RMS signal ?

[yphs_mst]

Hi everyone,

Just wondering, what's the difference between 'peak/standard' compressor and RMS compressor (Like DBX 165A)? I mean .. I know what RMS means, but ... Still don't know the technical side of it. Anyone can explain ?

[Rail Jon Rogut]

RMS vs. Peak Link

Compression Link

Rail

[yphs_mst]

Ok .. thanks for the link, that's really helpful, really ! I do understand it ... but let's get back to the compressor, what makes the peak and the RMS compressor sound different ? Well, not technically, but does anyone try to compare those 2 compressors ? Like ... how do you describe the difference by a same setting ?

I tried RMS compressor once (DBX165A), it sounded nice on overhead drum and snare. But I didn't have time to compare it to other things. Anyone ?

[inaudio]

Not to be smartass, but reading and being told are mere suggestions - you really need to use your ears! That said, peak vs RMS sensing is a bit like changing the attack time - slower is smoother - though they really work differently.


RMS means the average loudness is the reference for gain reduction, Peak means just that, the loudest peaks are the trigger to GR.

It is very interesting to me how "high end" forum can conatin such basic questions - and how I always feel I have learned the most about audio in the last five years - experience being the best teacher. It's always the last five years - as true now as it was 35 years ago - where the details begin to make sense from experience.

We used to learn this business by apprenticing or on the job. I did both, for a long long time...

Oh, and you really need to take the time to compare or you will still be guessing...

<L>

[yphs_mst]

Thanks ..

I wish I had access to those compressors easily ... ... i wouldn't ask this thing.

Thanks again anyway

[inaudio]

So, yphs (?!?), My implication wasn't to diss you, it is that these things take time to learn.

Do you have people around with whom you might spend some quality time helping them out and watching them work the gear? Taking a kind of apprentice type thing can be extremeely educational - think of it as free education rather than low paid work! I spent ten years sharing gis with another engineer, so neither of us had to hire idiots for grunt work, and we both learned a lot by working together... Just an idea.

I was lucky enough to find associates when I was younger who could explain and demonstrate things like this - but that is what forums like this are for in this day and age!

B est luck,

<L>

[jajjguy]

Originally posted by inaudio
RMS means the average loudness is the reference for gain reduction, Peak means just that, the loudest peaks are the trigger to GR.

I've always wondered over what time period devices are "averaging" when they detect RMS levels. Is there a standard? 10ms? 50ms?

[lefthando]

Originally posted by jajjguy
I've always wondered over what time period devices are "averaging" when they detect RMS levels. Is there a standard? 10ms? 50ms?

Great question!

[Tim Farrant]

Originally posted by jajjguy
I've always wondered over what time period devices are "averaging" when they detect RMS levels. Is there a standard? 10ms? 50ms?

The averaging will be done by the attack and release circuits. Slow attack and release is slow averaging time (obviously)!

But another note about this. Compressors have a threshold control so if you think about it, it will be the peak of the RMS signal that will at first cross this threshold.

Confusing isn't it!

Tim.

[yphs_mst]

Yea .. .confusing, but make sense to me now ! Thanks again !

[NeoVXR]

ok, technically the PEAK is just that, the maximum of the rectified signal. it has a quick rise time (but you can slow down the thing with the attack control) and needs to be kept up for some time, otherwise it would work just like clipping. so there is need for a remarkable time constant in the falling slope, and you can set it even slower.
RMS is built on an integral on the square value of the signal, and this integral has a time interval to accumulate, e.g. 50 ms.
this prevents instant reaction (but you can introduce "lookahead" in the digital domain) and has the advantage of better recognition of "loudness", esp. with a ear-corrected filter in the side chain.

cool edit (continued as adobe audition) yields an experimental tool in its built in compressor, that is very instructive in my opinion. you can switch modes, set time constants and look ahead, and draw transfer curves. but you need to be careful, crass or unusual parameter settings may cause irregular behavior with blips and spikes, similar to an analogue self made compressor that does not limit the parameter settings to reasonable values.