Default cflux / flux vs sep.sum_* methods

[RReverser]

ctest/test_image.c - which I assume is supposed to be a canonical usage example from C - is doing those steps in order:

• sep_background
• sep_bkg_array
• sep_bkg_subarray
• sep_extract
• sep_sum_circle

However, I've noticed that on practice sep_extract already populates cflux + flux fields (and the documentation confirms that - https://sep.readthedocs.io/en/v0.5.x/api/sep.extract.html), so I guess my questions are, 1) which model does those default flux values correspond to and 2) correspondingly, under which circumstances I would want to invoke one of the sep_sum_* functions instead of using those values populated by sep_extract?

Apologies if it's supposed to be obvious, but I haven't found an answer or mention of those fields in https://sep.readthedocs.io/en/v0.5.x/apertures.html.

[kbarbary]

Both the output of sep.extract and the sep.sum_* functions are referred to as "flux", so this is indeed confusing. The cflux and flux values from sep.extract are the sum of the member pixels of the object, which in general is some set of connected pixels above the detection threshold, but can have arbitrary irregular shape. The sep.sum_* functions are the sum of all pixels within an aperture, regardless of their value and regardless of whether they belong to an "object" or not. Does that help?

[RReverser]

Somewhat, yes, thanks... Can you provide a rough example where those values would significantly disagree? Would it be for "smudged" stars?