Problem with Generalised Hyperbolic Transformation

I have been experimenting with the GHT and it appears to be a very powerful tool but there is a problem when the image is translated into formats other than FITS. Here is a screenshot side by side of how an example image looks in Siril vs. when saved as tif, png or jpg:

It appears that the MIPS-LO parameter may be the culprit. It is around 0.27 and this is put into the FITS header as “MIPS-FLO” with the comment “lower visualization cutoff”. Maybe this isn’t taken into account when translating the image into other formats. If, on the other hand, this is just user error, I would like to understand what I’m doing wrong.

Also, the new symmetry point selection feature doesn’t work in 16 bit mode, and always sets the symmetry point to 1 regardless of the selection area.

It is because the cursors are not to 0 and max. See also Why are exported images so dark? in the FAQ. So yes, it’s because of MIPS-LO to some extent, but mostly because the display settings are wrong.

This one is an ugly bug indeed. Sorry

OK. The problem is that it is the tool itself that is setting the MIPS-LO as I adjust the stretch factor and intensity. If I then manually reduce MIPS-LO to zero, I don’t seem to be able to make adjustments in the GHT tool to get back to the same appearance. It does appear that I can manipulate the histogram afterwards and maybe get back to the same appearance. I also notice that the GHT tool does a great job of bringing out star colors but for some reason turns the nebula to gray (instead of mostly red if I do a normal histogram stretch). Maybe I need to separate out the RGB channels and apply the tool independently to each channel, but I would think that will destroy the color balance from photometric color balance.

This one has been fixed in the new version released today.

1 Like