Passing RMS on Audacity 2.1.1

Hi there,

I am submitting an audiobook audition and am having trouble passing the RMS part of the acx requirements. I am running Mac OS 10.6.8 which means I’m unable to upgrade to the newest 2.3.3 version of Audacity, limiting my tools. I have installed/applied the RMS normalization, Limiter and ACX check. What else can I do to pass? I’ve attached a sample so you can hear what I’m working with.

Thanks in advance, I’ve been able to do a lot of troubleshooting on my own just by reading the forum.

RMS Normalization should work… It’s not working for you?

See [u]Audiobook Mastering[/u].

It should be easy to adjust the volume (usually up) to meet the RMS, spec, but then you usually need to apply limiting to bring the peaks down into spec.

Boosting the volume will also boost the background noise and noise is usually the “big problem” for anyone who doesn’t have a soundproof recording studio.

My mic volume displays as all the way up. In fact, it doesn’t show that I can adjust at all (USB mic). Is there some other way to adjust the volume?

I don’t have the option of adjusting the volume on my USB mic. Frustrating. Any other ideas?


Somehow I managed to confuse the forum messaging system. Maybe it’s my Special Power.

There are some really exotic ways to mess up Audobook Mastering, but for the most part, if you apply the tools correctly and in order to a reasonably recorded voice track, you will pass both RMS (Loudness) and Peak. Nearly guaranteed. If you record badly, you won’t pass noise, but that’s not your problem right now.

I pulled down your test sample and applied the three mastering tools, in order, and your work easily passes ACX Conformance.

Screen Shot 2019-12-06 at 14.41.08.png
So the question is what did I do that you didn’t.

I have installed/applied the RMS normalization, Limiter and ACX check.

That’s not the Audiobook Mastering Suite.

The official posting is here:

And the abbreviated version is this:

When you read, your first action when you get to the last sentence of a chapter is File > Export > Export as WAV (Microsoft) 16-bit. That’s your protection sound file if “Anything Happens.”

MarysLittleLamb-Chapt1-RawRead-2019-12-06.wav For example. Do not use punctuation marks in filenames other than underscore and -dash-. Use the ISO version of dates as in that example. No slash marks.

Do Not use MP3 for anything until the last step when you prepare the work for submission. Use WAV. MP3 creates distortion each time you save the work, it gets progressively worse, and it’s permanent.

So go over the work again, or a test, and include all three steps. The first step, equalization, is going to be different when you shift to Audacity 2.3.3. That’s just the way it is in product upgrades. Do Not upgrade anything until we get this settled.

Make sure you have the track selected with the button on the left when you apply the tools. That may be what happened to you. As I said above, it’s should be nearly impossible to get RMS and Peak wrong with the Mastering Suite corrections.

Let us know.


Hi Koz, thanks for this.

In the Audiobook Mastering Suite, it says that Audacity 2.1.3 is required. I have 2.1.1 and can’t upgrade at this point to anything other (due to my OS).

Still, if it worked for you, it should work for me, right? I’m going to make a new recording and try the steps again. In 2.1.1 though, there is no “Low roll off for speech” option in the Equalization effect. So that messes the order up right off the bat. Is there an equivalent setting for 2.1.1?

Going in…


Is there an equivalent setting for 2.1.1?

Maybe, but there is a way to add Low Rolloff to earlier Audacity versions. It’s not that complicated. It’s a rumble filter.

Home microphones sometimes have sound errors down in earthquake, thunder and large truck tones. This filter is designed to get rid of most of them in addition to actual thunder, trucks and earthquakes. This can be important because those tones can mess up the other tools and, of course, most times you can’t hear that the problem is because the pitch is too low.

But your cat can.

Anyway, I need to drop for a while.

Don’t get all excited about posting to ACX when you fix this. You have some sound quality issues, too.


Sounds good.

I read about the rumble filter option in another forum and tried it as my first equalization step, no dice.

I won’t get excited about posting to ACX, but the cat will be pumped.

You sound like you don’t think the equalization/rumble will even fix my RMS fail. You think it’s a crappy mic?


the cat will be pumped.

There is a video of a room with three or four cats in it napping (as cats do). For no apparent reason, they all wake up, pause, and go nuts running for cover. Beat. Beat. And the room starts shaking from an earthquake.

