Spectrum Browser / Identify issue / Kinds of Input

[vnp85]

Last checked in 2.6.5 dev

Use case: finding exotic lines. Generally speaking, to know where the spectrum knob is pointing at. For example the upper hydrogens beyond CaK, up until delta it is easy and epsilon is a given. Or the Helium at 447.15nm. For the future, particular iron lines in the corona.

Input quality: I try my best to give crops that contain lines more or less horizontal, without significant bending, and without vertical features (sunspots, plages) contaminating the input, no saturation or clipping, no borders etc. The below observations come from using inputs that I consider to be as good as one can get.

1. Input uncertainties - what file should be loaded, pixelsize, proportions? Undesired result: window resizing, buttons and controls moving around, and getting lost on the UI. Selecting an input image, like a snapshot from a ROI in Sharpcap, pushes the spectrum browser's window in a way that its control buttons are moved around in funny ways. To solve, one has to close the window or resize it in counter intuitive ways (some autosizing image holder pushes center justiofied controls around in a left justified window frame?). Possible fix: the software should have a set of criteria to accept a spectrum image and reject the invalid input, or create a valid input under the hood (by cropping or horizontally-only resampling) etc. As a workaround, I would try several crop/horizontally resized input files on a trial and error basis - or put an irfanview at 100% next to the spectrum window to scroll until a visual match is found.

2. Invalid detection state, contrast/normalization criteria? Undesired result: scrolls into infrared with no error message, with a dark-ish overlay. Often times the identification fails, either because of 1. or some other reason, and the reference spectrum is scrolled to the bottom.

3. Loading another image, to identify, without closing the window after the attempts described above, yields uncertain results.