Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Flats issue

  • Please log in to reply
16 replies to this topic

#1 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 08 August 2020 - 05:54 PM

OK i have a conundrum i am, hoping you all can help me with.  I use a flat box i created with three of these in it (4x4 square):

 

https://www.frys.com...FromSearch=true

 

I was using software to control the scope and camera at the time, MaximDL, and everything else ran fine except the flats.  Normally it works fine for doing ISO 800 or lower shots but last night i tried to take a series of flats at ISO 1600 and the pixel values were all over the place. Whats odd is it started at the max pixel values i wanted for the flats on the first shot but then shot up from there and then started pogo-ing all over the place.  I am trying to take images at 65% of full well capacity. On my DSLR full well is 65,500 and 65% works out to around 42,000 or so.  I try to get it in the 38,000 to 42,000 range most days.  I have a Canon 60D (FS modded) and was using a LPR filer from Orion. I even reduced the time to minimum and STILL it showed the same issues. I managed to get good shots of M16 and M15 with it but the flats proved to be an issue and without them the images look awful.  my scope is still setup and i plan to do more imaging tonight but hope to find a solution to the flats issue before then.    What could be the problem here?    i brought the battery downstairs to charge afterwards figuring it was low and it was. Could the battery life have caused this?


Edited by bignerdguy, 08 August 2020 - 05:55 PM.


#2 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 08 August 2020 - 06:30 PM

Power supply is the main cause of headaches. Get a good, constant supply. You need consistency.

 

It sounds like the light sources you bought may have possibly flicker problems.. Check that.

 

You should have enough diffuser between the camera and light panel. I use regular printer paper. Two sheets at lowest brightness. You can adjust exposure time to get the range you want. The light must be even across the flat light source. I use a flat panel I bought on Amazon.

 

What are you analyzing the flats with. PI is excellent. Look at the histo and make sure you have the proper bit depth for your camera in whatever you use.

 

Can yo upload and link a flat here? I'll run it through PI and see what you have.


Edited by Dynan, 08 August 2020 - 06:31 PM.


#3 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 08 August 2020 - 07:53 PM

I have a custom made lightbox and it does have two pieces of copy paper between it and the camera. The led's don't normally flicker since the ISO 800 flats are usually consistent. I will have to reinstall DropBox on my PC to upload the file.  the Flats are all FITS files and are too large to upload here.  I will post the link tomorrow once i have done that. In fact i just went looking for a flat to post and noticed i had some ISO 1600 flats i had forgotten about.  Seems  the variation on these wasn't too far out there but there was a little.  there were one or two that were too far into the higher values to use and i may have to go back and reprocess my master flats for that session without those included.  So maybe it WAS the battery on my camera.  i would be using a external supply but i don't have one that works reliably at the moment.  the only one i have has issues so i use a fully recharged battery or two.  I will try doing the flats first thing tonight and see what happens.  Its almost time to get out there, thankfully all i need to do is pull the cover off everything and make sure nothing melted in the 100+ heat from earlier today...


  • Dynan likes this

#4 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 09 August 2020 - 12:36 PM

OK so last night i used a fresh battery just to do the flats.  I got the same result, each separate image was at a different pixel value, some really high others really low.  i decided to take 100 flats and average them all and see what i get.  the final image was exactly where it needed to be in pixel value: roughly 39,000.  This seems to work fine for my processing.  So it looks like i am going to need to take a lot of flat images and average them all to get the value i want.  The LED's shouldn't be flickering as there are not any major electronics running them, simply a 12v source attached with no controllers so not sure why the image levels are varying so widely.  



#5 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,731
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 09 August 2020 - 01:22 PM

What is your exposure length? You probably need to get it up to a second or so.

Mark

#6 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 09 August 2020 - 01:37 PM

Make sure you have no light leaks, Can you do this in a dark room with your light panel as the only light in the room?

 

What is your exposure length? You probably need to get it up to a second or so.

Mark

I expose flats at 3.3 seconds.



#7 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 09 August 2020 - 06:21 PM

What is your exposure length? You probably need to get it up to a second or so.

