I recently began using Sharpcap and want to make sure I understand a few things about flats.
I have a ASI294MC Pro and use the basic driver settings for the camera except I lowered the brightness to 10, and i normally change just the exposure time and gain. I shoot RAW16, 4144x2822, bin1, Debayer Preview=on, Output=PNG.
I create my flats with a tracing light panel over several layers of cotton t-shirt. This creates a histogram similar to the one in the documentation where the blue channel is to the left around 20% and the green, red, and white over to the right around 40-70%. The histogram appears this way with the red and blue channels set at 50.
If I up the blue channel I can move the blue curve higher and I can also adjust the red and green to better align these two and create a more compressed graph. This reduces the extent to which these curves extend towards the outer boundaries and compressed them more towards the middle of the scale. Is this what I want to do when creating flats? I have created flats recently using just a 50 50 on red and blue and my EAA pictures come out looking really good, but will compressing the curves more towards the center of the graph create a "better" flat?
Sharpcap flats and histogram
-
admin
- Site Admin
- Posts: 13353
- Joined: Sat Feb 11, 2017 3:52 pm
- Location: Vale of the White Horse, UK
- Contact:
Re: Sharpcap flats and histogram
Hi,
in my opinion, best practice for flats is to keep (as much as possible) all pixels in the image below 90% saturation, but other than that, as bright as you can go. You avoid anything above 90% to be sure of keeping clear of the region where the sensor response might not be linear (most CMOS sensors seem to be linear up to 99%, but why push your luck). You also want to keep everything clear of very low pixel values (<20% or so), since you can potentially get less accurate correction since the same size error in pixel value is relatively much larger at low brightness.
Now, colour cameras and serious vignetting or very dark dust specks can make it hard to obey both of these rules at once - at that point you have to balance things somehow, but do keep *everything* aside from perhaps a handful of pixels clear of 99+% on the histogram.
Finally, it's worth noting that the 294 sensor in both colour and mono forms have a special pitfall for flats. There have been issues with the ZWO models containing this sensor where the histogram will not fully saturate (all pixels at 100%) even under long exposure and bright illumination when the gain is between 120 and about 200. This issue can cause the histogram shape to look good for a flat when really the image is very overexposed. I would recommend being very careful in that gain range - if you do intend to work in that range, set the gain, turn the exposure up to several seconds (16 bit mode) and allow bright light on the sensor. If the histogram does not collapse to a single spike at the right hand side then do not use that gain value.
cheers,
Robin
in my opinion, best practice for flats is to keep (as much as possible) all pixels in the image below 90% saturation, but other than that, as bright as you can go. You avoid anything above 90% to be sure of keeping clear of the region where the sensor response might not be linear (most CMOS sensors seem to be linear up to 99%, but why push your luck). You also want to keep everything clear of very low pixel values (<20% or so), since you can potentially get less accurate correction since the same size error in pixel value is relatively much larger at low brightness.
Now, colour cameras and serious vignetting or very dark dust specks can make it hard to obey both of these rules at once - at that point you have to balance things somehow, but do keep *everything* aside from perhaps a handful of pixels clear of 99+% on the histogram.
Finally, it's worth noting that the 294 sensor in both colour and mono forms have a special pitfall for flats. There have been issues with the ZWO models containing this sensor where the histogram will not fully saturate (all pixels at 100%) even under long exposure and bright illumination when the gain is between 120 and about 200. This issue can cause the histogram shape to look good for a flat when really the image is very overexposed. I would recommend being very careful in that gain range - if you do intend to work in that range, set the gain, turn the exposure up to several seconds (16 bit mode) and allow bright light on the sensor. If the histogram does not collapse to a single spike at the right hand side then do not use that gain value.
cheers,
Robin