AAVSO report shows unusual comparison star magnitudes
Stacking program, astrometric (plate) solver, fits file viewer
Brought to you by:
han59
Menu
▾
▴
Hi,
I have run a 'measure all annotated' photometry which - with one comparison star - looks quite normal (see attachment). However, when selecting an additional star from the comparison stars list the graph displays an exact mirror image of the magnitudes of the first comparison star. Using other star combinations does not show this behavior.
Any idea what happened here?
cs Axel
Ho Thomas,
That is weird. Rather then trying to fix it can you try the very latest version I have been working on for the last two weeks ? There have been many changes including in the comparison star routine. Maybe that fixes the problem.
http://www.hnsky.org/astap_setup.exe
In this version transformation is fully implemented but you need at least one image using a different filter. I'm just in the testing phase.
Cheers, Han
Hi Han,
I will use the latest version and report back.
cs Axel
I just found and fixed a problem in the last development version. The abbreviation of the comparison star was missing in the report.
http://www.hnsky.org/astap_setup.exe
just ran 2025_06_23 but same result (attachment). I feel that color coding the different markers in the graph makes them easier to distinguish.
cs Axel
Can you share the first 20 or 30 files for testing?
The colour is now fully related to the filter. So you it can plot in B, V, R at the same time. For a single colour like your setup it will be green.
Han
Looking to your graph the mirrored curves are logical. The COMP star is an esemble of two stars. So all is referenced to the mean of the two stars. So the mean of the two COMP star should be steady by definition. It is only difficult to understand why at least one of the stars is varying so much. If you activate Gaia ensemble then one of the two should show a lot of variation. Why that is no idea at this moment.
I'll send you the files for testing.
download from https://we.tl/t-6MZDliuPDi
Hi Axel,
I suspect this is caused by the difference in sensitivity of the pixels. The stars have a varying position in the image. This sensitivity variation is normally be compensated by the flat& flatdark and dark. It doesn't look they where applied. If not can you apply them (calibration) and try again?
I see I still have to fine tune the photometry tab interface and make some changes.
Han
The images are dark corrected with the internal dark correction of the camera. I have still not been able to convert my NEF images in ASTAP correctly so there is no flat correction.
Can you convert the images I sent you and extract the green channel?
I had the impression you used a green filter. Are those image from a OSC camera? If so I have to know the Bayer pattern. But even then so the stars illuminate only a few pixels so the peak of the star could be at a red , green or blue sensitive pixel. This will make photometry even more difficult. The stars are pretty sharp, so magnitude measured using the green sensitive pixels will be more random and depending if the flux peak is located at a green, red or blue sensitive pixel. Only defocusing during imaging could help to distribute the flux more equally.
Han
I have been using an astromodified Nikon D750 with a 100/600mm refractor without any additional filter. I did not have the impression that the stars are particularly sharp with this lens but good to know - I will test how far I have to defocus to get better pixel coverage.
I have attached a masterflat made with that camera / lens combination in case that makes a difference.
cs Axel
I have calibrated you images but the flat is overcompensating. See attached. That would indicate the flat-dark value was too high. I think it is better to fix this first. After applying the flat the image background should be everywhere in the image the same.
What I do not understand is the resolution. The D750 resolution is 6016 x 4016 pixels. Your images are 1963 x1320 pixels. That indicates a bin 3x3 factor. So any green sensitive pixels are mixed with other pixels. To get good a photometric results you should first calibrate the full raw resolution 6016 x 4016 and then extract the green sensitive pixels resulting in 3007 x 2008 pixels image or something very close this. To fix your problem first this process should be in place. Can we test this first?
Cheers, Han
Sorry, I forgot that I have set the D750 image size to FX as that fov was large for the required field . The NEF image size should be 3936 x 2624 (see attached).
As I told I still have problems calibrating NEF images to FIT and extracting the green channel. I guess that is a failure on my part rather than that of the software.
Preparing other flats is no problem but in this case I am unsure about the exposure time. Usually I expose until the peak of the histogram is at 50% of the histogram - in this case it took a very unusual 30s. I could prepare a set of shorter exposures to find a better compensation. What would you propose?
cs Axel
That image looks good to me. If the NEF image is loaded in ASTAP, the libraw library tells me the Bayer patter is RGGB.
The maximum value is something like 16833 so I assume it is a 14 bit camera.
Yes a flat expose up to 50% is a good approach. 30 second is long but should not be technically problem technically. Why does it takes so long?
