SNR ratio, Photometric, Centalt, Airmass

[Pieter Vuylsteke]

Hi Han,

Just some questions that I do not find an anwer on :

Could you please explain how you calculate these values ?
Background ?
SNR Ratio : flux of the star divided by ? or ?
Photometric "poor" -> why is a measure "poor" ?
Centalt ? Meaning ?
Airmass ?

Kind Regards,
Pieter.

[han.k]

Hello Pieter,

SNR, signal to noise ratio:
For both bright stars (shot-noise limited) or skybackground limited situations snr:=signal/sqrt(signal + r x r x pi x SKYsignal) equals snr:=flux/sqrt(flux + r x r x pi x sd^2).

where r is the radius of the star, sd standard devation of the background around the star. flux is the star flux.

Photometric poor means:
Image not calibrated with a flat field. Absolute photometric accuracy will be lower. Calibrate images first using "calibrate only" option in stack menu. After a flat is applied this is normally documented in the fits header behind keyword CALSTAT.

Centalt, altitude of the object image at the middle of the exposure. CENTALT is a keyword used in the FITS header to document altitude.

Airmass. At zenith the airmass is 1,. At lower altitudes it increases. This is something the AAVSO likes to have.

Regards, Han

[han.k]

The forum doesn't like stars as multiplier. I will update the previous message with "x"

This is maybe clearer:
{See https://en.wikipedia.org/wiki/Signal-to-noise_ratio_(imaging) and https://www1.phys.vt.edu/~jhs/phys3154/snr20040108.pdf}
snr := signal/noise
snr := star_signal/sqrt(total_signal)
snr := star_signal/sqrt(star_signal + sky_signal)

[Pieter Vuylsteke]

Hi Han,

Aha, I do calibrate my images in PixInsight, so the photometric "poor" is to be ignored.

Also : if my SNR i e.g. 800, that means my signal is more than excellent ? From what point on it becomes "poor" ?

In fact, if I look at my check star, the magnitude changes +/- 0.02 mag, for photo's taken with a ZWO 2600mc cooled camera, calibrated, debayered and green channel extracted in PixInsight. And then analysed in Astap.

Is there a way to know that that 0.02magn. change is to be considered as the most accurate to be hoped for ?

Kind Regards,
Pieter.

[Pieter Vuylsteke]

Hoi Han, again...

I see in the pdf the formula :
delta magnitude= 1.0857 divided by SNR

My SNR is between 700 and 800, so the expected delta magnitude should be about 0.001 .
But, in reality it scatters +/- 0.02 magnitude.

I did the test on a fixed star for the first 10 photo's of the sequence, and get the values from the attachment.

The database V17 gives 9.55

Am I doing something, or understanding it in a wrong way ?

[han.k]

Yes ignore "poor" . The CALSTAT keyword for the status of flats and darked applied is an established method. PI should have added that,.

The 0.02 magnitude variation is about the limit you can achieve. They only thing what could help is adjusting the aperture. Using an OSC camera doesn't help either. In professional setups, i see sometimes close to 0.01 magnitude noise. I'm not an expert how to achieve that, but an observatory on a mountain would help.

I didn't see the attachement,

You should not debayer! Only process the raw and extract the green pixels, not the green channel! Debayer will result in a much larger absolute error.

Han

[Dominique DANIEL]

Hi you will be kind to tell me how to get the SNR information from ASTAP for an image file
Best, Dominique

[han.k]

The SNR of an image is more or less undefined. What you can measure is

1) the noise of the background
2) SNR of a star.
3) Limiting magnitude. This is the most useful measurement.

1) Can be measured by the viewer popup menu, "show statistic within a rectangle"
2) This is displayed in the status bar of the viewer when you move the cursor to a star.
3) Solve the image and go to viewer TOOLS, "MAGNITUDE (MEASURED) ANNOTATION" . See the value at the bottom of the image.

Han

[Dominique DANIEL]

Thanks Han
for this very complete answer. 3 cases are all interesting. My first
interest was for SNR of a Star. I just discover that I have to insist and
turnaround the star to make appear its SNR value.

Thanks again
Dominique

Le mar. 11 janv. 2022 à 09:24, > a écrit :