Stellar grouping issues
rachel wrote on Jul 18, 2006
I was wondering how PSF selects what stars it includes in a group and if there was any way this could be changed. The field I am working on contains a density of stars such that allstar doesn't seem to be able to subtract them. Some of the stars only have 1 pixel of separation. It leaves behind clusters of stars and this seems to be caused by some kind of grouping issue.
I have set psfrad=3 and fitrad=1.57 which is equal to the FWHM of the brightest, unsaturated star. Is there any other task or parameter that determines grouping of stars? In daopars, there is a parameter called maxgroup that allows no more than 100 stars fit per group (this can't be made any higher). Is there a possibility that because of this parameter and the proximity of the stars to one another, there is not a way to fit and subtract these stars?
Thank you.
-rachel
I have set psfrad=3 and fitrad=1.57 which is equal to the FWHM of the brightest, unsaturated star. Is there any other task or parameter that determines grouping of stars? In daopars, there is a parameter called maxgroup that allows no more than 100 stars fit per group (this can't be made any higher). Is there a possibility that because of this parameter and the proximity of the stars to one another, there is not a way to fit and subtract these stars?
Thank you.
-rachel
Philip Massey wrote on Jul 18, 2006
My own guess is that this is a problem related to sky subtraction, not grouping, For one thing, "allstar" does not use the standard grouping used by NSTAR
(and, I believe, by PSF). So, it might be an issue related to sky determination.
You might try setting daopars.fitsky to "yes".
But, if you want to experiment with the grouping issue, you can readily
set the maximum number of stars in a group to > 100. I don't think this makes
sense, but you can do it. Simply do a
daopars.maxgroup.p_maximum=300 (say)
and then you will be able to epar daopars and set maxgroup to a larger value.
But, I don't think this is going to get you where you want to go.
The only stars that are going to affect the PSF are those within a PSFrad. I've never heard of restricting this to only 3 pixels, but I sure hope you don't have
>100 stars within 3 pixels of a star you're picking for a PSF star!javascript:emoticon(':roll:')
Is there really no light coming from the brightest star you're interested in
beyond 3 pixels? If there is, you need to make the PSFrad to whatever this value is. Sounds like your data is horribly undersampled, and that daophot
isn't a good solution.
---phil
(and, I believe, by PSF). So, it might be an issue related to sky determination.
You might try setting daopars.fitsky to "yes".
But, if you want to experiment with the grouping issue, you can readily
set the maximum number of stars in a group to > 100. I don't think this makes
sense, but you can do it. Simply do a
daopars.maxgroup.p_maximum=300 (say)
and then you will be able to epar daopars and set maxgroup to a larger value.
But, I don't think this is going to get you where you want to go.
The only stars that are going to affect the PSF are those within a PSFrad. I've never heard of restricting this to only 3 pixels, but I sure hope you don't have
>100 stars within 3 pixels of a star you're picking for a PSF star!javascript:emoticon(':roll:')
Is there really no light coming from the brightest star you're interested in
beyond 3 pixels? If there is, you need to make the PSFrad to whatever this value is. Sounds like your data is horribly undersampled, and that daophot
isn't a good solution.
---phil
Last post on Jul 18, 2006