Are you sure a dark is subtracted?. Better to test it with. Only with a dark and flat-dark normally the flat correction is perfect. You could try with and without dark & flat-dark. The result should be a perfect flat background. Lets try again with new images.
Just made a series of flats and flat darks (25s, ISO400, NO internal dark correction) but applying them to my light frames shows the same over-correction of the images. I have attached an original flat and a flat dark, perhaps you can make out what goes wrong here?
cs Axel
The dark is also essential. The dark pixels values are around 600. That counts if the image values are around 650.
So
1. DCS5869 to lights tab
2. DCS5954 to dark tab (I misused the flat dark for this test. Better to have some darks of 30 seconds)
3. DCS5984 to flats tab
4. DCS5954 to flat-dark tab
The I select stack option calibration without demosaic. Or just used the popup menu option calibrate in the photometry tab. Then the result has a equal background. See attached result:
So besides lights, darks, flats and flat-darks. The flat-dark are added to the master flat. Once you have a master dark and master flat all next images will calibrate well.
Han
After this process you can extract green.
The light frame is already dark-corrected by the camera internal dark reduction. So I just added the flat and the flatdark in the respective tabs. I checked Calibration only in the Stack method and in the photometry tab popup menu used the calibrate option but received:
18:55:06 █ █ █ █ █ █ Can't process this file type. First analyse file list to convert to FITS !! █ █ █ █ █ █
When using the analyze button I received:
8:55:32 Converting G:\astro\Veränderliche\Boo\Boo V417\lights_DSC5869.NEF to FITS file format
18:55:32 Error reading G:\astro\Veränderliche\Boo\Boo V417\lights_DSC5869.NEF.fits
I guess I am making some stupid error!?
cs Axel
I can not explain that. Somehow conversion to FITS fails. I did some testing here and made a similar path "Veränderliche" but it should all work.
Is the G: drive a network drive?
Does it have enough space?
Can you try a local drive like C: ?
Can you open the NEF file in ASTAP?
If so the flat correction without flat-dark should work. I have a little doubt if it will work well but let's test it.
Checked on a laptop (both software and data on the same external SSD) and on a desktop (software and data on different HDs), no network drives involved. Same error on both with astap 2025.06.23 (attachment).
It works when images are copied to c: (with software running on c:) so it appears to be a simple access problem
cs Axel
software and data on C:
it works as expected:
12:58:31 Stack method Calibration only. No de-mosaic
12:58:31 Analysing lights.
12:58:31 Converting C:\bilder_DSC5869.NEF to FITS file format
12:58:33 Grayscale stack (classify by light filter unchecked)
12:58:33 Stacking (Calibration only. No de-mosaic), HOLD ESC key to abort.
12:58:33 Analysing flats.
12:58:33 Analysing flats
12:58:33 Converting C:\bilder_DSC5984.NEF to FITS file format
12:58:34 Converting C:\bilder_DSC5954.NEF to FITS file format
12:58:35 Averaging flat dark frames.
12:58:35 Adding flat-dark image 1 to flat-dark average. C:\bilder_DSC5954.NEF.fits
12:58:36 Combining flats.
12:58:36 Adding flat image 1 to flat average. C:\bilder_DSC5984.NEF.fits
12:58:37 Combining flats and flat-darks.
12:58:37 Applying the combined flat-dark on the combined flat.
12:58:38 Saved C:\bilder\master_flat_corrected_with_flat_darks_CV_1xF_1xFD_25sec_2025-06-26.fit
12:58:38 Master flat(s) ready.
12:58:38 Calibrating individual files only.
12:58:38 █ █ █ █ █ █ Warning, could not find a suitable dark for 30 sec, temperature 999°, gain 400 and width 3968! De-classify temperature or exposure time or add correct darks. See report in column ISSUES in tab dark.█ █ █ █ █ █
12:58:38 Loading master flat file C:\bilder\master_flat_corrected_with_flat_darks_CV_1xF_1xFD_25sec_2025-06-26.fit
12:58:38 Calibrating file: 1-1 "C:\bilder_DSC5869.NEF.fits" to average. Using 0 darks, 1 flats, 1 flat-darks
12:58:39 █ █ █ Saving calibrated file as C:\bilder_DSC5869.NEF_cal.fit
12:58:40 Calibration of the individual files is complete. New files are posted in the results tab
12:58:40 Completed. Resulting files are available in tab Results and can be copied to the Blink, Photometry or Lights tab.
The calibrated file however has little contrast and the masterflat is uneven. This is perhaps my fault, I will check the input images.
in ASTAP the calibrated file displays a grid pattern on the screen