›

Darks and Bias - old question for a new era???

auroraTDunn
in PixInsight
Is there any advantage of taking more then 64 darks and 64 bias frames with modern CMOS camera's noting I'm using a ASI6200MM Pro.
I spend months gathering data on singular projects and don't want to skimp out on such important calibrations. Just not sure iif it buys my anything. But if it does then its worth the disk space and processing time since I have already spent more then 100 hours on some of my projects and have one where I'm working on over 100 just for one filter!

Note I also use 64 flats, I'd take more of those two, but what if I don't get the same exposure times, thus some are 'brighter' then others?

Comments

  • BlockHeadBlockHead
    64?
    That seems excessive. When you take multiple frames you are making multiple measurements. It really depends on how noisy your camera/sensor is. Today sensors are not *that* noisy. So you are really looking for good statistics in minimizing the noise (so you are not adding it unnecessarily to  your data when you calibrate). There is math that you tell you how many frames is enough. For our purposes of pretty pictures (not measuring values to within 1%)- I would think 15-25 is pretty good. You can't hurt yourself by integrating more though. 

    Concerning flats- this is different. With flats you have many many photons! The noise is poisson from the photons and pretty much overwhelms the electronic signature (though you still need to calibrate raw flats with bias/dark). So there is an argument that you need *less* flats for our pretty picture sake. 7-15 is a really good number. 

    So...I would say take more of the biases and darks and be safe. 64 is excessive though. Flats, you can save some time with those. It doesn't matter with flats if some are brighter than others... the total photon count (when integrated) is all that matters... the more the better- but you can gather  A LOT of photons with flats quickly. 

    -the Blockhead
  • auroraTDunn
    You know, I don't even use Bias frames in WBPP as I don't optimize my master darks. Am I wrong in doing that, not optimizing my darks?  I keep forgetting that and yet I add them in anyways. Then ignore the fact that when I look at the diagnosis output and graph that I'm not using them (which is how I remembered tonight as I was getting setup to run WBPP tomorrow). I got into that habit, of using bias frames, a year or two ago when running WBPP on Windows was pretty much broken so switched to Siril and ended up using the bias in Siril.

    I'm not sure where/when I got so committed to jumping from my old norm of 24 to 36 and now 64 frames. It was something I read years ago. But I really need to experiment with fewer frames as at 177MB per frame it adds up both is storage and compute time.

  • BlockHeadBlockHead
    You are correct. There are two uses for a bias. 
    1. Calibrate Flats
    2. Scale/Optimize Darks

    Do not do #2 in general and don't do it without understanding the effects of doing so. Never optimize darks to calibrate Flats.

    With respect to #1- if you have matched darks to calibrate your flats- this is not problem. However, most cameras today can use biases to calibrate all of the short exposures that you produce for flats. This is a savings in storage as well since you can just use a masterbias to calibrate your Flats of any (reasonably short) duration.

    -the Blockhead
Sign In or Register to comment.