Thanks for the feedback Rolf, and again apologies for the mixed up on the initial post.
Yes, and here.
I was able to replicate the noise stripes then, and was trying to replicate with the current setup, but did not get them to show… yet.
Yes absolutely. I do not have an optical table for sure, and I am well aware of the issue. While my setup is not perfect, it is workable.
And I know the issue with plastic… so had to add some weight to the camera. Again, not perfect setup, but workable.
Great suggestion. The captures are presently stacked as they are captured. I am thinking this could be a capture error detection check.
The averaging is done pre-resolve. What I meant was that when the image is scaled 8 times, the resolve scaler is effectively doing spatial blending of adjacent spatial pixels. That is what I initially perceived as more detail in the single capture 12 bit, as blending of the noise reduces it and is perceived as a higher resolution texture… when in fact is just reduced color noise.
On the next snailscan2 update, I will go deeper into the current software and pipeline.
Thanks for the feedback and the great suggestions.