How do I measure noise in a RAW or JPEG image? I can’t find any menu to do this? Also, in viewing spectra, it shows red, green, and blue peaks, but it shows the green peak twice. Why is that?
Would you please describe in a little more detail what you have already done and maybe include a representative image for reference?
We would be more than happy to assist.
As for the two peaks in the green channel, it simply means that you have two clusters of green pixels, one at each location.
Thanks. I did not know the two greens in the Bayer pattern were analyzed independently.
Peaks in an RGB channel have nothing to do with the number of green pixels in the Bayer mask of a camera. You cannot see the Bayer mask. The RGB image has already been interpolated.
What Bob explained is that there are 2 clusters of green intensities in the histogram.
What I am talking about is the fact that there are two green pixels in each color set of RGB on the Bayer sensor. That would account for the two peaks.
No, it would not, for the reason given above.
Then why are there 2 green peaks and only 1 red and 1 blue?
Maybe your image has an area of a given green intensity and another region with a different one. Or it has some imperceptible banding or uneven illumination. It is difficult to say without seeing the image.
What you can do is stretch/compress the contrast or move the threshold limits interactively and see which areas get highlighted around those peaks.
By the time you get the RGB image, every pixel has been “de-mosaicised” (i.e. the RGB values have been computed based on the Bayer mask).
You can also post your image here for others to see.
This only relates that there are two clusters at different intensity levels but somewhat equal pixel numbers. Does that clarify any? If not keep asking.
Alright, I see what you are saying. The two green peaks are dependent on the specific photograph. So, one could end up with several peaks of red, green, or blue.
Yes, that is correct and happens quite often. But you are sharp enough to check and respond very well. Good work.