What I'm trying to do is read a rooms noise, smoothed over some time (say a 5 minute average), and have the sound level dB normalized (can be buckets really: 0-10, 10-20...70dB+) so that they direct a linear actuator to extend or contract such that its position matches to the corresponding noise level.
Many people have tried to do this sort of thing and they all have failed.
This is mainly because the perceived noise is not well related to the measured noise. I remember several years ago some one trying to measure how noisy their school canteen was and when it exceeded some threshold a “please be quie “ sign would light up. Six months of fiddling with averaging and thresholds failed to find a soloution. At best this is a lot more difficult than you think.