Inconsistent with Negative Conversion and Roll Analysis On

According to the scene (inside the courtyard and in the street) I shot all my film with my FM2 at the same speed and same aperture because during then, the light was constant (or it seemed to be). Once I set my speed during a scene I don’t touch it. The only setting I’m changing is the aperture because with a shift lense you can’t see through when the aperture is at f/8, it’s too dark, so I just set it at max aperture to check my frame and focus, and then get it back to f/8. That’s when I might sometime shot and forgot afterward I kept the max aperture. Nether the less, it happens only on a few shots (on this film specifically on 3 shots, like the one you pointed out).

And during my scanning, I also shot all my film with my Fuji XS20 on manual mode with the exact aperture and speed.

I think we have to admit all that the cloudy and rain condition when I took the pictures were not good and maybe because the shots are all similar (in terms of framing and subject, same building) the color differences are even more obvious.

As @nate pointed out above, taking a photo with the same setting and changing light will, in the end, produce different looks. Film is not linear and r, g and b sensitivity curves are not identical. Any change of parameters in capture necessitate some adjustment if output should be close to identical.

The non-linearity of film can be beneficial too - as we can learn from e.g. Ansel Adams’ zone system.

Yep, that’s the most likely explanation because the difference in the densities of the negatives of the same scene are very large (at least a few stops) and the lighting appears similar, and I can see from the metadata that you had the same digitization exposures.