Noise Floor

Hi there,
I’m extremely new to the sound engineer game and audio in general so please forgive my ignorant questions. I’m starting to record an audiobook. I’m using Garageband as my DAW and I’m exporting high-quality WAV files into Audacity to master each chapter file. I have installed the ACX Check plugin. I have set the Amplitude peak at -3.2 db so I am passing the Peak level test. But I continue to fail the RMS and Noise floor ACX checks. My RMS is hitting -41 db (too quiet) and my Noise floor is hitting -237.44 db (which is way too low as it should be -60 db max). Mo matter how much I turn up my gain on my id4 Audient interface while recording my vocals and try different techniques, I continue to fail the tests. I have tried installing the RMS Normalize plugin but can’t seem to add it to Audacity’s list of plug-ins list to utilize it. I have also tried using the “Factory Preset” to use the “Low roll-off for speech” tool. Nothing is working and I’m super frustrated because I can’t start recording my audiobook until I pass these tests.
Please help and please keep your instructions very simple as I’m a baby in all of this.
Thanks for your time.

Step One. Do not read a whole book before you resolve the technical issues. There are errors that can’t be fixed later and you might be reading it all again.

This has all been rejiggered for use in Audacity 2.4.2. The tools (except ACX Check) are built-in now.

We publish an audiobook mastering suite of tools. That will guarantee Peak and RMS (loudness) and if you read in a quiet, echo-free environment may get you all the way past technical testing.

Don’t let that scare you. It’s just an exhaustive description of how to use three tools.

Screen Shot 2020-10-20 at 10.02.39 AM.png
The tools clean up after each other, so don’t take them out of order, add any, or leave any out.

You can shoot your voice in any app you like, or even not use the computer at all.

Announce a clean voice test (no processing) and post it on the forum.

Read down the blue links. They’re very short and they address common errors.


You almost told us which microphone you’re using. We may have tricks for using it if we knew what it was.


A word on decibels. They seem to work backwards but the numbers are all negative, so it’s OK.


That suggests a noise-gate has been applied at some point in your recording-processing chain.
The gate could be in the hardware device, (a physical dial on the interface), or software thereafter.

If the interface or associated software is gating it will be possible to switch it off.

If the interface or associated software is gating it will be possible to switch it off.

Which snaps us back to knowing what your microphone is and how you have it connected.


Thanks so much for your advice. I’m using an Aston Origin condenser mic. This mic has got a built in pop filter but I’m also using a cheap $15 pop filter to try to reduce my “Ssss” sounds. I have figured out a way to pass all of the ACX checks now (thank God - that was not easy) but I still hear a lot of pop and clicking in my vocal tracks. I tried turning down the gain on my interface to avoid hearing these as obviously but the “Ssss” are still a problem. I did some research and found that Audacity has a “De-esser” and “De-clicking” plug-ins and added those to my Effects choices. But I’m wondering if I need to be careful about the “Macros” order that I created for my audiobook file. My husband helped me create a Macro chain that first does the “de-esser”, then the “de-clicker”, then the filter curve, then the RMS Normalizer, then the Limiter. Does this order make sense or should the De-esser and De-clicker be in a different order (if I need to Compress or EQ my audio files at some point in the mastering process)?
Thanks again for all of your help and advice.

Yes. I was trying to originally record my vocal takes in Garageband and had put a Noise Gate limit of -64 dB but now I have decided to record straight into Audacity as they have better plug in options that are easier to navigate. Forgot to mention that I’m plugging my Aston Origin into an Audient iD4 interface using XLR cable (using Phantom power on my interface).

Before you go riding off in all directions, we need to make sure you’re using the microphone properly. The instructions claim you don’t need any DeEssing and the microphone has no harsh gritty sound.

It is strongly recommended you record with proper volume and lip spacing and do everything else in post production. It’s not a good idea to apply gating and other corrections during the performance. We can’t take those effects out later if something goes wrong.

Remember the voice test?

Stop changing things around, announce one of these, and post it.


As we go.


I have attached a voice test as you suggested. Thanks again for all of your help.

I heard back from the sales engineers and yes, you are supposed to be speaking into the side grill, just up from the company logo.

Or I guess it would be just down from the logo if you decided to hang the microphone. Up or down doesn’t make any difference to the sound.

As we go.


I’m trying to decide which stories I want to hear in that voice. Jane Austin, perhaps.

The volume is super quiet. When you look at the blue waves in the timeline, they’re supposed to be something like this.

Screen Shot 2020-10-21 at 1.27.17 PM.png
Your waves are almost straight flat lines. Your Audient interface has a flashing light volume meter on the front. -36, -18, -12, -6. You should adjust the MIC GAIN so you regularly get -12 flashing with occasional -6 when you speak. Nobody will be shocked if the knob runs all the way up.

You should be about a Hawaiian Shaka away from the microphone for normal announcing

or a power fist away if you have a blast filter.

It’s noisy. I made the first second louder on purpose so you can hear the noises. Is that your computer cooling or air conditioning back there? Can you tell the computer is on just by listening?

Some of these fixes may affect the voice quality, so let’s get these down first and then worry about tongue ticking or sibilance.


Did I frighten you? It’s a new user mistake to assume we need to apply a long list of corrections, filters, and effects to everything. I submitted a simple sound test to ACX and it sailed right past basic technical acceptance. I failed because I can’t read. Different problem.

Were you cable to get your speaking volume up? It doesn’t have to be perfect, but it does have to be closer than your test. There are a couple of tricks to getting a little more volume.