SharpCap Forums

I've been experimenting with what we have coined Ultra Fast (UF) EAA and here is the topic on CN.

https://www.cloudynights.com/topic/8928 ... acked-eaa/

What I've experienced so far when taking a flat in SharpCap 4.1 is the outer area of the frame, which is likely black due to the smallish central image, is completely overexposed when applied. I'm using an IMX585 OSC sensor (Uranus-C) and have tried using an ROI of 2180x2180 and the flat when applied looks good in the center and is overexposed everywhere outside the center. I've been using the flat with bias option.

Should I reduce the ROI to cut out more of the vignetted field? Is there a way to edit a flat and get rid of the outer "off'-calibration" area?

Hi,

interesting thread - once you see it as a 0.2x reducer it makes sense more quickly It's surprising how little distortion you have given the extreme nature of the reduction.

For the flats, there isn't anything quick and easy that you can do with SharpCap alone to adjust them, but with other processing software you could introduce a floor level to the flat frame.

Most flats are going to be 16 bit data, so values from 0 to 65535. The problems near the corners are caused by very low values, leading to high correction. If you load the flat into some image editing program (that can keep 16 bit data) and make all the pixels darker than some value (lets say 2000) get brightened up to value=2000, you will have limited the flat frame's correction in the darker regions. Maybe 1000 is the magic value, not 2000, or maybe it's 4000, but you get the idea.

In a traditional photo editing application, a flood fill from the corner with an appropriate tolerance would probably do the trick. I expect you could also do it in PixInsight or Siril or some other application that lets you do pixel mathematics on images.

cheers,

Robin

Robin,

Thank you for the explanation. It sounds like FITS Liberator should be able to make the changes you suggest. I'm sure there is an obvious change in values just outside the properly illuminated image circle. And a combination of being more selective with an ROI, combined with editing the FITS file sounds very doable.

Mike