Photometry experiment - failed! Now Resolved

[Mr Spock]

I'm going to have to give this another go. I hasten to add this is only an experiment! The results I'm getting are 'wild' though. I am hoping to do some photometry of variable stars 

The image I'm looking at is in Auriga, originally of a comet. Star A is HD52552. I used a very small section to be relatively free of any gradients.

So we have from CdC: A= 6.99, B=7.72, C=8.29, D=8.95, E=9.29, F=9.26.

For the image I separated out the green layer and used IRIS to determine the magnitudes. I got:

A=6.99 (calibration star used to set the level), B=7.526 (7.72), C=7.888 (8.29), D=7.974 (8.95), E=8.686 (9.29), F=9.505 (9.26)

As you can see, the results aren't even close. Strangely, looking at the image, I see similar to what IRIS has recorded. So, it must be the image? I'll have to try again and take new images! I do realize when doing this for variable stars you are comparing the variable to a single star (with a control), so if these results are repeated on multiple images then you will get a valid curve. Still, I'd like to understand this better.

This is my first attempt, so much to learn I think. Tips welcome 👍

 

[vlaiv]

How was image taken and processed prior to measurement?

All measured values seem higher than expected, except for F star - which could mean gamma correction.

Was image taken by phone at the eyepiece or something like that?

[Mr Spock]

DSLR with a 150mm f2.8 lens. The only processing is to take out the green layer. 

Rather than use an extracted green layer, I've just loaded the original RAW file which is unprocessed. That gives A=6.99, B=8.007 (7.72), C=8.449 (8.29), D=8.578 (8.95), E=9.356 (9.29), F=9.857 (9.26).

A good indicator should be C and D as they have identical background levels. The catalogue has them 0.66 mag apart but you can see on the image they look close and they measure 0.129 apart. 

Still wild. 

Addendum. I've just opened a different RAW file, one that has never been opened in any other software. That should rule out seeing or scintillation effects. Results from that are A=6.99, B=7.907, c=8.545, D=8.466, E=9.435, F=9.807. Not much different so I'm not understanding what is going on 

[vlaiv]

Do you mind posting that raw file for inspection?

[Mr Spock]

D72_7773.NEF

[vlaiv]

Brightest star might have been clipped / over saturated and that will produce inaccurate results.

Here is surface plot of star profile of A:

looks like core of the star has been saturated.

What you can try is to skip the star A and work with the rest as it seems to be only one that is saturated like that (the brightest).

Instead use B as reference and see what values you'll get for C, D, E and F that way (if they are more accurate).

In general, you want to use shorter exposures to avoid saturating star cores and then stack some of those to get better SNR, possibly even defocus image a bit as that spreads the light into doughnuts instead of them having sharp peak.

 

[vlaiv]

It is definitively clipping.

I used dcraw to extract raw numerical values from the image and for star A I got this:

Max value is clipping value for 16bit integer. You can still see Bayer matrix pattern in this raw data, but it is very likely that saturation is in green as it is the most sensitive of the three.

[Mr Spock]

I redid the test using C as the control. I now get:

A=6.72 (6.99), B=7.63 (7.72), C=8.29 (control), D=8.27 (8.95), E=9.10 (9.29), F 9.75 (9.26)

I also looked in Stellarium to see if the catalogues used were consistent - they are not. This gives A= 6.95, B=7.70, C=8.25, D=8.40, E=9.05, F=9.25

Hmm. I'll take some more images of recognised VS fields and see what I come up with. And maybe ask for advice on how to do this properly from the BAA VSS. This was only really a test to see if the software worked but it's just thrown up confusion.

Meanwhile I'll stick to visual estimates. I'm good at that 

[Mr Spock]

Ok, I took a step back and had a look at the data and how it was being collected. It was clear their are catalogue errors, so I sorted that. Discounted star A as it was over saturated. I then looked at the sample aperture settings in IRIS to see if I was collecting the correct data. Finally I used two different images and took five separate samples of each star and averaged them. Here's what I got...

7772
	Actual	1	2	3	4	5	Ave	Diff
B	7.72	7.669	7.789	7.797	7.837	7.669	7.7222	0.00
C	8.29	8.298	8.152	8.25	8.159	8.069	8.1556	0.13
D	8.40	8.446	8.403	8.403	8.403	8.281	8.3572	0.04
E	9.19	9.428	9.414	9.03	9.509	9.321	9.3104	-0.11
F	9.40	9.535	9.565	9.535	9.405	9.405	9.459	-0.06
7773
	Actual	1	2	3	4	5	Ave	Diff		Combined
B	7.72	7.734	7.72	7.662	7.801	7.702	7.7238	0.00		0.00
C	8.29	8.312	8.338	8.23	8.257	8.312	8.2898	0.00		0.07
D	8.40	8.344	8.422	8.535	8.312	8.455	8.4136	-0.01		0.01
E	9.19	9.342	9.342	9.424	9.076	9.039	9.2446	-0.05		-0.09
F	9.40	9.444	9.444	9.429	9.444	9.455	9.4432	-0.04		-0.05

 

As you can see, all the results meet the +-0.1 mag I was looking for, and data set 2 is better than +-0.05 - perfect!

What I need to do now is make sure I have enough images of each target to eliminate any seeing or scintillation effects, get the data capture radius settings right, use the right catalogue, and enjoy the results!

[Mr Spock]

Tonight was all about camera testing. I'm using a Nikon D500 and Sigma 150mm f2.8. I wanted to see what it was capable of in my light polluted conditions. Just at the point the background is becoming visible, the best settings are 30s @ f4, ISO 400.
The original RAW file has stars detectable down to mag 14.2. Here's a 'snap' from that image. I'll do some testing on the RAW tomorrow. Target area is Arcturus, chosen because of its location and VSS chart availability for T Boo.

Even got one of those darn satellites...