Flat Wizzard: Dynamic Exposure does not work well for smallband filters on OSC Camera

[Michael Ring]

OK, this is not the standard usecase I have to admit…..

I have bought Antlia Filters and will sooner or later buy a monochrome camera.

Up until that I use the Sii Filter so that together with my L-eXtreme Filter I have complete data for hso images. I also will use the Ha Filter for RGBHa images of Galaxies.

Problem is now that the auto configuration for Dynamic Exposure does not work for my OSC Camera, it completely blows out the red channel when creating flats the Ha/Sii Filter (Have not yet tested Oiii, but will likely be similar).

I did a manual test, when I expose 30s with gain 90 with my ASI071MC Pro with the Ha Filter and the histogram for Red is at 50% so the flats show nicely the small vignette of my Scope.

But Dynamic Exposure calculates 100 seconds for the first try and then says flats are too dim and increases exposure.

In fact, when I expose for 90 seconds the resulting image is already completely blown out in the red channel. I verified this in pixinsight, the 30s exposure has red values of arround 0.5 as expected and the longer one has 0.99 on red.

To some extend the problem also exists for the L-eXtreme Filter, the flats are at 0.75 on green channel, blue is on 0.55 and red on 0.12 when I use Dynamic Exposure. Because of the low red the green channel gets 50% overexposed but not blown out, so the flats likely still work well.

The distribution of signal on RGB will of course depend on the spectrum of the light that is used to do the flats, I use one of those round flat lightsources that are recommended for diy flats-panels that people usually buy from aliexpress.

A possible solution could be to check luminance for each channel seperately when trying to find the correct exposure, but I am not an expert on Flats……..

[Stefan B]

When your flat panel produces a flat that is 0.75 on one channel but 0.12 on another, then it is next to impossible to determine a good exposure time. A different light source that is more even is required. The flat wizard will just take the average of the image to determine the time.

[Michael Ring]

For a smallband filter no light source will work, for Sii and Ha there will always only be light on the sensor in the red channel, no matter what light source you choose, similar for Oii, there will be green and blue but no red.

As there is no manual mode (at least I did not find it) there is no way to use the Wizzard and I’d have to resort to Sequencer.

‌

[Stefan B]

Removing version: 2.0 Beta (automated comment)

[Stefan B]

Nothing really that can be done about, as the logic works on the overall mean. When looking at histograms per channel it also won't work as which value to pick then, when the channels are vastly uneven.

Best to just lower the overall target mean to work with so that one channel won't be blown out