Processors that detect RMS do in fact calculate "true" RMS, which takes into account the contribution of harmonics. RMS for a complex waveform is, by definition, the sum of the RMS values of each harmonic. 

reaComp comes to mind as one compressor with RMS detection, and features a variable RMS window. The free TDR Feeback Compressor also uses RMS detection, IIRC, as does Stillwell's Bombardier. Ozone's multiband compressor is switchable between peak and RMS. 

But my favorite compressor - for everything - is FabFilter's Pro-C. It's not obvious looking at it, but it does feature an RMS-detecting mode; they just label it "classic style".

EDIT: just thought of another one with switchable peak/rms detection - Klanghelm's DC8C. It's cheap and has about every bell 'n whistle you could want, including a unique S-shaped release curve that's supposed to help make it more transparent.

Hi Dave,

Peaks are more or less instantaneous.

It seems like as soon as you introduce RMS detection that there must be some sort of duration defined with which to esteem an average over.

So, how does that work in practice? Are these compressors look ahead? Do they look behind and anticipate (perhaps with a slow attack potential)? Do they run more effectively as offline processes?

Thanks for any info you can share.


best regards,
mike

It is a good question as I am sure many have wondered what the difference is especially if a compressor provides both modes of operation.

http://www.uaudio.com/web...uly/text/content2.html

http://www.doctorproaudio...s-compressors_en.shtml

http://www.izotope.com/su...multiband_dynamics.htm

Yes, Mike, RMS detection implies a time delay equal to the RMS window, typically 50ms. That would be considered a medium attack for a peak-sensing compressor, but imposes the minimum attack time for an RMS-sensing compressor.

You cannot achieve a fast response with RMS detection, and large peaks can and do sneak past. This could be mitigated with lookahead, which Pro-C does not have, and I don't know offhand of any compressors (except limiters) that do. But RMS-based compression is used for volume leveling, not peak limiting, and so is usually preceded by a peak-sensing compressor to shave off peaks before the signal gets to the leveler.

Thanks!!!


:-)

bitflipper


Yes, Mike, RMS detection implies a time delay equal to the RMS window, typically 50ms. That would be considered a medium attack for a peak-sensing compressor, but imposes the minimum attack time for an RMS-sensing compressor.

You cannot achieve a fast response with RMS detection, and large peaks can and do sneak past. This could be mitigated with lookahead, which Pro-C does not have, and I don't know offhand of any compressors (except limiters) that do. But RMS-based compression is used for volume leveling, not peak limiting, and so is usually preceded by a peak-sensing compressor to shave off peaks before the signal gets to the leveler.

Spot on reply. There are tons of compressors out there and many of them allow switching between peak and RMS detection.