Mark

Um, 3.3 seconds would horribly over expose my flats to the point of over-saturation.  the light box is bright enough that a setting in MaximDL of .0031 seconds is enough to properly fill out a ISO 800 shot.  i did do a test indoors with the equipment and a dark room and did the same last night outside, the results were the same. however i do notice the ISO 800 flats do vary a little bit but not as widely as the ISO 1600 ones do.  Even at a setting of zero seconds in MaximDL (sets maximum shutter speed as far as the camera is concerned) still gets the same values as .0031 seconds does.  Its weird.

 

Really here what i need is a Full well capacity of 65% of maximum for proper flats to work.  Last night when i took them they were all over the place so i decided to take a bunch, in this case 100 of them. i then averaged them together and got a single flat that sits around 37,000-39,000 value which works for what i need.

 

Here are links to two stacked Flat files, one color adjusted and one without color adjusting.  Both have a mean value somewhere around 32K-39K according to the information window in MaximDL.

 

Adjusted:

https://www.dropbox....justed.fit?dl=0

 

Non-Adjusted:

https://www.dropbox....er RAW.fit?dl=0

 

I have Dark and Bias adjusted both.

 

And here are about 30 of the raw flats if anyone wants to play with them.  I don't know what zipping them will do to file quality but if you want to there they are.  smile.gif

 

https://www.dropbox....-0030F.zip?dl=0


Edited by bignerdguy, 09 August 2020 - 06:34 PM.


#8 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 09 August 2020 - 06:58 PM

FIRST problem. Flats need to be GRAYSCALE!!! You need to be using RAW exposures.

 

Your shutter speed is WAY too fast for flats. At that speed, you're in the Bias Range.

 

Slap paper in between until you get to at least 2 seconds. Then upload a flat and let us look at it.

 

Here's what PI says about your flat:

 

FLAT A1.jpg

 

You're getting 600-ish edu out of 17 Million! No nearly enough light for a flat.

 

I don't know what software you have, but you need to get the flat to greyscale (black and white).

 

The histo should have your three peaks (RGB) in the middle of the histogram.

 

Try these things and post again!


Edited by Dynan, 09 August 2020 - 06:59 PM.


#9 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 10 August 2020 - 09:41 AM

Actually i believe the software converts it to mono as it integrates it.  The raw flat has color as that is what i shoot in but the program converts it to mono as it applies it.  I can also manually convert it to mono if needed.

 

Question: should i be shooting in greyscale to begin with on the flats?



#10 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 10 August 2020 - 09:54 AM

What you uploaded was NOT greyscale.

 

Yes, shoot greyscale if possible. Flats are useless as RGB.



#11 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,731
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 10 August 2020 - 10:11 AM

Question: should i be shooting in greyscale to begin with on the flats?

The terminology you are using is potentially confusing.  You should be shooting your flats as raw and you should also be shooting your lights as raw.  The lights should be calibrated by the flats before you debayer.

 

I didn't take a look at your raw flats (are they actually raw?) because I don't have the bandwidth to download 3Gb of data.

 

Mark


  • Miguelo likes this

#12 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 10 August 2020 - 12:37 PM

Here is a single flat RAW image file if you want to look at it.  Yes i shot this in color.  Sorry about the large download i figured a few of you might want to try to combine them and see what results you got from doing so.. 

 

https://www.dropbox....-0001F.fit?dl=0

 

When i shoot images i have a lot of options to choose from.  I use JPEG mono when trying to find the object, and RAW Color when actually shooting the photos.  JPEG mono takes less time to download and is faster when i want to locate and center an object.  i shoot everything else in RAW color but it gets saved directly to a FITS file by MaximDL so all the RAW data is still preserved and the FITS header with metadata gets added to it.   I will try next time trying to shoot it in mono and see what that does.  As for the length of time of the shot, the light box is pretty bright.  i have been trying to limit the brightness to 65% of full well capacity but the histogram i see in MaximDL looks different that what you are seeing in other applications.  Here is what i see on that same file when viewing the histogram:  

 

Screenshot.jpg

 

So trying to take a single shot any longer will max the histogram out completely and render it useless.  This is what i see when i look at the pixel values.  This is the information window in MaximDL from a point near the middle of the image:

 

Untitled.jpg

 

 

As you can see this one is a bit higher than the value i want however when i average all of my flats i get a value of 39k-42k which is about right according to everything i have read about this program.  Here is the histogram in Photosshop CS3:

 

Photoshop.jpg


Edited by bignerdguy, 10 August 2020 - 12:38 PM.


