Measuring maximum RMS amplitude

I’m switching from Audition to Audacity for final checking of materials.

I used Auditon only to check the maximum RMS amplitude and i cant seem to get the same values in Audacity. My workflow is setup to have -18dB maximum RMS amplitude in Audition, and I cant seem to find a tool to do the same thing in Audacity.

I have checked out the RMS, Wawe Stats and even ACX pluging, but I dont get the same numbers as in Audition.

Hi Smoking1950

Try this.
EFFECT-RMS-NORMALIZE.jpg
RMS NORMALIZE.jpg

Do the measurements from RMS, Wawe Stats and the ACX plug-in agree?
Another one that you can try is “Contrast” (See: Contrast - Audacity Manual)

Another test: The RMS level of a full scale (0 dB peak) Sine wave is -3.01 dB.
Which of the tools give the correct answer?

Reference:
https://en.wikipedia.org/wiki/Root_mean_square
RMS of Sine wave = A/(sqrt(2.0))
where “A” = amplitude.
When A = 1.0, RMS of Sine wave = 1/sqrt(2) = 0.707107
Converting from linear to dB:
0.707107 linear = -3.0103 dB

How different are they and who’s higher?

RMS (loudness, area under the curve, effective energy) can be different depending on your assumptions. ACX Check can be different from ACX’s AudioLab because I believe they don’t measure silences and we do. Ours is simpler but can be slightly off.

Then there’s the assumption that RMS equals loudness and strictly speaking, it doesn’t. It gets close and it was a handy measurement that everybody understood. There is an RMS value for the electrical power at your house. RMS insists on measuring sound so high pitch that only dogs can hear it and so low pitch it bothers cats, but humans can’t hear it. This can get you into trouble with home microphones which can have rumble and low-pitch distortion. You can’t hear it, but takes part in sound processing for show delivery to a client.

It’s the reason that Audacity Audiobook Mastering starts out with a filter to get rid of low pitch trash.

That brings us to LUFS which does take all those things into account. What’s the possibility Audition is measuring LUFS but just calling it RMS because that’s what everybody is expecting.

And that brings us back to who’s higher and do all those others match each other.

Koz

maximum RMS amplitude

That makes me nervous. RMS isn’t like Peak values where you can point to one small portion of one musical note and declare the waveform value is some number. RMS is more like a fuzzy average of everything in the performance.

Koz

Yeah, me too. But thats the literal term they use in Adobe. Im not editing the sound files, the clinets send me their files I just check them before I send em to the next persone in chain.

All of the tools in Audacity gave the same nuber -21.6dB or so, and I think you are correct. In audition my target is -18dB

So this is what I get when I scan a file in Audition, the target is -18dB show on the maximum RMS aplitude
2021-09-27 14_36_18-Window.png
And this is what I get from audacity Wawe stats
2021-09-27 14_37_18-Window.png
I tried different files and I cant get the number to match. One file matches the -3dB difference, but this one doesn’t

Look closely at the interface and documentation for Audition’s use of “RMS”. Does it mention “AES-17”?

Unfortunately, in 1996, the American Audio Engineering Society (AES) decided to define 0 dB RMS 3.01 dB lower than the mathematical / scientific / engineering definition in their “AES-17” document. I suspect that the Audition measurement is “-18 dB (AES-17)”, which is equivalent to “-21 dB RMS” as defined and used for over a century in mathematics, physics, engineering and other fields.

But why do i get -18,04dB in Audition and -23.5dB in Audacity?

Thank you all for the advice and help, you are really a great community.

So, to explain a bit why am I doing this, I have to say the final yes if a file is safe to be played in our theatre. For years we used Audition and refrenced the maximum RMS value (nobody knows why). I hoped to make it a bit more understandable for the new colleagues and really wanted to use Audacity for this. How would you valited a file is not “to hot” for your setup?

I don’t know what Audition does.

I do know that Audacity calculates RMS such that a full scale square wave has an RMS level of 0 dB, which is the same as used in engineering / math / science / everywhere except AES-17 and Bob Katz’s “K-system”.

Professional theatre systems are usually calibrated to produce a “Sound Pressure Level” (SPL) that falls within limits specified by local or national law for both continuous sound and short sounds. SPL measurement requires a “SPL meter” to measure the actual loudness of the sound.

Because theatre frequently has sounds that are very loud but short (example: gunshots / explosions), the average level of a recording is usually pretty low. These days, the loudness of a digital recording is normalized to somewhere around -27 to -31 LUFS. This allows sufficient headroom for short loud sounds.
(Note that this is a much lower level than is typically used for podcasts and similar).

Audacity has a “Loudness Normalization” effect that allows you to amplify (or attenuate) a sound to a specified level (measured in LUFS). See: Loudness Normalization - Audacity Manual

But why do i get -18,04dB in Audition and -23.5dB in Audacity?

Audacity isn’t giving you the overall RMS, not the peak. This is close to Audition’s total & average RMS.

[u]dpMeter[/u]* can show you the maximum RMS as well as the overall (Integrated) RMS. It’s the dB-K settings and I think it’s offset/biased by a certain amount. You might have to read the dpMeter user manual and the K-metering documentation and/or just compare the readings to Audition to make your own correction.

RMS has to be measured over-time and the time-window could be different in different applications giving you different “peak RMS” readings. Also, the overall RMS might be an average of (windowed) RMS readings or an RMS of RMS readings, because otherwise the software has to sum the squares of millions or samples.


\

  • If you install dpMeter with the 64-bit version of Audacity, install the 64-bit VST2 version of dPMeter.

All of the tools in Audacity gave the same nuber -21.6dB or so, and I think you are correct. In audition my target is -18dB

You could fake it. Use the -21.6dB Audacity number in Audacity from now on. If it’s consistent, that can be your “Audition Equivalency Value” (AEV)

Also a reality check, have you ever tried to hear a 2dB change? Nobody in the theater is going to know.

Koz

For features we do check the SPL and it’s a complete different process and department. We use this rude method on quick checks on ads. You never know what they could send you, so we used the peak values to make sure there isnt a huge bump in the audio, just to ad a safety layer.

Yeah I thought I could be done with this in a day, x=y and be done with it, but I think I’ll need to rewrite our guidelines and adjust to a more logical way of verifying the RMS and peak of our materials.

Try checking the RMS on some “good” ads (where you are happy with the loudness).
My guess is that the RMS will be around -21 dB, but whatever the figure is, use that as your guideline.