Author Topic: How to determine if you are taking the best flats  (Read 1768 times)

Offline rtemen

  • Member
  • *
  • Posts: 56
    • View Profile
How to determine if you are taking the best flats
« on: 2017 November 29 06:31:35 »
I have been trying to learn the best way to determine that I have suitable flats.
I have a Canon T3i that I use for my shots.  I have a lightbox for taking my flats.

I seem to be finding two different philosophies while trying to google info on determining the best flats.

1. Take flats so that the histogram hits at about 1/3 of the way from the left.
2. Take flats so that the median just a touch over half the maximum, around 51%. (this data is in the Statistics for the shot in PI)

What I seem to find is that the two methods cannot meet. 
For example, if I put the histogram at 1/3 of the way, the percentage is like 80% to 90% or more.
Or, if I start changing the exposure to bring the percentage down towards the 51%, then the histogram is off the scale to the right.

So, what is the proper way to determine that I have the best exposure for my flats?

Rich

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4471
    • View Profile
Re: How to determine if you are taking the best flats
« Reply #1 on: 2017 November 29 08:44:59 »
the most correct way is to build a light source for which you can control the brightness. then, keeping your exposure length constant, increase the brightness of the light iteratively and take a series of images at each brightness. then plot the # ADUs at the center of the frame as a function of the light intensity. the plot should be linear, but at some point at the brighter end of the light, the plot should depart linear behavior.

that is the point at which the sensor becomes nonlinear, and you should always strive to have the brightest part of your flat below that ADU value.

assuming you can change the brightness of your lightbox, you should be able to easily do this.

the problem with the back-of-camera histogram is that it is post-gamma stretch and so will wildly overestimate the brightness of the frame. dont forget if that you open your CR2s as pure raw they can only ever have 16-bit values between 0 and 16383 - canon sensors are 14-bit and the top 2 bits of the pixel value are always 00 when opened as pure raw. so when viewed as a floating point number, you will see 0.0 - 0.25 with 0.25 being full saturation.

rob


rob

Offline rtemen

  • Member
  • *
  • Posts: 56
    • View Profile
Re: How to determine if you are taking the best flats
« Reply #2 on: 2017 November 29 09:45:33 »
Hi, Rob. Thanks for the quick reply.
Holy cow, I am so much a newbie at so much of this, I am embarrassed to admit.

What is   # ADU at center of frame  and where/how do I find that info?  What does ADU stand for?

I am able to vary the brightness of my lightbox.  Currently it is just at 1650 Lumens, 4,000K color temp.
Is this default value too bright? I can easily enable the box to control brightness.

Rich

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4471
    • View Profile
Re: How to determine if you are taking the best flats
« Reply #3 on: 2017 November 29 14:04:22 »
ADU means "analog-digital units" and means the raw pixel output of the camera. most astro cameras have 16-bit A/D converters so you get 16 bit numbers out of the camera. ADU is also sometimes referred to as "DN" (Data Number). canon cameras have 14-bit A/D converters as mentioned above.

PI normalizes to a floating point number between 0-1.0 by taking the ADU count for a pixel and dividing it by the maximum possible integer for the bit depth of the image. you can convert back to ADU by multiplying whatever you see in the readout by 65535 if you are talking about 16-bit integers. or you can switch the readout cursor to 16-bit integer, which is easier.

by center of the frame i mean the center of your flat images. generally speaking all telescopes suffer some kind of vignetting at the edges of the frame, so the brightest part of the frame is always near the center. since it will be brightest, that's where the flat will become overexposed first. it's probably OK to just eyeball the position, since you're going to want to back off from the point of nonlinearity by a good amount just to be safe. you can make a preview over the middle of the image and use the Statistics tool to see the max value in that area.

as for whether or not the lightbox is too bright, that depends on the sensor and how the brightness is controlled. most of these panels run on a 10-20khz square wave, so if your exposure is too short, your flat may not have even illumination. so you want to make sure your panel is dim enough so that you can take an exposure that's longer then multiple cycles of whatever is driving the panel. 20khz has a 50 microsecond period so to be really safe i'd want to make the panel dim enough that you get 1000 cycles, or 50mS (or 1/20s)

rob

Offline dld

  • PixInsight Addict
  • ***
  • Posts: 114
    • View Profile

Offline sharkmelley

  • PTeam Member
  • PixInsight Addict
  • ***
  • Posts: 233
    • View Profile
    • Mark Shelley Astrophotography
Re: How to determine if you are taking the best flats
« Reply #5 on: 2017 November 30 02:48:56 »
If you take DSLR flats with the back-of-camera histogram 1/3 from the left then your flats will be very underexposed i.e. a lot noiser than they need to be.  I take mine so the histogram is well over to the right.  Typically I increase exposure until it is just saturating in the middle of the frame then back off by 1-stop.

Mark
« Last Edit: 2017 November 30 03:39:36 by sharkmelley »
Takahashi Epsilon 180ED
H-alpha modified Sony A7S
http://www.markshelley.co.uk/Astronomy/

Offline pfile

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 4471
    • View Profile
Re: How to determine if you are taking the best flats
« Reply #6 on: 2017 November 30 11:25:27 »
yeah despite all the OCD testing i've outlined above, what i used to do when using a DSLR was just let the camera compute the exposure, then manually dial in that exposure plus +2EV, and that came out fine.

rob

Offline rtemen

  • Member
  • *
  • Posts: 56
    • View Profile
Re: How to determine if you are taking the best flats
« Reply #7 on: 2017 December 12 06:44:34 »
Hi, all.
Thanks for all of the help.

Rob, I did build myself a variable light box, and I am getting the proper exposures that everyone has been coaching me on.

Visit my page to see the lightbox that I built.
http://astronomy.richtemen.com/lightbox.html

Thanks again, everyone.
Rich