›

flat frame exposure

Van Grafton
in General
Rookie question here:  What is the proper exposure for flats?  If PixInsight calibrates/converts all image data into a 0 - 1 scale, it should follow that if I get a flat exposure whose histogram is exactly in the middle of the scale should read as a .5 value in PI.  However, I'm not seeing that correlation.  Every video/tutorial I've watched advises to ensure that flats expose to 1/3 - 1/2 from the left on the histogram scale.  When I follow that instruction, I see .05 values on my flats.  I recently attempted M31, and got the most complete data set I've captured since beginning this hobby.  I didn't get stellar results, and received input that suggested that there was possible an issue with my flat frames, which, after looking closely, seemed underexposed. 

Seeking advice on the proper way to expose flats, and the proper way to verify within PI, if possible.

TIA

Canon 5D MKIII
Canon 70-200mm telephoto lens
Celestron Advanced VX mount

 

Comments

  • BlockHeadBlockHead
    Van,

    Yes, you should measure 0.3 - 0.5 for your flats to be properly exposed. It isn't critical though... if you have 0.2... it is OK! But 0.05 is not enough light. 

    There is a little math issue though. It gets a little complicated.
    Your image could be a 16-bit image (saved by your camera or the software that took the image). But your camera is only 14-bits. Do you see why this matters? This means the maximum value you would see in your image (from 0-1) is .25 ...and IF you expose for 0.5 natively this means you expect to see a value of 0.125 for GOOD flats. But your 0.05 is too low. You need to double your exposure time (approximately).

    It probably isn't the major source of your issue... but hopefully this brings some awareness to the meaning of the numbers.

    Again, to repeat- if your 0.05 flats are not calibrating the frames well (removing vignetting and dust)- the light level isn't the reason...there is another cause.

    -the Blockhead
  • Van Grafton
    Thank you for the reply!  I have attached the finished image here in case you would have time to examine it and let me know what problems are there.  At the end of the day, I actually like the image as it came out, but it's way off the mark in terms of astrophotography perfection, for sure!

    Happy Thanksgiving!
    Andromeda 1.jpg
    1447 x 965 - 447K
  • BlockHeadBlockHead
    HI Van,

    Based on this image, I agree there is an issue- but you need to widen your scope of consideration. It looks like the signature of a light leak or scattered light. Did you take your dark frames in the dark? What did you use to generate your flat fields?

    -the BLockhead
  • Van Grafton
    For the first image above, I capped the camera and shot the darks right after I finished with the lights.  I'm using a tracing panel for the flats, which for that image, were under-exposed.

    I attempted the target again and have attached the registered result, with DBE applied. I used 230 samples for the DBE process.  No color calibration has been done.  New flats were taken, and are giving me .14 average values based on your advice.  Darks were acquired in the same way as the first image, by capping the lens.  Blinking through all the new calibration images doesn't reveal anything that I can tell would be causing this issue.

    I did remove my clip-in CLR LP filter, on advise from a respondent to a post on the Cloudy Nights forum, who stated that, for broadband targets, those filters do more harm than good, and surmised that the halo I'm getting could be caused by that filter.

    New image integration:

    55 subs at 135s
    55 darks at 135s
    35 flats average value of .14
    60 bias frames at 1/8000

    Thank you for the input!  Is there a link to your Patreon that you could share?

     

    Andromeda2.jpg
    2314 x 1548 - 471K
  • Van Grafton
    Additionally, I am attempting the Photometric Color Calibration, and, since this last update to PI, I don't see several of the parameters in the process now that you mention in Fast Track training.  See attached screen shot for what I see.
    Screen Shot 2022-11-25 at 9.17.32 AM.png
    1004 x 1578 - 491K
  • BlockHeadBlockHead
    LOL
    I know what you see... we are all on the same page.
    I am currently making FastTrack Training *AGAIN* from scratch to capture all of the new changes and updates. In addition, I will be explaining SPCC as well.

    Please look for videos to be published....at least some... by the end of the day.

    -the Blockhead
  • Van Grafton
    :). Thank you again for the response!  Your tutorials have been THE VERY BEST I've found anywhere!  I'll keep an eye on the website for the new videos! Many, many thanks!
  • BlockHeadBlockHead
    Thanks... look below the original FastTrack for the work in progress.
    -the Blockhead
Sign In or Register to comment.