Using FWHM scores to compare telescope performance

Somewhere to share your expertise in using SharpCap
Post Reply
Posts: 309
Joined: Mon Aug 26, 2019 5:50 pm

Using FWHM scores to compare telescope performance


Post by timh »

A simple observation on the interpretation of SC FWHM values when using an OSC that caught me out - which I think (please do challenge if wrong!) that I have now interpreted correctly? Probably not any sort of revelation to most folk but just in case...

Here the short summary for those who don't want to read the entire saga below -

The process of debayering OSC frames (as I believe SC does on the fly) blurs them slightly and increases FWHM values (because interpolation adds signal into neighboring pixels?). This relative increase in blurring upon Debayering is greater in undersampled than in oversampled frames.

Corrollaries would seem to be that 1) mono cameras will tend to offer better sharpness than OSCs in wide view undersampled images and conversely, 2) when using an OSC there is a 'sharpness' case for slightly oversampling?

Below are the observations that led me to the above via trying to use FWHM to understand why my refractor set up was apparently underperforming in terms of sharpness relative to a Newtonian.


For a long time I puzzled over the fact that the Sharpcap FWHM (pix) values for frames from my 110 mm Apochromatic refractor and from my 200 mm Newtonian were inconsistent. Using the same mount - with guiding at usually better than 0.7 arcsec RMS - the conversion to FWHM arcsec (FWHM_arcsec = FWHM_pixels x ( pixel_size_um / focal_length_mm) x 206.3) yielded values typically in the range 2.5 to 3.0 arscec for the Newtonian and about 3.5 to 3.9 arcsec for the refractor.

Newtonian = f 1000 mm, camera pixel size 4.63 uM --> ~0.95 arcsec/ pixel (F 5.0)
Refractor = f 616mm, camera pixel size 4.63 uM --> ~ 1.54 arcsec/ pixel (F 5.6)

So for my typical seeing (~ 2.5 arcsec ?) the Newtonian samples at 2.6X and the refractor slightly undersamples at 1.6X

Given that the FWHM arcsec score results from addition in quadrature of the seeing + mount + optics I at first suspected some nasty problem with the refractor optics (i.e. since the mount was the same and any seeing effects would not be consistent).

Then the following observations helped to clarify things..

1) Refractor FWHM values dropped by 30% from ~ 3.8 to ~2.9-3.0 ' when using a MONO rather than OSC camera
2) On average Newtonian FWHM values also dropped but only slightly from ~ 2.75 to ~ 2.5 when switching from OSC to MONO
3) Most tellingly, I then measured the FWHM of saved (undebayered) refractor OSC frames to be about 30% lower than after debayering them in PixInsight (at which point the values became about the same as originally indicated by SC)

So in the end it was good to see that everything worked out to be consistent -- i.e. that my average seeing must be somewhere around 2.5 arcsec -- and that even after adding the reducer/ flattener the apochromatic refractor was performing more or less as it should. Overall the Newtonian plus 4.63 um camera combination remains somewhat sharper - which probably comes down to its higher sampling rate.

User avatar
Site Admin
Posts: 8092
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK

Re: Using FWHM scores to compare telescope performance


Post by admin »

Hi Tim,

thanks for sharing - that all makes sense (and SharpCap does indeed debayer RAW images before calculating FWHM, since otherwise the grid pattern can cause issues with the calculation). It's nice that the measurements line up so well with the theory too :)


Post Reply