jo)
Registered User

nebulosity. is offline
 
Join Date: Jul 2013
Location: Cecil Plains QLD
Posts: 1,228
Quote:
Not sure what you mean by RGB aligned as this is a colour DSLR so the output is already colour. Do you split the image into colours and process separately? (Some DSLR imagers do that).
What I did is adjust the RGB levels until each was exactly the same, this is why its now white.

Quote:
The dust is removing nicely but there still seems some vignetting present, at least in the small thumbnails.
Yes I wondered about that, are flats supposed to remove it as well?

Quote:
Also you need to clean your camera - that's pretty dusty and cleaning your gear regularly is a lot easier than flats!

My current procedure for flats is to take several and then max/min clip combine in CCDstack then when I apply them to the image I select the bias subtract option.

There is a paper by Richard Crisp on flats. He is a semi conductor engineer and an expert in this field. Apart from the complex maths etc (best to skip over that unless you are into maths) it is quite informative.

He basically states that with CCDs (it should be similar for DSLRs) that they are very linear (meaning they respond in a predictable straight line type way at different light levels). He recommends taking a quite bright flat and a flat dark subtract (a dark of the same duration and temp as the flat) to give the minimum amount of noise the flat introduces into the image.

Flats can be a tricky area of processing so its worth a thread by itself as there are bound to be different approaches from various imagers here.

Greg.
Good to know, thanks

Jo
Reply With Quote
  #17  
Old 08-11-2013, 08:35 PM
nebulosity.'s Avatar
nebulosity. (Jo)
Registered User

nebulosity. is offline
 
Join Date: Jul 2013
Location: Cecil Plains QLD
Posts: 1,228
Quote:
Originally Posted by Octane View Post
Doesn't matter what colour the flats are. But, grey works best.

You want to watch the exposure on the back of your camera's histogram.

You want it to peak somewhere between 1/3rd- to 1/2-way across the X-axis.

Use LiveView Exposure Simulation (if your camera has it) for working out exact shutter speed required. Or, just view the histogram as you take test shots.

H
Thanks H, have been doing what you said and it seems good.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 08:36 PM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Testar
Advertisement
Bintel
Advertisement