Just spent over an hour listening, switching, measuring, and taking notes while going between S1 (default non-adjustable module), MAM (optional adjustable module), and B1 (new non-adjustable module). I'm not going to post FR measurements since my setup is not calibrated and I don't want for people to be confused when looking at amplitude peaks. But the relative measurements support exactly what I'm hearing with U12.
S1 vs B1 - I hear a noticeable reduction in bass quantity where the bass flattens out between 60-100Hz. You still hear a textured well extended sub-bass and a decent impact of mid-bass punch, but the quantity of the mid-bass is reduced, the attack is a bit slower, less aggressive, the decay stays the same, and overall you still have this analog quality articulate bass. With reduction of low end impact, now mids have more room to breath and overall sound becomes more balanced. I even hear more sparkle and some improvement in treble airiness.
Pretty much what I hear is aligned with MAM module adjustment close to fully open position. Keep in mind, with MAM module fully closed you start at the same level as S1 default non-adjustable module (max bass quantity). As you continue to open it, somewhere 2/3 of the way (based on what I hear with my U12 and MAM combo) you hit the minimum low end quantity, and as you continue to open it up the bass quantity starts to go up. The amount of bass with B1 module is somewhere in the middle between the default S1 (max) and semi-open MAM (min). Also, B1 has less isolation than S1, which is aligned with what I hear when adjusting MAM.
In a very simple analogy, MAM is like an EQ (not literally) which you can adjust to fine tune the low end quantity/impact. S1 and B1 are like two presets where you don't have to adjust anything, just select whichever you prefer and enjoy it! B1 "preset" hits a sweet spot for me
If I'm not mistaken, S1 vs B1 default module selection will be available for future A/U models orders. Also, my observation is based on U12 and S1/B1/MAM evaluation, and I can't say for sure if exactly the same will be applicable to other models. Hopefully more people will chime in with their observation.