Aperture photometry: different radius for each star

no you can not change it.   Once it is set then it is the same for all of the stars in that image and for all the images you are analyzing in the same batch.

Thanks Barbara.

Luigi

Sorry, but I'm not yet convinced that we can't change at least the middle and outer circles in the same image.  I think that the inner circle should be the same but we can change the middle and outer circles. Often, in ensamble photometry, it is difficult choose a set of circles that is ok for any star of the ensamble. If a star have a near companion, I enlarge the gap (distance from inner and middle circle) so that the companion falling into the gap.
I do not understand why we can not make these changes. The software subtracts the measurement of average background made in the annulus (distance from middle and outer circle) to the measurement made inside the inner circle. The average of background in the annulus is not significantly modified from the gap dimension. I did some tests and the results confirm what I think.
Any advice would be welcome.

Regards,

Luigi

Hi Luigi,

Remember that a star is not a point source, but instead has a profile.  Some of the spreading of light is due to the atmosphere; some is due to the optics, mount tracking, etc.  Basically, the star profile never ends - think of the Sun and how far its light spreads.  Back in the 1970's, we measured star profiles out to several degrees.

This means that setting measurement apertures is a compromise.  You want the maximum signal (large aperture) but with the minimum sky background included (small aperture).  There is a sweet spot where you maximize signal vs. noise; that sweet spot is different for different brightnesses of stars as the relative contribution of the sky background changes.

It also means that, as you increase your measuring aperture, you continue to increase the star signal as you are capturing a larger fraction of its total light.  If you use a different measuring aperture between the target and the comparison star, then you are measuring different fractions of total light and the magnitude difference will change.  Your measure cannot be compared with someone else (there will be an offset), and if you use different aperture sizes on the next night, your next measure can't be compared with the first since you would have a different offset.  The only way things work is to use the same measuring aperture for the target as you do for the comparison, as then you are measuring the same fraction of each star's profile and you get an accurate differential measure.  (Well, there are aperture corrections that can be used to account for using different aperture sizes between target and comp, but I won't get into that here).

What about the sky annulus?  Recall from above that the sky annulus actually includes the continuation of the star profile, so if you use a small annulus, you will include a larger contribution from the star wings (you are closer to the star's center) and therefore get a brighter sky per pixel than if you use a larger annulus.  Again, as long as you use the same annulus size from one star to the next, that is a fractional contribution, and subtracts the same fraction of the star profile.  If you change the annulus size between stars, you subtract more of a star's wings from the object with the smaller annulus than from the object with the larger annulus.  This effect is real, though significantly smaller than the measuring aperture change since you are talking about the wings of a star profile where there isn't much flux.

How large are these changes?  It depends, on the original sizes of the apertures and annuli in comparison to the seeing disk of the star.  The simplest solution is to avoid the problem altogether and just use consistent measuring apertures and sky annuli for all stars in a single image.  When you move to another image to analyze, you can of course change these radii for the new image, just not within a single image.

Arne

Hi Arne,
thanks very much for your answer.
I thought that the contribution from the star wings was negligible at distance from center greater than 2.5 FWHM.
However, I have done a test.
I have choosed an isolated star on my image, set the inner circle to 2.5 FWHM and I have measured the background average in the annulus for some gap values. I report the results in attachment file.
How you can see, the background average remains more or less constant. The relative magnitude fluctuates very little. I have repeated the same test for other stars getting the same results.

Regards,
Luigi

Hi Luigi,

You have not given me sufficient information to understand your test.  You quote your inner aperture in fwhm, but the gap in some other unit - is this pixels?  What is the relationship between pixels and fwhm for your system?  How wide is your sky annulus - you only indicate the gap width? I assume the values are ADU/pixel, or is this the sum within the annulus?  How bright was your original star in peak ADU?  All these things matter.

Arne

Hi Arne,
the FWHM is approximately 5 pixels. I have set the inner circle to 12 pixels, and the annulus (distance from middle and outer circle) to 4 pixels.
I have used Maxim DL. It give me the value of the  "background average" (Bgd Avg) in the annulus, and the "Intensity" (sum of all pixels within inner circle less background). The value of "Intensity" is approximately 50000.

I put an example of window information in attachment.

Regards,

Luigi

Hi Luigi,

Then I guess I remain confused.

For your most recent post, you show "intensity" as 49690.  You show the sky background, which I assume is ADU per pixel, as 24833.  You indicate an aperture size of 12 pixels (I assume radius), which means the average value per pixel of your star flux is about intensity/pi*12**2 = 110 ADU.  So per pixel, the sky background totally dominates your measure.  This latest post does not seem to have anything to do with your original test, where you show the gap between the measuring aperture and the sky annulus changing from 1 to 7 pixels.  Did you leave the sky width to be the same or what?

Perhaps it would be best if you would go back to the beginning, and explain in detail what you are trying to do and what your tables represent.   I cannot give you advice based on the limited information that you have posted.  From what I've seen so far, you are in a background dominated situation, where the star profile is lost in the noise and so that is why you don't see much difference when changing the size of your gap.  This is not a very common situation, and having the sky value at half of your total well depth is definitely not normal.

Arne

Hi Arne, Roger,

Those are the setting of image:
Camera: DSLR Canon 600D
Lens: 85mm
ISO: 200
Aperture: f/4
Stack of 6 images with 10s exposure: 6 x 10s = 60s
No bias, no dark, no flat, only remove bad pixel map.

I have done a test as suggested by Roger. Sorry, I don't have a bias frame, but I have a dark of 10s. I have loaded dark in Iris and then read the background with "statistic".
Mean: 2049.35; Max: 2075.0; Min: 2024.0; Sigma: 5.91

Then I have loaded one of those images (the first) and read the values for background and for two isolated stars of different brightness with "statistic".
Background:
Mean: 2063.68; Max: 2090.0; Min: 2041.0; Sigma: 6.31

Star 1:
Mean: 2079.54; Max: 2162.0; Min: 2039.0; Sigma: 24.75

Star 2:
Mean: 2108.18; Max: 2413.0; Min: 2029.0; Sigma: 79.69

Finally, I have load the stacked image in Maxim DL and I have done a test with the two stars, measuring the Mag., Intensity, SNR, Background average, Background deviation. I have reported the results in the attachment file.
How you can see, the two stars show a different behavior. With some fluctuations, for star 1, the background average increases slightly for increasing values ​​of the gap. For star 2,  the background average decreases slightly for increasing values ​​of the gap. Anyhow, the difference in magnitude are very small (for star 1 it is 0.04, for star 2 it is 0.006).

Regards,

Luigi

Hi Roger,
yes, I have extracted the two green layers from each image, then stacked this two green layers and finally stacked all images.

Hi Arne,
sorry, I forgot to say that the measure for the two stars are made with inner radius= 12 pixels, annulus=4 pixels, and gap between 1-10 pixels.
The FWHM for two stars is about 5-6.

Regards,

Luigi

Thanks, Luigi.  Also, thanks, Roger, for your detective work on why the background was so high.

In your file "Photometry 3", you show two stars.  Star 1 is similar to the star in your "Photometry 2" file; Star 2 is much brighter.  The Star 2 shows the typical star-profile variation, where you are including the wings of the star in your sky annulus.  As you move that annulus outward, less flux from the wings is included, which is why the background is smaller.  For Star 1, it is much fainter, so that change-with-radius is obscured by the scatter in the sky background itself.  Trust me, though - the variation still exists; it is just harder to see.  A simple demonstration is to look at a radial profile plot of your star - you will see a typical Gaussian-like profile with a bell-curve like shape, where it decreases as you move out in radius, but will continue to be present for a long way (more obvious with brighter stars).

While changing the annulus inner radius might not make a major change in your photometry for some stars (the faint ones), it will make an obvious change with the brighter ones.  Rather than making a judgement call as to where this crossover occurs, it is far easier to leave things constant within a single image.  Note that almost all software packages do a pretty good job of removing stars that fall within the sky annulus, so just because you see a faint star in the annulus for one star is not a reason for changing the radius of the annulus.

Good luck in your quest for precision photometry!

Arne

Hi Arne,

thanks very much. I will try to follow your advice.

Regards,

Luigi