#13 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 10 August 2020 - 02:53 PM

Here is greyscale analysis of my flat. This is a n ASI533MC image, luminance only. (In PI you can extract Lum) Histo is showing all peaks close to half scale and ≈ half maximum for 14-bit.

 

FLAT A2.jpg

 

If you look closely you can see that the EDU is close to Half of Maximum (shown by the 14-bit dropdown).

 

I don't believe you can use RGB for flats.(Shown in the next post)


Edited by Dynan, 10 August 2020 - 02:56 PM.


#14 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 10 August 2020 - 02:55 PM

FLAT A3.jpg

 

PI is requiring Grayscale.



#15 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 13 August 2020 - 09:26 PM

I did a little testing indoors today with my flats.  I found several things out:

 

1. My reducer is the source of the horrible vignetting I am seeing.  Removed it and the bright center in my images went away.  Gonna have to find a solution for that. I have one of those Meade f/6.3 reducers which is really F/4 due to incorrectly manufactured optics.  That might be the cause of this.  I can easily image at F/10 but with a 2000mm Focal Length that means it is heavily magnified and wont be good for too many DSO's.  I am going to see about getting a replacement but for now will continue to use it.

 

2. I had wondered why my flats and lights were so badly red as in past usage they were blue since i use LED lights in my flat box.  LED's, especially white ones are actually blue/white and should show up in my images with a blue shifted tint.  However as you all saw the flats i posted were quite red.  Turns out i had the wrong filter in position #1 on my filter wheel.  I actually had a V-Fring filter in that position but had it labeled in my astro software as the LPR.  The LPR allows more blue in the image and the V-Fringe a bright Red.  So i have since fixed that on the filter wheel and will see what happens next time i do imaging.  The weather on Friday of this week (tomorrow) is supposed to be possibly OK till around midnight so i will take it out and see what happens.

 

3. Someone in another thread suggested people using Canon Cameras use the AV function on the camera to take flats as it auto detects the light level and sets the shutter correctly.  When i used it for the flat with the V-Fringe it was setting a setting of 1/250th.  The other thread suggested if it sets it to 1/4000th then the light source is too bright so i think the light box is OK.  When i used the LPR filter the setting went to 1/125th which is even better as it means the LPR was working fine.  Also the light levels didn't vary quite so widely on the LPR as they did still on the V-Fringe.  I think the IR levels were throwing the sensor for a loop (i don't believe V-Fringe blocks IR?).



#16 Dynan

Dynan

    Skylab

  • *****
  • Posts: 4,396
  • Joined: 11 Mar 2018
  • Loc: NOLA

Posted 14 August 2020 - 06:00 AM

Good catch(es)! Backspacing is critical for reducers and flatteners.


  • bignerdguy likes this

#17 bignerdguy

bignerdguy

    Messenger

  • -----
  • topic starter
  • Posts: 483
  • Joined: 31 Oct 2019
  • Loc: Lewisville, TX

Posted 14 August 2020 - 07:20 AM

Good catch(es)! Backspacing is critical for reducers and flatteners.

Yeah i know.  the funny thing is i measured the back spacing quite a while ago and it is really close to where it should be.  I mean it is within 2mm of perfect for the reducer.  it was the closest i could get it with the hardware i use.  if i use the LPR i believe the vignetting will be low enough to cancel out when i process the photos.  I plan to take my scope out tonight and see what happens.  Weather outlook has gotten better over night so hopefully i can spend all night on imaging and get better results.


  • Dynan likes this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics