how much can you rely on the FWHM from CCDStack

[Somnium]

i was looking through some of my data i noticed something strange. some of the images were really bad but when i look at the FWHM score in ccdstack it isnt represented. for example, both the attached images are take with the same filter, the first had a FWHM of 7.9 and the second was apparently worse at FWHM 7.96. does anyone know the reason for this?

[rally]

Cant be sure, but for a reliable FWHM the star being analysed needs to be in a clear background - yours is full of nebulosity - so maybe it chose a poor candidate star/s?
I agree, visually the lesser one is much worse than the other which doesnt make much sense.
I'd be throwing away the subs over 3 or 4 ! - so 7.9 becomes a moot point.

Different programs use different search and calculation algorithms too - so you can expect different results between say MaxIm and CCD Stack
Why not examine each image with MaxIm - pick a single star in a clear background area and see what you get.
It could be that the bloated FWHM was being confused with the nebulosity background and so the algorithm excluded the wider parts of the bottom of the stellar profile ! - Best explanation I can offer, sorry.

[Somnium]

no worries, i just thought it was interesting. the FWHM here is pixels, not arc seconds. 7.9 works out to be 3.5 arc seconds. i will be lucky to get images in the 2 arc second range due to my horrible seeing.

i did also notice that when i register the image, the FWHM can change dramatically ...

[rally]

I meant arc secs not pixels - so your subs are safe !

FWHM is usually defined in arc secs although some packages can report it in either, if you know the image scale, pixels would yield the same value after basic some maths.