Taking Flats with a DSLR

[Carole]

Well after doing DS images for a little while now I have found out I am doing my flats wrong, which may account for some of my poor results compared to the length of exposure.

I understand the principle about what it does and the orientation of all lenses etc.  What I am a bit unsure about is how to achieve this 1/2 bucket, and am reading conflicting info.

I have read that I should take the flats at 100 ISO in one place, and in another I have read it should be the same ISO as my actual images which is normally 800 ISO.  Also

Metering mode to average.  
(This ensures that the camera will expose for the entire illuminated field).

Just wondered what those using DSLR thought about ISO and Metering Mode.
Think I am going to make myself a light box. 

Carole

[mickw]

I've assumed that the object is to see any grot in the optics so that would mean leave iso same as image, just reduce exposure time so you can actually see the grot.


Sitting back waiting to be shouted at

[RobertM]

It shouldn't matter about the ISO or shutter setting; flats are all about removing artifacts of the optical system.  the electronics are dealt with by the Darks and Bias.

At the risk of invoking an argument on flats, you should really do your flats with a peak at 2/3rds and not half.  The result will be cleaner and better able to take account of any severe vignetting.  Since the sensor will still be in the linear part of it's range you are better off using the available signal to improve the signal to noise ratio.

Robert

[MarkS]

I agree with Robert.

Also, since ISO does not matter, use ISO 100 to give the cleanest (i.e. least noisy) flats.

Mark

[Carole]

I've assumed that the object is to see any grot in the optics so that would mean leave iso same as image, just reduce exposure time so you can actually see the grot.

Well so had I Mick, which is what I had been doing, but as no dust or vignetting had shown up at all on my flats so far I need to find out what I am doing wrong.  

I've found an article describing how to do it which I tried yesterday and the flats were more like what they should look like, but the flats had to be stretched quite a lot to show any grot.  This same article says do them at 100 ISO which is what I did yesterday, but I've also read elsewhere that it should be at the same ISO as the imaging subs, so now I don't know which is right.

Currently building a light box.

Just read your posts Robert and Mark before posting this.  Thanks, that's cleared that up. 

Carole

[Mac]

Also, since ISO does not matter, use ISO 100 to give the cleanest (i.e. least noisy) flats.

you should do your flats at the base ISO of your camera, which is either 100 or 200.

this is normally the lowest setting that is not Lo.

My Nikon is 200 but i have a 100 ISO setting which is shown as Lo1, this is achieved in the same was as the higher settings,
i.e. changing the amp gain.

which is what I had been doing, but as no dust or vignetting had shown up at all on my flats so far I need to find out what I am doing wrong

lets think about this.

Your camera is probably between 12-8Mp

unless you have a huge lump of crud in the optical system then you are not going to see this unless you magnify your image to 1:1
and then look across the whole of the image, bit by bit, so it is taking the image of the dust.

As for the Vignetting, as you said it is there but the difference is not normally noticeable unless it really is serious,
but as you change the  values it then shows up.

so your flats are working, just you cant see the bits subtle difference the computer can.

Mac.

[MarkS]

It's very difficult to see dust on a flat unless it is an extremely severe example.

To see the dust in your flat do the following:
Take the flat and perform a Guassian blur of around 20-50 pixel radius.
Subtract from the original flat (or use Photoshop difference layer with opacity around 90%)
All the big dark blobs that appear are bit of dust.

Mark

[Carole]

Thanks for your help every-one I think I know where I am going now.  

I think my original flats were over-exposed as there is no crud on them at all whichever way I stretch them, if any-one with more expertise than me wants to examine the previous Masterflat it's in my dropbox on:
http://dl.dropbox.com/u/5730788/MasterFlat_ISO800.tif

The ones I did yesterday following new instructions do have some dust particles and vignetting showing on them after a fair bit of stretching, so I hope I'm now heading in the right direction.  

I finished the light box and tried it out.  I works OKish, need to experiment a bit more, the flats I did yesterday with the home-made diffusion mask into the daylight sky worked better.

This is the one I took yesterday after a lot of stretching to show the dust particles and vignetting.  

[Mac]

This is the one I took yesterday after a lot of stretching to show the dust particles and vignetting. 

See it is working.
You dont need to see the image you have stretched.
All you need to know is it contains the information about the crud,
Same as the bias images, they will be black, but unless you stretch them to kingdom come
you wont see the information contained within (which you are not supposed to see.)

Might update the RTFM post to include these in a little more detail.

Mac

[MarkS]

The best way to produce a flat is to imge the dusk sky (tonight was perfect) because it is very distant and diffuse.
Next best thing is a diffusion plate over the end of the scope.
A home made light box is a thing of last resort - unless reasonably sophisticated it will never provide even illumination.

The only comment I have about your excellent flat produced yesterday is that the bright patch is not central in the CCD.  This almost certainly indicates that your camera is not sitting centrally in the light path.  Could be a drooping focuser or a screw not tightened properly.  Whatever it is, it probably means that the camera is not "square on" and is likely to cause a little added distortion to stars at the extreme left or right.  This might be linked to the chromatic aberration you were seeing on stars central in your field - maybe both the focal reducer and camera are not square on.

[Rocket Pooch]

Opps I thought you meant this http://tinyurl.com/33j62qu

[Carole]

the bright patch is not central in the CCD.  This almost certainly indicates that your camera is not sitting centrally in the light path.

Yes I noticed that.  
I'm sure all the screws are tight.  
I'll have another look at it.

You dont need to see the image you have stretched.

Yes I thought that might be the case, but I just wanted to see the crud for reassurance that the flats were taken OK.  

Opps I thought you meant this http://tinyurl.com/33j62qu

Carole

[RobertM]

Opps I thought you meant this http://tinyurl.com/33j62qu

I bet there's a lot of crud in those flats...

[PhilB]

Come on, Chris, you know better than that. The http://tinyurl.com/33j62qu link clearly shows a STACK of flats or MASTER flat 

Mark, I'm interested in your dislike of light-boxes. I was considering making one of these, largely because in their image processing book, Richard Berry and James Burnell state "....light-box flats are by far the best for amateur astronomers..." However, your method clearly works very well. Could you put a little more flesh on the bones please?

[Rocket Pooch]

Come on, Chris, you know better than that. The http://tinyurl.com/33j62qu link clearly shows a STACK of flats or MASTER flat 

Sir,

With refrence to your comment above I would at this point like to say that this is indeed a single flat frame, however, it has very high signal to noise ratio.

So ner!

Yours Truly




Mr Trellis