I know what does it mean Root Mean Square. But I can’t find any information what is typical portion (length) of audio signal for RMS envelope detector in dynamic audio compressor.

Of course I know it depends on compressor design, there are many compressors with RMS detection. But I wonder is there any reccomendation what gives resonable good results? Like for example should it be longer than 1 second or less? I mean some compression for audio mastering or mix busses?

I don’t have any idea, should it be just little bit of second like 0.1 second? Or better more seconds like 2.0 or even 3.0 seconds?

Could anyone give some advice? I would be very appretiate.

And what is also interesting, how about attack and release time for RMS detector compression? It seems to be stupid to use shorter attack time than length of RMS portion pf audio. But maybe it is OK?

  • 1
    $\begingroup$ well, as you noticed, it depends on how dynamic you want your compression to be. I'd expect voice comms compressors to be more aggressive than say one used in a symphonic metal production, where the build-up of volume might be an important quality of a song, just as a stupid example. With typical vocoder frames often being in the time scale of 20 to 40 ms (+ 50% overlap), I'd just, from a laziness of implementation point of view, expect the RMS averaging to be on the same order of magnitude, but: [citation needed]. $\endgroup$ Oct 27, 2019 at 13:05
  • 1
    $\begingroup$ it really should be adjustable. and it might depend also if your compressor has look-ahead level sensing. $\endgroup$ Oct 27, 2019 at 18:11


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.