Mure DeNoise Detector Settings

[tdgm]

Hello

I have a question on the value for the Flat frame exposure numbers. When I run the script it comes uo with a number ~9458. But hen you say that in your opinion that the expsoure should be between 30-70% of full well which for 16 bit camera (65553 ADU) would be starting at 30,000 not 9000 range so what is that FLAT FRAME Expouser number?


Steve

[mschuster]

Hello Steve,

Is 30,000 e- camera full-well depth? If so, then a 9,458 e- exposure is OK, that is just larger than 30% of 30,000 e-.

In comparison, if Flat frame exposure were say 2,000 e- that that would indicate under-exposed flats.

The script does not know full-well depth, as it can depend on camera bit-width which is also unknown, so it just provides the measurement for you to double-check.

Flat frame exposure is trying to provide the typical flat-frame exposure in e-. More precisely, it is the median of the center-cropped and bias- or dark-subtracted flat frames in DN multiplied by gain in e-/DN to obtain a value in e-.

Mike

[tdgm]

Hi Mike

1st Off Happy Thanksgiving


Also thanks for the reply. I checked the full well capacity specs on my canera and there site syas >17,000. I also checked for my frame of mind and divided the flat frame exposure by your calculated gain and cane up with a DN of 31526 which is a little less than the 50%b of adu I was shooting for, so thanks again and have a clear great holiday season


Steve

[Pompey Monkey]

I don't use the Detector myself:

The median noise value can be found by multiplying the total DN for the camera by the image noise (from FITS header or Noise Evaluation script).
The gain (e-/DN) can be found by dividing the full well depth (in electrons) by the total DN for the camera.

Although these derivation rely on the manufacturers information, I find that they work well.

[niccoc1603]

Hi can you make an example? I have an ASI1600MM and manufacturer gain/RN values are reported on ZWO website (although they are in 12bit scale so they should probably be adjusted to 16bit)