Author Topic: Hardware vs software binning  (Read 1592 times)

Offline darrynl

  • Newcomer
  • Posts: 6
    • View Profile
Hardware vs software binning
« on: 2016 May 04 06:53:49 »

I am narrow band imaging faint objects with a SBIG STL-11000M using iTelescope. For each filter I'm doing 7 images of 10 minutes. Increasing the exposure time is not an option.

Binning seems to really help increase the signal. With this CCD is there any advantage to using hardware binning over software binning?

thanks in advance


Offline gvanhau

  • PixInsight Old Hand
  • ****
  • Posts: 345
    • View Profile
Re: Hardware vs software binning
« Reply #1 on: 2016 May 04 11:17:10 »
I'm no an expert on this, but as far as I know, there are two advantages in using HW binning vs Softw.

The first is that in Hw binning you only have once the readout noise, in softw binning you have the readout noise from each pixel you sum up.
The second advantage that you may have, is that fractions of ADU that are lost in soft binnin may be counted up by HW binning.

Geert Vanhauwaert

Offline mschuster

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1087
    • View Profile
Re: Hardware vs software binning
« Reply #2 on: 2016 May 04 17:15:18 »

You might try noise measurement using DarkBiasNoiseEstimator script. Capture two good unbinned bias frames and another two hardware binned bias frames. Run software binning on the unbinned frames to generate a third pair. Run the script on each pair, note the (detector) Temporal (Gaussian) noise value in DN, and convert to e- using the appropriate gain value.

On my detector (KAF-8300) I measure about 8, 12, and 16 e- Gaussian noise respectively on the three pairs (binning here is 2x2). Hardware binning has less noise than software binning, but not as low as unbinned.

On your relatively short exposure narrow band frames, detector noise likely dominates sky background noise. Assuming this is the case, with my detector, hardware binning gives a 16/12 or 4/3 factor improvement in dim target SNR over software binning.

Since detector noise also dominates dim target shot noise, it would take a 4/3 longer total exposure with software binning to equal the SNR of a hardware binning exposure.

« Last Edit: 2016 May 05 06:54:20 by mschuster »