Affiliation
American Association of Variable Star Observers (AAVSO)
Mon, 05/03/2021 - 19:38
Hello all,
I have started doing photometry on variables with my mono cmos camera and telescope. This is a 14 bit camera. For processing I use Muniwin. Muniwin to my knowledge only works with 16 bits. And the preproccing output of the individual frames in Pixinsight also produces 16 bit files.
My question: i know the gain (e/ADU) and readnoise (e) specifications of my camera, but those are of course related to 14 bit. Is it correct to multiply these values by 4 to get a 16 bit analogue for the gain and readnoise values which are needed as input in Muniwin?
Thanks
Hi Remco,
The readnoise stays the same.
The gain may change. You need to determine how your camera vendor/driver handles your 14-bit camera. All vendors produce 16-bit FITS files for their cameras, no matter how many bits are in your analog to digital converter. There are two main ways of storing your camera results:
- the software does nothing to the sensor values. That is, data values in the FITS file range from 0 to 16383. You can determine this by saturating your sensor, such as exposing it to room light or daylight. If you don't see any data values greater than 16383, then the gain is what is specified by the vendor.
- the camera software takes your 14-bit numbers and multiplies them by 4. That is, data values range from 0 to 65535. If this is the case, then the vendor's gain value is DIVIDED by 4. That is, if the vendor's gain was 1.25 electrons per ADU, then the new value is 1.25 / 4 = 0.31 electrons per ADU.
I hope that helps!
Arne
Thanks Arne,
My camera produces 14 bit output. When i look at the histogram of the frames it goes to 16.383. I process my data in PixInsight and that produces 16 bit files from it after they are calibrated as a result.
As you said, the read noise stays the same. In hindsight that seams logical because it is an absolute number of electrons the camera produces per pixel. I will leave that unaltered then.
So the one question that remains is, do i have to convert the gain specification from the vendor (e/ADU) x4 to get from a 14 bit to 16 bit representation?
CS, Remco
Hi Remco,
If the numbers in your 16-bit image run from 0-16383, then you don't have to change the gain setting at all. How many bits are used to store your 14-bit value is not important - you could store the image in 32-bit integer, for example, just wasting storage space but not changing the values.
Be careful with PixInsight. It is a great deep-sky imaging tool, but has its own internal method of handling images, usually representing them in normalized floating point (values between 0 and 1). It also has a great number of filters and smoothing functions that are wonderful for improving extended objects, but modify the original image and so can distort photometry. Use it in the most basic form when calibrating photometric images.
Arne
thanks again Arne,
I will have a look at what part of the 16 bit storage is used for my 14 bit files then :)
Thanks on the heads up on PixInsight. when doing photometry I only used it to calibrate the frames with darks, flats and bias and then register before inputting the aligned files in Muniwin. I will have a look if Muniwin comes up with different outcomes when i calibrated the frames in Muniwin itself.
CS Remco