Image Format for Flats - PNG FITS or ?

[GaryS]

Hi
When I create flats for photometry, I always make them as .Fits images.
When I've been doing flats in Sharpcap for EAA observing, I generally do them as .PNG files, without really thinking, since that's the setting I have selected for image capture.

Is it a given that flats should have the format of the images they are intended to calibrate? It seems logical but I've never run across a discussion that addressed this issue - maybe its just too obvious that they should be matched in type with the images they are intended to correct - just checking.
thank you !
Gary

[admin]

Hi Gary,

it shouldn't really matter - the important thing is getting the image data back out of the saved file accurately, and SharpCap can do that for both FITS and PNG for 16 bit image data.

cheers,

Robin

[GaryS]

Robin
Yesterday I made a few versions in both png and Fits - they looked incredibly different when I opened them in Maximdl.

The fits version showed the vignetting and a few donuts where dust is, just as I’d expect, while the png version showed a totally homogeneous texture across the entire image.

I was using my ASI294mc but I’m not clear whether that’s actually 14 or 16. There are some subtleties about bits that I don’t understand.
Gary

[admin]

Hi Gary,

the 294 is a 14 bit sensor, but SharpCap automatically scales up the image data from the 0..16383 range to 16 bits (0..65535).

If you want to share one file of each format (via google drive, etc - they will be too large to upload here), I will happily have a look to see if there is anything we have missed or are not noticing.

thanks,

Robin

[GaryS]

Hi Robin:
Here's a link to my Dropbox which will take you to a Folder labeled "For Robin".
In the folder there are 2 Flat Files - one prepared in .PNG format, the other is in .FITS.
I'm used to Flats that are mono but for some reason the FITS version appears to be in color.
Cheers
Gary

https://www.dropbox.com/sh/sgtnv1f4hkx3 ... PX0Za?dl=0

[admin]

Hi Gary,

making a monochrome flat frame is optional - see https://docs.sharpcap.co.uk/4.0/#Creati ... t%20Frames

I suspect your FITS flat was taken with this option unticked, meaning that the bayer pattern shows in the flat frame and using the flat will automatically colour balance the image. In general I prefer mono flats, but others have success with the colour ones.

cheers,

Robin

[GaryS]

Robin
I think I’ll have to redo this test. I may have left the box unchecked but I always intend to check it. I may also have taken the color Flat with Maximdl- I’m just not sure.

I was pretty interested to see that the png fits taken with SC shows none of the expected optical issues such as vignetting and dust motes which do show up in the the fits version of the flat. How can the png version correct for those optical issues ? ….. perhaps I need to make at lower adu values than the recommended 30-50% of full well which I’ve used as an adu target. I’ve noticed that I get different adu readings from SC and Maximdl. Of course I have no idea where SC takes its readings but with Maximdl, I sample an area that is just barely shy of the entire screen and use the ‘average’ adu value.
Gary

[admin]

Hi,

the PNG shouldn't really be any different to the FITS file, but some viewers make a mess of 16 bit PNG, so that might be part of it.

SharpCap aims to keep the flat frame exposure so that there are no (or very few) pixels above 90% saturation and also very few below 10%, as it is pixels at or near saturation (or black) that cause problems in the correction.

cheers,

Robin