The Nikon D5300 has a well-earned reputation as one of the lowest noise DSLR cameras used for Astrophotography. Now that I have a new Nikon D7500, I was keen to see how it compared to the D5300 in terms of the level of read noise and the extent to which that read noise is non-random ( and thus needs to be removed using a Master Bias frame to prevent it summing up during image integration).
So here goes ...
A single bias frame Nikon D5300: ISO400, 1/4000th second:
This may look pretty bad but really the extreme stretch is bringing out the very small variations across the frame:
Standard Deviation: 3.53 ADU
( note: in this context, 3.53 ADU means 3.53 "units" on a real number scale ranging from 0 to 16,383 ( ie. a real conversion of a 14 bit digital scale ))
The master bias frame looks like this:
Nikon D5300: ISO400, 1000 x 1/4000th sec frames
Standard deviation: 0.48 ADU
The bands at the bottom are each 1 ADU brighter than the one above.
Now for the D7500 ...
A single bias frame from the Nikon D7500: ISO400, 1/4000th second:
It is immediately clear that the single bias frame is cleaner. The statistics confirm this:
Standard Deviation: 1.37 ADU
And the master bias ...
Nikon D7500: ISO400, 1000 x 1/4000th sec frame
Standard deviation: 0.07 ADU
The band at the bottom is 1 ADU brighter than the background.
The improvement is very obvious
Std. Deviation single bias frame: 1.37 versus 3.53
Std. Deviation master bias ( 1000 frames ): 0.07 versus 0.48
In graphical form ...
The read noise in a single frame from the D7500 is around 40% of that in one from the D5300. This should give me greater flexibility to reduce exposure times and still ensure that the read noise is only an insignificant component of the overall noise.
I will need to consider further the impact of the very low level of pattern noise in the Master Bias; it is so low that I will think about whether or not I still need to calibrate my lights with a Master Bias ( particularly for long exposures when the noise is dominated by light pollution and thermal noise ).