Subject:The LUFS vs. RMS Metering Mystery

Posted by: Weka

Date:4/11/2013 2:10:53 PM

Hi Guys
I've recently begun using LUFS EBU R128 to measure master levels. All good.
However, during the learning process, I downloaded a pink noise file very reliably purporting to be at -11.5 dB RMS.

The Sound Forge Normalise/scan meter reads 11.3 RMS
Sound Forge Statistics gives me -14.68 RMS
My Tone Boosters LUFS meter is giving me -14.4 LUFS
Klangfreund LUFS gives -14.3 LUFS

Is anyone able to explain the extent of the differences in these levels between the RMS' & the LUFS'

Many thanks
Weka

Subject:RE: The LUFS vs. RMS Metering Mystery

Reply by: Weka

Date:4/11/2013 4:45:21 PM

Right. A little more experimentation has shown me that the discrepancy between the Normalise/scan level and the LUFS' was due to me using a mono file (silly boy!)

However I'm still curious about the considerable difference between the normalise/scan RMS readings and those provided by the statistics. Anyone?

Subject:RE: The LUFS vs. RMS Metering Mystery

Reply by: musicvid10

Date:4/12/2013 8:56:48 PM

RMS measures average levels.
LUFS indicates average loudness.
Apples and oranges. There is no correlation whatsoever between the two.

Here's a little reading to get you started.
http://tech.ebu.ch/docs/techreview/trev_2010-Q3_loudness_Camerer.pdf

Subject:RE: The LUFS vs. RMS Metering Mystery

Reply by: Weka

Date:4/13/2013 2:01:13 AM

Yes, I pulled myself up in the second post (though the way I understand it is that there should not be, in practice, much more than 1dB between the RMS reading and the LUFS).

My question was more about the significant difference between the Sound Forge normalise/scan RMS readings and those provided by the SF "statistics" RMS reading. I'm just curious.
Cheers
Weka

Subject:RE: The LUFS vs. RMS Metering Mystery

Reply by: musicvid10

Date:4/13/2013 7:17:15 PM

1. I've never read that. RMS does not take compression directly into account, so I imagine there must be a lot of exceptions. Again, there is no direct correlation.

2. Everything being equal, there should be no difference between Statistics and Normalize RMS scan levels, save for a small rounding difference. However, if the Normalize scan settings are anything but [0,100%], of course the readings will be different, maybe by a lot.