Error calibrating using flats

Hi Phil @philsav8,

In which APP version was this? in APP 1.061?

Did you create the master frames in APP 1.061?

Perhaps you can share your master frames so we can have a look why it's not working as expected?

Kind regards,

Mabula

Hi Mabula,

I redid the processing but this time with 10 lights, darks, flats, bias frames and still the same issue with the flats. I think I could be missing a step/setting?. I basically run through the tabs 0-6

This was done in APP 1.061.

I have uploaded the master flat/dar/bias frames plus some lights to Google Drive: 

https://drive.google.com/drive/folders/1o6KTpieYqcr35DhfcIKWdUeytsKCLWEp?usp=sharing

thanks Phil

Hi Phil @philsav8,

Thank you for sharing your data.

I have downloaded the data and I think I know what the source problem is. But... it's best if you can provide more information on your data:

In your initial post, you state that the data is from a ASI 294 colour camera.

So I would expect your light frames to be FITS frames in 16bits. They are 8bits TIFFs now and even compressed with LZW, that really will not be a good starting point.

https://astronomy-imaging-camera.com/product/asi294mc-color

The camera has a14bit ADC, so by storing at only 8bits (and compressed) a lot of information is lost. You should try to save the captured data in 16bits and uncompressed, preferably in FITS. (but TIFF should work as well)

Do you have original FITS frames? Or can you share how you captured your data, what capture software did you use?

Were the masters created from 8bits TIFFs as well? (it does seem so, because the masters are 8bit as well).

In the meantime, I will have a look at what I can get out of the data that you provided 😉

Mabula

@philsav8,

This is a screenshot of the masterdark with it's histogram.

Both the Masterdark and MasterBias masters are very important for flat calibration to work well. Without a good Masterbias or Masterdark or both, flat calibration is doomed to fail.

Clearly, these frames are not suitable at all to be used as Masters. This has to do with the data being only

• 8bits instead of 16bits.

The LZW compression shouldn't be a problem, since it's lossless, (but I still wouldn't use it then I think)

The 8bits has reduced the original bitdepth with a factor of 14bits - 8bits = 6bits = 2^6 = 64x.

That means that all data values between 0 and 64 in your original data are either 0 or 1 now or

That means that all data values between 64 and 128 in your original data are either 1 or 2 now etc....

This has a great destructuve effect on the quality of your calibration masters, there is severe posterization in the masterframes and valuable information is lost which is needed for calibration.

So this will only work with your camera if you capture the data in 16bits and use that in calibration 😉

Mabula

Hi Mabula,

I understand my mistake, totally my fault. I have just started using SharpCap to acquire the images (as well as a new scope)  and to keep things simple and to ensure I could see the captured images easily to check focus etc on my laptop while acquiring, I saved as jpg. Never again  😀 The error with the flats makes perfect sense as well. I w=ondered why there was such posterization.

Thanks for your help in pointing me in the right direction 

Phil

Hi Phil @philsav8,

No problem 😉 Yes, especially JPGs are terrible. This is because the JPG compression algorithm is very destructive in low light areas, almost all data will simply be gone....

So save your captures in 16bits Fits, I assume Sharpcap can do that and then all should be fine 😉

Mabula