Omg @ben.wes you went godmode! Well, you are really sophisticated. I think yours is a much cleaner solution (mathematically speaking), but it affects cpu cycles a lot.
Your intuition about the filter was definitely the best.
I dont need to explain my solution, I just positioned the filter in the right spot and simply exploited the feedback in the correct way (or at least I think).
And here's the grand finale

Gran Finale.mp4

Thanks for everything Ben, you will definitely be mentioned in this work!

WOW @ben-wes that's a great looking solution! Thanks for the answer
I tried to create the idea, having two smoothing factors.
The first concerns the transition between peak and rms response.
The second, as you well thought, with a lop~. The smoothing is quite evident, as is a frequency shifting/spectrum translation (probably phase problems caused by the filter). I would have to study quite a bit about rephasing the signal
I'm really out of ideas, as I would like to stay in the plugdata environment to take advantage of the patch in the DAW.

Freqshifting.mp4

Hello everyone! I was wondering if there was the possibility to dynamically change the Hann windowing to change the update rate of the spectrum.
What could be a feasible solution to move from a fast-peak spectrum to a slow-human one such as in the TDR prism?
Thank you for your attention!

rmsslowhuman.mp4