RMS Normalizing

There is a new “RMS Normalize” plug-in available that may be useful to people engaged in producing audiobooks.
It is available here: https://forum.audacityteam.org/t/rms-normalize/40902/1

Hi Steve- thankyou for this.
Without any fiddling, I’m at an RMS of -31.1 dB and peak amplitude of -6.6453. I need to adjust this to fit the ACX perameters of between -18 and -23 and peak -3.
Can I simply use the RMS normaliser to within the spectrum (say -20) and then hard limit the peaks to - 3?
I suspect its much more complicated than that…?

Sorry- forgot to attach a clip for your reference- any help on this would be greatly appreciated…

Your recording level is a little low. That means that you will need to amplify the recording, and that will raise the noise level as well as the signal.
A good level to aim for in your recording is for a maximum peak level of about -6 dB. That gives you a waveform of about half the height of the track when in the default “Waveform” view.

Have you tried just amplifying/normalizing that audio clip to a peak level of -3 dB (as required by ACX) and measuring the RMS level? Try it. I get an RMS of -21.5 dB, which sits close to the middle of the specified (-18 to -23 dB) range, and there are no massive peaks that stand out as being much higher than the rest. In other words, no compression or limiting required.
firsttrack007.png

Cheers for getting back to me Steve, really appreciate it- The actual track is 30 minutes long so I think there’s a couple of discrepancies that are altering with the values.
Yup, my first port of call was to amplify to -3db peak and the RMS worked out at 27.4. Here is a screen shot of the full waveform after amplifying to -3db peak. I’m going to need to compress first right? Because there’s a couple of odd peaks? Or am I best off Isolating those peaks and reducing their amplification before raising the whole thing? It’s been a year since I last did this and I’m rusty I’m afraid!
Chapter1wavafterAmptominus3dbpeak.png

Could you post a short sample at about the 12:30 mark. I see there is a big spike there.

Also, when you make a screenshot for the forum, please reduce the size of the Audacity window first so that the image fits better in the forum page - it saves a lot of scrolling for those trying to read your post and view the image.

Thanks Steve- this is the spike with the amplify to -3b.

:slight_smile:

Taking a somewhat ‘unscientific’ approach, I’ve roughly drawn by eye a line showing a kind of “average peak level” of your recording as a whole.
envelope.png
I notice that there’s a period of a few minutes where the peaks get higher and higher, then lower again until it finally settles down at about 15:00 back to a similar level as the first 10 minutes. I’m not able to tell from the images ‘why’ that is - perhaps you were becoming enthralled in the drama, which is a good thing from a ‘performance’ point of view, but not so good from a recording point of view.

Before you do anything, ensure that you have a good backup copy in WAV format, preferably “32-bit float WAV”.

I’d suggest that you try a bit of compression to even out that rise in volume.
It’s difficult to judge the settings from just a short sample, but as a starting point I’d suggest something like this:
fullwindow-Compressor-000.png
I’d advise not going much above 4:1 for the compression ratio, and keep the attack time at a couple of seconds or more. The control to twiddle with is the “Threshold”.
You will need to normalize to -3 dB after using the compressor.

I know this topic is quite old, but I was wondering if there is a way to use this RMS-Normalize function in a chain? I’d would like to batch process multiple files with it, and having to open each file and run it can take a long time…

Thanks!

Yes, once the plug-in has been installed it can be used in a Chain.

perhaps you were becoming enthralled in the drama, which is a good thing from a ‘performance’ point of view, but not so good from a recording point of view.

This is where zero latency monitoring and a good pair of sealed headphones comes in. You’re much less likely to have wide volume swings when you’re listening to yourself in real time. It’s instantly obvious that you’re fading out when your headphone volume dips at the same time.

Highly desirable instead of trying to fix it in post production.

Unfortunately, your recording system has to be designed for that. You can’t plug your headphones into the side of the computer.

Many of the newer production interfaces and mixers have Zero Latency Monitoring and a headphone connection.

What you’re really supposed to be doing is watching the Audacity recording meters out the corner of your eye while you record.

Koz