AES-17 RMS metering versus DAW RMS

[bennychico11]

A new scope/metering plugin came out for Pro Tools (here for those who are interested) and I noticed something in the manual that kind of confused me. Digi states that it uses "true" RMS metering and explains that it's different than the AES-17 RMS standard method in that a -20dBFS peak sine wave does not show -20dBFS RMS in Pro Tools...but rather it shows -23dBFS.

Doing some brief research on the web, I read that this might have to do with Pro Tools and other similar programs (I have another metering plugin that shows the same thing) referencing a square wave instead of a sine wave. Sure enough, a square wave shows -20dBFS for peak/RMS.
Any idea as to why it is done this way? Is there a mathematical reason why it would be -3dB off the suggested standard RMS scale? How many other programs or digital gear are like this...?

 

[mshilarious]

How many other programs or digital gear are like this...?

Wavelab shows a full-scale sine wave as -3dBRMS. I mean, hey, that means you can get it 3dB louder, man Square waves rule!

 

[SouthSIDE Glen]

Doing some brief research on the web, I read that this might have to do with Pro Tools and other similar programs (I have another metering plugin that shows the same thing) referencing a square wave instead of a sine wave. Sure enough, a square wave shows -20dBFS for peak/RMS.
Any idea as to why it is done this way? Is there a mathematical reason why it would be -3dB off the suggested standard RMS scale?

Benny,

Part speculation, part mathematics here. Take it for what it's worth unless/until someone can provide a perhaps better answer...

A square wave is going to show a higher RMS because it (theoretically) immediately jumps to it's maximum level (there is no "slope", per se) and stays there for a full half-cycle, then immediatly jimps to maximum level in the opposite polarity. In other words, theoretically it is always at it peak value. (In reality there is almost always some slew and other distortions, but the theoretical model is close enough.)

An unclipped sine wave, however, is only momentarily at it's peak value before it starts sloping back down (or up) to the next wave peak.

Therefore, because RMS is a measurement of an average value *over time* (that time usually being longer than one cycle), it will always measure the average level of a sine wave as being less than that of a square wave because it is spending much less time at it's peak value. So if PT is using a square wave as the reference, it's measurements will appear lesser than those who use a sine wave as a reference, because referencing against a square is referencing against a literally higher-values waveform.

As to why PT is using the square wave measurement, I don't know. Part of me says they are perhaps catering to the RMS Wars crowd in a couple of potential ways, but even *I* am not quite so jaded as to want to believe that ; I'd like to believe they have a more technical reason for it.

G.

 

[bennychico11]

Thanks Glen
Now I understand the idea behind square vs. sine...and should have guessed that to begin with {slapping my forehead for not seeing that before}
Although, I think this may be the reason Bob Katz was saying there is a discrepancy between VU and RMS meters.

I'm going to try and read over the AES-17 white paper this weekend to see what it says about digital audio standardization.
I'd appreciate anyone else who can possibly give me information on the idea behind using a square wave instead of a sine wave for reference. Apparently several programs do it this way (Wavelab, PT, Sonar and various plugins). It's probably for a much far simpler reason than I'm seeing. And of course, it won't change anything I'm doing...just interested.

But maybe this is one of those times Glen needs to remind me to get the hell away from the meters before I break something.

 

[SouthSIDE Glen]

But maybe this is one of those times Glen needs to remind me to get the hell away from the meters before I break something.

Now, now, Benny, I was going to be vey nice and not even go down that road here. We've already been there, done that, got bored on that one

I will make one comment about the RMS vs. VU metering. This may not be what Katz was talking about, I might be highlighting the wrong part of the subject, but one inherant difference is that RMS is based upon a specific mathematical calculation, whereas VU can vary depending upon exactly what 0VU is calibrated to on the particular machine (it may be calibrated to line level in going into the preamp, it may be calibrated to reference tape saturation, etc.) Add to that, if you're talking about analog needle VU meters, you have the whole physical ballistics thing affecting the reading as well.

G.

 

[bennychico11]

Thanks again Glen

 

[flatfinger]

WL6 lets you choose;
Preferences/General/ then checkbox.......

" Use AES17 standard for RMS values"

Default = selected.



This was posted by David Spearritt at the cubase forum:

Posted: Fri Jul 21, 2006 11:06 pm Post subject:

--------------------------------------------------------------------------------

Not sure if that Katz article is all that useful on the "philosophy" behind AES-17. The K-system doesn't really have that much to do with AES-17.

AES-17 is a standard on testing digital audio gear and its main context here in Wavelab is to allow the RMS value of AES-17 metered sine wave to equal the Peak value. This is mathematically anomolous, but is a helper function to try to get the audio industry talking the same language when confusing peak and RMS metering. Most people never quote Peak or RMS when they speak of levels, so if AES-17 is used for metering, it doesn't matter, they will be talking the same levels (albeit only for sine waves though). Also an RMS meter is more consistent than a Peak meter as the latter depends more on ballistics.

To me, I prefer the mathematical accuracy of WL original meters and always quote RMS or Peak when I talk about voltage levels. I also tend to be focussed on dBFS levels when mastering rather than RMS as I leave the dynamics largely alone in my classical chamber music projects.
_________________
David
http://www.lodestarrecordings.com.au