You think it’s a crappy mic?

Not yet. You have some odd symptoms that need to be resolved before we go much further.

I found the original production files in the back of the sock drawer.

LF_rolloff_for_speech.xml (299 Bytes)
LF-Rolloff.txt (1.79 KB)

The XML file is programming you can add to your Effect > Equalization to produce the rumble filter. It’s correct in spite of being a different name. That was the original development name when only the techies had to worry about it.

The text file is something I apparently wrote to explain how to install it [blowing off dust].

When you get done installing it, run it according to the instructions in the mastering suite and it should give you a curve that looks pretty much exactly like the above picture file. It’s not important to hit the Length or quality of the curve exactly. 5000-ish is OK.

None of this is exotic except inside Audacity. This rumble filter appears as a simple switch inside many field sound mixers and you can even buy this filter as a physical thing to plug your microphone into for noisy field sound work. It’s sometimes known as a wind filter. It’s our implementation of that tool (Steve, I think) and it’s proven to be insanely handy for audiobook work, particularly with New Users, noisy environments and ratty home microphones.

Post back if you get lost anywhere. I will just write at full steam until somebody stops me.


You can open up LF rolloff for speech in any text editor (Text Edit) and read it. The instructions are in semi-English Extensible Markup Language (.xml).

If you do open it up to look at it, don’t save anything.


On to the theater issues.

I think you’re too close to the microphone. You have that speaking and breathing directly into my ear effect but that’s nothing compared to the P-Popping.

Open your posted MP3 and listen carefully. “Miss PPeregrin’s home for PPeculiar children.”

That’s a rejectable offense right there.

You might back away from the microphone. The general recommendation is one Hawaiian Shaka spacing.

…or use a pop and blast filter with closer spacing.

Those wind screens are different depending on your microphone.

Or you can try moving the microphone.

The microphone (or the sensitive parts) should be straight in front or up slightly.

It should never be opposite your chin or lower. That’s where your breathing sounds go. Please note that’s where most home microphone desk stands put the microphone.

Screen Shot 2019-12-07 at 9.45.21.png
Run away.

You can also try moving the microphone around your head.

That one has the advantage of leaving space right in front open for the script. That one is more susceptible to moving your head and voice tonal problems—but much less likely to pop Ps.


The coffee is just now kicking in. You don’t need Audacity 2.1.3 for the equalization tool, you need it for RMS Normalize.

That’s a lot harder. RMS Normalize was designed for audiobook mastering. There are no tools I know of which can work directly on loudness. Without that, Audiobook reading is a hit and miss operation with experimental volume changes and inspections and eventually settling on a least-worst sound track.

Track to track volume matching is a nightmare.

If you claim you used RMS Normalize, that could be where your voice volume went. RMS Normalize just didn’t work right on the wrong Audacity.

Well…shucks. I don’t remember the really old spells. The sock drawer isn’t going to help.


You can try this. Start with the clip you posted.

Select the track by clicking just right of the up arrow at the left of the track.

Effect > Equalization according to the instructions in the mastering document and your fresh, exciting, new LF curve.

Effect > Normalize according to these settings. This is the old Normalize.
Remove DC, Normalize to zero. > OK

Screen Shot 2019-12-07 at 12.22.08.png
Then measure the results with Analyze > Contrast > Measure Selection. Write down the Volume number > Close.

Screen Shot 2019-12-07 at 12.22.44.png
Subtract -20 from that number. I got 2.2 in the example.

Effect > Amplify > Amplification and put that difference value in the top box. Allow Clipping > OK.

Screen Shot 2019-12-07 at 12.23.32.png
Now run Limiter from the mastering instructions.

That should get you within spittin’-distance of perfect RMS volume. My clip passes ACX Check.

You may need to lie down for a while.

Screen Shot 2019-12-06 at 14.41.08.png

Mac OS 10.6.8.

If you do decide to upgrade your Mac, you should know that the most modern machines, OS-X 10.15.1 and higher (Catalina), don’t support Audacity. So you could improve yourself so much that everything stops working.

If this was easy anybody could do it.


Ok, wow. That’s a lot. I downloaded the new LF curve no problem - thank you. Still failed :slight_smile:

I will back away from the mic (geesh, I even have a pop-filter, go easy…)

I’ll dive in on the rest and report back. I’m glad you warned me about updating OS because I literally just decided that I was going to do that. Maybe I’ll slow my OS roll a bit and figure it out here first.

