FabFilter

Hi guys, I'm testing out FabFilter Pro-L against my Neutron and as much as I was amazed by the former, the difference in RMS output puzzles me out quite a bit. Especially how huge it is.

I use both on the master bus with the exactly same dynamics in front of them aimed at -6db RMS.

Neutron limiter gets me there with about +3 db gain on the output while the Pro-L requires +7.7 db and introduces significant distortion.

If I level match both against an external LUFS meter, Pro-L's RMS is hovering around -12 db while Neutron is still around - 6 db and they do sound similar.

What am I missing here?

Hi Medito,

In Pro-L's normal meter scales, (16dB, 32dB, 48dB), the RMS calclulations are just that: root-mean-square of the input samples. There's no pre-filtering, no weighting of any sort, etc. Neutron probably uses a different type of RMS-like loudness metering (EBU R128 / ITU-R 1770) where pre-filtering (K-Weighting) is used, and channel signals are combined differently, resulting in a different measurement.

In a future Pro-L update, we will surely support the current widely used metering/loudness standards like ITU-R 1770, and EBU, and will drop the use of "plain" RMS calculations like Pro-L uses now.

Ok, so the weighted RMS is supposed to deliver what exactly? Loudness metering closer to human perception, as contrary to strictly mathematical calculation of pure RMS?

I do find Pro-L's K system metering highly informative and useful but if I understand what you say about the pure RMS correctly, there is no real informative value and I can/should have the plugin switched to K system all the time.

Hi Medito,

I think there's some confusion. Pro-L's normal metering (16dB, 32dB, 48dB) uses plain RMS calculation, no pre-filtering.

Pro-L's K-Scales (Bob Katz's K12, K14 and K20 scales) follow the K-Metering specs and -does- use a form of pre-filtering + a 3dB offset. So the K-Scales from Pro-L surely offer valuable information! But it might be different from loudness readings that use a ITU-R BS.1770 scale...

... Note that the 'K' from Bob Katz's K-Metering scales has no relation to the 'K' from K-Weighting I mentioned earlier (which is described in the ITU-R BS.1770 specs).

Ok so what would be the reason - if any, to use the plain RMS over the K-Metering?

It's a matter of preference.

When we first released Pro-L, the ITU metering standard were not as common as now, otherwise we would have used the "momentary" standard for the regular RMS meters. For example Pro-C 2 does use this standard for its RMS meters.

When not using one of the K-scales in Pro-L, the RMS level is best interpreted as a relative measure of loudness rather than as an absolute value. Like Floris said, we will improve this in a future update.

Cheers,