Jump to content

SkySurveyBanner.jpg.21855908fce40597655603b6c9af720d.jpg

How do you analyse a Flat??


Recommended Posts

As it says really - How do you analyse a flat frame to see if it is suitable for use as a flat.  I have bought one of those new fangled LED lightboxes and have taken some flats with it.  How do I check them out.  I've soon some folk show spectra and/or histograms of their flats but I'm not sure how they get them.   Any suggestions?

Here is a JPEG of one of them:

post-4502-0-39779100-1402772507_thumb.jp

Link to comment
Share on other sites

  • Replies 31
  • Created
  • Last Reply

I thought that is what it should look like - It was taken through my normal imaging set-up ie Canon 1100D and Evostar 120mm Achromat.  The vignetting should be there surely?  I agree is seems a bit low on the red channel and a bit heavy on the blue.

Am I right in thinking that the flats used by DSS with a DSLR colour image are monochrome?

Link to comment
Share on other sites

Hi Roger,

I'm not sure I follow what you're talking about here. Have you tried applying the flat ? Michael has shown that it at least looks like a reasonable flat. Use it and see ?

Secondly .... I went on your blog about a minute or so ago and Avast is claiming it has a virus attached. Could be a false positive I don't know,

Dave.

Link to comment
Share on other sites

Avast often shows a false positive on blogger blogs - so I'm told, I've never seen it as I use a different AV software.

As to the flats - I'm just trying to establish whether you can "define" what a good flat should be.  In the sense "can you measure it"?  I have seen folk display "histograms" claiming they show the "even illumination" of the flat.  I'm just not sure what they mean.

This is what PS shows as the three RGB histograms:

post-4502-0-65540900-1402775547_thumb.jp

Link to comment
Share on other sites

If your flat is nice and even with your refractor then it's probably duff. Every flat in every scope has to be different so I think I know what you mean but I'd dismiss it. Next time you image just take a reasonable number of flats and use them.

Re DSS. DSLR OSC images are reported as 16 bit grayscale.

Sorry about the Avast scare !!

Dave.

I was slow typing. ChrisH is correct

Link to comment
Share on other sites

I've just had a play with the contrast and I can say the flat isn't very flat, the illumination is uneven:-

post-4502-0-39779100-1402772507A_zpsabbe

Which is the whole point. In my view this hard stretch reveals a credible flat. It has recorded a believable image of the light path. The centre is brighter than the corners but there is no gradient top to bottom or side to side. These rarely exist for real so if you see them in a flat be suspicious.

If you took a flat that was flat there would be no point whatever in applying it. It would have no effect on the image because there would be nothing wrong with the light path. You can google tha flats of famous imagers to see how alarming they are! They look awful, but that is the point. What is awful in the flat is corrected by the flat. QED.

This is a flat from an APOD rig.

O%20FLAT%20web-M.jpg

Bad, isn't it? Don't panic, Mr Mannering!

Olly

Link to comment
Share on other sites

Must be me, it still doesn't look even, the right side is significantly lighter than the left - unless that is normal?

The bright centre is slightly offset, you are right, but even using high end focusers, in my case Tak and Feathertouch, I would say the OP is well within what seems credible. Of course, if the flats don't work then they don't but I see no real reason to be suspicious at this stage.

Olly

Link to comment
Share on other sites

As it says really - How do you analyse a flat frame to see if it is suitable for use as a flat.  I have bought one of those new fangled LED lightboxes and have taken some flats with it.  How do I check them out.  I've soon some folk show spectra and/or histograms of their flats but I'm not sure how they get them.   Any suggestions?

Here is a JPEG of one of them:

attachicon.gifFlat.JPG

It is all too easy to get bugged down with trying to analyse flats, darks , bias  etc. Your flat looks normal to me, it shows the unevely illuminated corners and some dust specks. Stack 50 or so of these and the master will correct a lot of the ills in your  subs. As for the histogram, depending on your camera, keep it between 2/3 to 1/2 way up the from left or If your capture software allows it aim for a value of the about 23000 ADU on the scale as you move the cursor around the center of the frame ( the brightest part ).

A.G

Link to comment
Share on other sites

Don't forget to calibrate your flats while stacking. They need dark subtraction, but the darks in question can just be a master bias or set of bias frames. For short exposures without the build up of thermal noise a bias and a short dark will be equivalent. Uncalibrated flats over-correct, as a rule, and produce inverse vignetting and inverse bunnies.

If using a reflector you might find that shorter flats exposed to maybe 15000 ADU work best. You just need to experiment sometimes.

Olly

Link to comment
Share on other sites

With my DSLR set-up I use "bias" frames taken at the shortest exposure time possible (1/4000 sec).  I'm not familiar with ADU values - i assume they are to help setting the exposure time.  i simply set the camera to "Av" mode and let it decide on the exposure time for the flats.  Using my new variable brightness LED screen I can get the camera to use exposures from 1/5 sec down to about 1/125 sec.  However from 1/125 sec and shorter I start to get black areas on the flat due to the shutter.

These are the test flats:

Flat at 1/5 sec

post-4502-0-37543400-1402829259_thumb.jp

Flat at 1/10 sec

post-4502-0-16985200-1402828919_thumb.jp

Flat at 1/60 sec

post-4502-0-55094800-1402828919_thumb.jp

Flat at 1/125 sec

post-4502-0-94979500-1402828919_thumb.jp

Flat at 1/200 sec

post-4502-0-30068400-1402828920_thumb.jp

Link to comment
Share on other sites

With a DSLR, setting it AV and letting it get on with it is the easiest way. You are aiming for a peak in the histogram 1/3 to 1/2 way across from the left hand side. ADU values are just the numerical measure of the histogram values. E.g. For my camera it runs form 0 to 65k ish so for a 'peak 1/3 of the way across' I aim for an average ADU of 22k. Same thing displayed differently.

Your flats look fine to me :)

Link to comment
Share on other sites

With a DSLR, setting it AV and letting it get on with it is the easiest way. You are aiming for a peak in the histogram 1/3 to 1/2 way across from the left hand side. ADU values are just the numerical measure of the histogram values. E.g. For my camera it runs form 0 to 65k ish so for a 'peak 1/3 of the way across' I aim for an average ADU of 22k. Same thing displayed differently.

Beware of DSLR histograms. The Canon 1000D, which is a 12-bit camera, only displays 8 bits in its histogram.  In particular it drops the top two, so if you are on the right hand side then you have about 1024 counts, whereas the full-well is about 4096.  If you only exposure to a third of the way across you are wasting vast amounts of well depth.

NigelM

Link to comment
Share on other sites

Beware of DSLR histograms. The Canon 1000D, which is a 12-bit camera, only displays 8 bits in its histogram.  In particular it drops the top two, so if you are on the right hand side then you have about 1024 counts, whereas the full-well is about 4096.  If you only exposure to a third of the way across you are wasting vast amounts of well depth.

NigelM

I didn't know that, thanks :) I had worked out that I can expose to at least half way and still retain reasonable star colour, so it's nice to know there is a sensible reason for that.

Link to comment
Share on other sites

I'll guess that you didn't calibrate the flats with their own darks?  You have to do this, but you don't need to take dedicated darks. A master bias will do fine. Just take 20 dark frames of the shortest exposure possible, average combine them, and that is your master bias. Use it as a dark for flats and it should solve your problem. Uncalibrated flats usually over correct, which what is happening to you here. 

Olly

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.