Marc provided the correct methodology. The reason is by taking all your raw data: light, flats, darks, bias and calibrating, registering, integrating you end up with a linear integrated “master” to start adjusting to bring out the subject. The master is linear and you could think of it as a line out signal from an amplifier where the signal is a pre-defined strength etc and independent from your volume knob. So the signal can go through various units via line level connections and sound as loud at all points. You can then mix in other audio sources at line level and it will sound just as loud etc.. Essentially here the integrated master has the signal at a standard level so they usually always look black at first glance, its only once you start stretching the levels do you start seeing the signal which is the image itself.
So you process each panel for a mosaic and then those masters for each panel are now all same scale, amount of distortion and all signal and noise is consistent across each panel so each panel can be joined without patchy discolouration. Which is why its important to NOT adjust levels making the data non-linear as patchy will occur as a result. Its a strong point of PixInsight is the screen transfer function which lets you do a levels adjustment for display purposes only allowing you to work on the data in its linear state such as registration where you can easily see matching stars and build your mosaic up in its linear state. Once complete it’ll be huge and odd shaped and look all black. NOW you can start level adjusting it to a non-linear state for display tastes and if you did everything properly at all stages its a flat consistent field with no patchiness between panels. Distortion in your optics though multiplies out from the middle of a mosaic so the more panels you want the join the more difficult it becomes if you dont remove distortions in each panel set. Its a long process. AstroPixelProcessor is the only program I’ve used that will let you through a ton of subs in at once and it can create the largest possible mosaic and use flats, darks, bias in the mix where needed. But its easy to fail due to the massive amount of resources required in the process.
If you are wanting to try on the free or cheapass then find yourself a repeatable process to go from linear to non-linear, something you can save as a preset to reuse on each panel and save to a jpg, png or tiff file to load into microsoft ice or similar to stitch together. Ice is the simplest option for you to learn. Try starting small, take subs of any patch of sky, stars only preferably then move you view to the side a bit trying to cover 50% of the stars from first set then take your second set. Learn from there.
|