Bias Subtraction documentation mismatches implementation?

[kolecki4]

The bias_subtraction function in modules/image_processing/src/alg.py indicates that it takes L0 files as inputs, with one being the master bias.

However, the master bias files in /data/kpf/masters/ on Shrek are 2D images. If the pipeline is performing bias subtraction by subtracting 2D master bias files from L0 image files, this could result in the fixed-pattern noise we've been dealing with.

Here are plots of the master bias file and a single bias image from the same date:
(Note the different color scales between the L0 and 2D products.)

Master bias file from 20240906:

Single bias image from 20240906:

L0:

2D:

[awhoward]

@kolecki4, does a row-cut or column average of the bias frame give the same fixed pattern noise that we're seeing in flats?

[awhoward]

@RussLaher -- I saw your PR related to documentation recently. Thanks. Does that close this issue?

[RussLaher]

Yes, that PR covered it. This issue can be closed.