ASTAP stacking reports mismatch between Flats and Biases?

[Bob Kidd]

Hello, Han.
ASTAP stacking reported the following:
Warning: Flat not calibrated with a flat-dark/bias (keywords CALSTAT or BIAS_CNT).
And I I saw another related warning flash by in the console during the time the stacking process was underway -- Warning flat pixel colour values differ too much. Use white light for OSC flats!!. Will compensate accordingly."

This is the first time I've experienced this, but I only recently updated my version of ASTAP.
In case it matters, I use the ASIair app to produce my calibration files. I also use an LED tracing pad placed over the scope for the calibration light source.

Many thanks in advance,
Bob

[Bob Kidd]

Han, I placed a Bias and a Flat here on Google Drive.

https://drive.google.com/drive/folders/1xhGH3KmXbqyREqVOT4fhJiyLTQbcCjH-?usp=drive_link

Is it possible that I had the LED tracing pad dimmed too much when shooting my Flats, thereby affecting the color temperature?

[han.k]

Hi Bob,

I duplicated the flats twice and tested it with your single bias. So the flats are to be add to the flat tab and the bias to the flat-dark tab. Then you get this log

20:34:44 Analysing flats
20:34:55 Averaging flat dark frames.
20:34:55 Adding flat-dark image 1 to flat-dark average. D:\temp\bob\Bias_1.0ms_Bin1_gain100_20250216-150931_-4.9C_0001.fit
20:34:55 Master flat creation, warning the flat and flat-dark have unequal values for FD_exposure|
20:34:55 Combining flats.
20:34:55 Adding flat image 1 to flat average. D:\temp\bob\Flat_15.0s_Bin1_gain100_20250328-093814_-5.8C_0001 - Copy (2).fit
20:34:56 Adding flat image 2 to flat average. D:\temp\bob\Flat_15.0s_Bin1_gain100_20250328-093814_-5.8C_0001 - Copy.fit
20:34:56 Adding flat image 3 to flat average. D:\temp\bob\Flat_15.0s_Bin1_gain100_20250328-093814_-5.8C_0001.fit
20:34:57 Combining flats and flat-darks.
20:34:57 Applying the combined flat-dark on the combined flat.
20:34:58 Saved D:\temp\bob\master_flat_corrected_with_flat_darks_CV_3xF_1xFD_15sec_2025-03-28.fit
20:35:05 Master flat(s) ready.

And this is in the header of the master flat file:
FLAT_CNT= 3 / Number of flat images combined.
BIAS_CNT= 1 / Number of flat-dark or bias images combined.
CALSTAT = 'B'

The unequal exposures is not so important.

So it should all work if the you load them in the flat tab and flat-dark tab. Can you try again?

Cheers, Han

[Bob Kidd]

I experimented increasing the brightness of my LED tracing pad and shot new Flats at a sub-second time, ran the same ASTAP sequence and still got the warning: "Warning flat pixel colour values differ too much. Use white light for OSC flats!!. Will compensate accordingly."

To be clear, I got a stacked result both this time and in the instance that precipitated my original post.

So the "Warning flat pixel colour values differ too much ..." message can be ignored?
Just wondering.
Bob

[han.k]

Hi Bob,

This warning will popup if the red, blue and green channel signal differs too much in the flat. In principle the colour of a flat should be around white. If you make them using the twilight sky they can become too blue and if you are using electroluminance panel for flat note that some loose their red after aging.

If you debayer a raw flat you will see it.

It is just a warning. If you can not correct the flat light source, you can just ignore it. But by correcting the light source it you will get a very minor better result.

Han

[Bob Kidd]

Wow, that is very interesting, Han. You are a fountain of knowledge!
Sincerest thanks,
Bob

[Bob Kidd]

Han, one more note on this topic. You got me questioning the quality of my recent flats and I may have discovered something causing trouble in my setup. I have been running my LED tracing pad at a greatly diminished brightness. I now wonder if that low-power setting caused worse color balance. I shall now try the LED panel at full brightness and simply add more layers of T-shirt diffuser.
Best,
Bob

[Martin Dowd]

Ive been using an LED tracing pads ( Huion A3 and Huion A2 dimmable light pad ) to take flats on my Newts for years and years for both OSC and Mono data sets without issue . Some LED pads do exhibit either dominant wavelengths in the Red band or Blue band at low illumination levels. Obviously mine don't as my Ha flats are dimmed down to low level ( 10 % to 15% illumination and exposure time of 2 to 5 sec ) and my flats calibrate out fine with Flat Darks of the same exposure , Gain and temperature. On the other hand my Red , Green and Blue Flats require 80% illumination and exposure time of 0.1 to 0.5 sec ) and Flats calibrate out fine with Flat Darks of same exposure, Gain and temperature. I initially used a white T Shirt ( one layer ) but in the past 2 years Ive used a square piece of 3mm opal acrylic diffuser between OTA and Huion light pad , in fact ive duct taped the diffuser to the light pad .
My Flats are captured using APT and my lights suffer from severe vignetting caused by my TS Optics Coma Corrector. ASTAP does a beautiful job at stacking and aligning my data sets. Kudos to Han !!!
Maybe try a Huion dimmable light pad ??

Martin

[Bob Kidd]

Martin, thank you so much for your detailed response. Very kind of you. I have noted the Huion brand for future reference, though I already ordered an A3 size pad by HSK before I saw your note. Will hope for the best after moving to A3 from the A4 I was using on smaller diameter scopes (that was definitely needed) and adding layers of "Fruit of the Loom" until I can use the light pad at or near full luminosity.
One other note: I also saw about 10 percent of my flats were inconsistent (much too dark) when I tried shooting them with my A4 tablet dimmed low. I think I was inadvertently capturing flicker from the tracing pad light source!
Best Wishes,
Bob