Seriously, thanks for all the help. It’s stellar.


I passed. And died. I think my machine started smoking a little. If I did, I would too.

Can I get excited now?!

Now, I’m going to try a new sample and (good lord) stay a little further away from the mic. I’m a speech path and heard all those of those terribly over-aspirated /p/s too. Don’t get me started on the /t/s.


So, now can I build a chain for that process – or I’ll probably have to measure that contrast each time, right?

Also, my current mic is borrowed, so I’ll be buying a new one. Do you have a favorite for audiobooks?


I did this using Windows plugins in Audcaity, (I don’t speak mac) …

I think this demonstrates the best you can squeeze out of your current mic.

I’ll probably have to measure that contrast each time, right?

Analyze > Contrast is the only tool in that bunch that runs in RMS. All the rest of them work in Peak, the tips of the blue waves. That’s why it was such a big deal when RMS Normalize was developed. The entire middle of that forum posting collapsed into Effect > RMS Normalize > OK.

There is even a Macro to smash the three mastering tools into one command, but it’s not recommended because of instability in Effect > Equalization. And that was the death of Effect > Equalization. The latest Audacity has that tool in two individual tools that were intended to be much better behaved.

Screen Shot 2019-12-08 at 1.20.11.png
See in the blue waves, they’re not all the same shade of blue? The darker blue are the peaks and tips of the show. Those are easy to calculate and work with, but they don’t have a particularly good relation to sound. The lighter blue is the loudness or “work energy” values. Those are a pain in the neck. Those are ‘area under the wave’ calculus which happens to work out to perceived loudness.

And that gives you two of the three ACX values. How high up and down do the tips of the waves go (Peak) and how loud are they (RMS).

The third value is how loud is the sound when you stop talking? How noisy is it?

I’m sure you noticed by now that the bouncing sound meter and the blue waves don’t match. The sound meter is in dB (proper for sound), but the blue waves are in percent. You just have to know that 50% waves corresponds to -6dB sound. Nobody else does that. That’s an Audacity oddity we hope to solve in a future release.

Your posting is remarkably quiet for a home reading. What is the microphone? That’s my “white knuckle” theory. When you get something that works, hold onto it with white knuckles.

Do you have a favorite for audiobooks?

I bet you’re waiting for a model number, right? In my opinion, new users shouldn’t be using USB or other computer microphones. If you inspect the forum postings, a large majority of them are to resolve troubles with USB microphones. “What’s that mosquito whine sound behind my voice?”

I call that the “Yeti Curse.” It’s not exclusive to them, they just happen to be an insanely popular microphone, so statistically, they appear more.

The last sound test I did was a stand-alone sound recorder, the Zoom H1n.


It passes ACX with simple mastering and no noise reduction.

That’s no joy to work with either, but it’s stable and all the forum pages and pages of USB microphone complaints vanish. Boop. Gone. People are using other variations such as the H4n.

That picture is a little Hollywood Illustration License. Replace all the wood with my quiet bedroom. That’s oblique microphone placement (not straight in front) and the paper towels get the microphone up to about nose level.

Not everybody pops for condenser microphones.

That’s Joe Rogan and his Shure SM7. Terrific microphone, but it frequently needs a Cloud Lifter sound booster, always needs a sound mixer and USB interface or USB microphone interface or sound recorder. It’s a straight, classic analog dynamic (moving coil) microphone. Full Stop. No computer interface.

I have an Electro-Voice RE-15 (no longer made). Same type of microphone, sounds terrific, but has all the above shortcomings. I have a small sound mixer and my older Macs have analog inputs.

Home condenser microphones almost invariably have a Crispness Boost which they insist sounds more “professional.” You can’t turn it off. You end up taking the crispness out of your voice in post-production with special DeEsser software because it hurts your ears and scares the cat.

Whatever you choose, you need to stick with it for a whole book. ACX places great emphasis on having your chapters match.

You can send ACX a short test. Your first exposure to them doesn’t have to be massive sound files and months of work.

Someone came up with the best-named microphone in the world. Professional Broadcasting Studio Recording Condenser Microphone. Actual name. The microphone is complete trash and someone found you could buy one for $14 USD.