(wow) Words Of Wonders Level 2259 Answers – CNers have been asking for a Cloudy Nights gift box over the years, so here you go. No donation is required, so enjoy.
I spent about 8 hours on Monday night collecting H-alpha data, which I ended up throwing away because, try as I might, I couldn’t get a good focus even with 10 second exposures. Last night with the same vision prediction and about the same guiding power (about 1.1 arcsec RMS) with the OIII filter I had no problem getting excellent focus with nicely sharp stars and objects.
(wow) Words Of Wonders Level 2259 Answers
I originally thought it might be the OIII filter versus the Ha filter. Maybe SCTs have a hard time focusing on red, just like refractors have a hard time focusing on blue? But data from almost a year ago with the same scope, camera, Ha filter, etc. they had very nice sharp stars and objects (and what I ended up using for my final image).
Collection: Recent Local Coverage Related To Covid 19
PixInsight shows the FWHM of stars on Monday night as excellent? Some images have a FWHM of less than 1.0 arcsec.
And PixInsight also shows that the FWHM of all images from a year ago is the same or worse than Monday night. Visually, not even close. The stars of Monday night are puffed up and the motive is blurred. I thought FWHM was a good indicator of focus and vision? But in this case, apparently not.
I’ve attached the image analysis from the PixInisght subframe selector, a comparison of a year ago and Monday night along with screenshots.
Maybe the FWHM numbers are “correct” in a relative sense from the same night of imaging, but I don’t think they’re relevant for comparing different nights.
Aes E Library » Complete Journal: Volume 11 Issue 2
One assumption is that out-of-focus stars are slightly fainter. And I showed that the dimming star will have a smaller FWHM due to the detection limit of the optics and will jam the outer rings of the PSF because they simply won’t show up on the sensor. But I think they would have to be significantly darker to get the effect I’m showing here. (?)
I think this shows that FWHM is NOT a good indicator of focus. Perhaps HFD works better as an indicator (as many claim). But I’m still interested in these analytical algorithms and what they show.
I don’t see any stars that look like this in individual or stacked images (and why I don’t “trust” these numbers)
So maybe individual images have too much noise / too little signal for the program to accurately determine (?)
Is One Brand Of Sewing Machine Better Than Another? • Erin Says Sew
When I compared the FWHM numbers between different nights on the same target, it was never clear to me that the group of stars selected by the analysis routine was exactly the same in both cases. If another set of stars is selected, their total FWHM value may differ from the previous set.
Another factor that can change the FWHM estimate can be the background level, which varies with sky conditions and the position of the object in the sky. Changes in the background level can change the pixels that the algorithm considers to be in an individual star, even if the central values of those stars have not changed significantly since previous estimates.
When I compared the FWHM numbers between different nights on the same target, it was never clear to me that the group of stars selected by the analysis routine was exactly the same in both cases. If another set of stars is selected, their total FWHM value may differ from the previous set. Another factor that can change the FWHM estimate can be the background level, which varies with sky conditions and the position of the object in the sky. Changes in the background level can change the pixels that the algorithm considers to be in an individual star, even if the central values of those stars have not changed significantly since previous estimates. – David F
The January images definitely show “narrower” stars – I would also expect the FWHM to be smaller for the earlier data.
Never Have I Ever But For Bangtan I Did
Is there a way (as David suggested) to select a specific (same) star to analyze in both images?
I also wonder if the eccentricity (also as above) is relevant – the stars seem quite round to me, certainly better than 0.7 or so….
The January images definitely show “narrower” stars – I would also expect the FWHM to be smaller for the earlier data. Is there a way (as David suggested) to select a specific (same) star to analyze in both images? I also wonder if the eccentricity (also as above) is relevant – the stars seem quite round to me, certainly better than 0.7 or so….
I opened each image in AstroImageJ, used the same files as in the original post, and selected the same discolored star.
The Word Of Zul
The December star has a 75% larger FWHM than the January star (10.89 pixels vs. 6.23 pixels, or 5.3 arcsec vs. 3.1 arcsec).
The December image seems to have a lot more background noise. Perhaps it was a transparency issue rather than a vision issue (as the guidelines seemed roughly the same both nights)
Thanks for the suggestion to choose the same star. I feel much better now (and know not to put too much faith in PI numbers on individual images, especially noisy ones).
But are the stars watching me? This was my last image from January 14th (Ha) and December 22nd (OIII) https://www.astrobin.com/uzaxji/ I don’t see any stars that look like this in individual or stacked images (and why would trust ” not these numbers)
We Have To Somehow Give Twice As Much, For Half As Much, All While Hiding Parts Of Ourselves To Come Off As Friendly To Them
If you look closely at your minor stars, there seems to be a difference between them. My guess is that whatever is causing the bulge (seeing, etc.) is masking the greater eccentricity of the bright stars.
But in the last stacked images, the FWHM and eccentricity numbers look very believable and believable. Maybe the individual images have too much noise/signal for the program to accurately determine (?) Note: I have 0.49 arcsec/pixel so the two routines in PixInsight match.
These FWHM diagrams seem to show what I mean. Eccentricity was smallest at the highest FWHM…as FWHM decreased, eccentricity increased
Judging by the numbers alone, the difference between the images is not that big… (seventh of a pixel is not that big of a difference)
Contest #801 Summary
>>>>>>If another set of stars is selected, their total FWHM value may differ from the previous set.
That sounds reasonable. But as far as I know, the FWHM of a star is the same for large bright stars and for faint small stars (in a given photo, on a given night, assuming a flat field, etc.). Therefore, it should not matter which stars are chosen.
I suggest you open both linear images. Use STF to stretch one of them onto the screen and pass the parameters for THAT stretch to both images via the histogram process. (Stretch one of them with STF, drag the “New Instance Icon” – the small triangle in the lower left corner) to the bottom of the Histogram Process window and expand it. Then click on the second image and drag the same new instance symbol from the histogram process onto the second image.)
My guess is that the FWHM calculation does the FWHM in the X direction and again in the Y direction and then takes the average of the two, which accounts for the possible correlation between eccentricity and FWHM (and possibly a non-linear correlation since the ellipse equation is non-linear).
Inlander 8/06/13 By The Inlander
Judging by the numbers, the difference between the images is not that big… (the seventh pixel is not that big of a difference) Is it just the selected stars (as in): >>> >>>If you select a different group of stars, their total FWHM may differ from the previous set. That sounds reasonable. But as far as I know, the FWHM of a star is the same for large bright stars and for faint small stars (in a given photo, on a given night, assuming a flat field, etc.). Therefore, it should not matter which stars are chosen. I suggest you open both linear images. Use STF to stretch one of them onto the screen and pass the parameters for THAT stretch to both images via the histogram process. (Stretch one of them with STF, drag the “New Instance Icon” – the small triangle in the lower left corner) to the bottom of the Histogram Process window and expand it. Then click on the second image and drag the same new instance symbol from the histogram process onto the second image.) (Remember to remove all screen parts from both images.) What is the result? Alex
So far I’ve only analyzed calibrated, not stretched, images.
