Sample WAV for ACX - Produce a Simple Audacity Forum Voice Test

I followed the directions here:

Audacity 2.3.3
Windows 7 on a Dell laptop that is in the room with the Blue Yeti Microphone
The gain is set to the 9:00 o’clock position, my mouth is the standard pinky-fist-thumb distance away, with a pop filter in between
The laptop is down below and behind the mic, resting on acoustic foam. The laptop fan is on low.
No processing applied to the audio file.

When you have time, could you tell me if this sample can work for ACX mastering? Thanks!

could you tell me if this sample can work for ACX mastering?

I can. No.

I think you’re recording your laptop microphone instead of the Yeti. It’s very noisy and you sound muffled and like you’re recording in a small wooden box. I made the two second Room Noise segment intentionally louder.

Find out where your laptop microphone is (mine is just left of the left-hand shift key). Make a test recording and scratch that microphone and scratch the grill on your Yeti.

Audacity > Edit > Preferences > Devices > Recording. Which thing or device do you have selected? It doesn’t have to say “Yeti.” If there’s nothing there even close, close Audacity with the microphone plugged in and then re-open it. Did the microphone show up? The microphone should be there and working before you open Audacity.

There’s a number of other problem with the clip, but that’s the main one.

Don’t promise anybody weeks worth of reading until we stamp out all the problems. The current record for longest forum posting is Ian who lives in Hollywood (a real city). He had a noisy apartment on a noisy street and it took us over a year to get it all working.


Well that’s embarrassing. You’re right, I accidentally recorded from the laptop mic. I could have sworn that Audacity automagically defaulted to the Yeti when it is plugged in, but I forgot to check. I apologize for wasting your time with that file, but there’s nothing like a public humiliation to sear a lesson into memory. That’s something I won’t forget to check from now on!

If I can impose once again, here’s the recording using the Yeti:

Same parameters as before except I moved the laptop into the adjacent space so it’s not quite so close.

There. That wasn’t too hard.

Screen Shot 2020-02-15 at 7.24.20.png
The specification for noise is -60dB, but it’s a good idea to pass it by at least -65dB (quieter).

There’s a little fudge built into Peak as well. The official specification is no louder than -3dB. The suite delivers -3.5dB (quieter). Nobody can hear a half-dB difference and that gets you past the MP3 conversion errors for file delivery.

I applied Audiobook Mastering Suite…

Screen Shot 2020-02-15 at 4.16.02.png
… from here…

The forum instructions will tell you how to get all the tools if you’re missing any.

Mastering will only get you past the basic sound recording part of acceptance. You also have to pass Human Quality Control who will judge you on theater qualities such as stuttering and mouth noises. There’s no Audacity analysis tool for those, and you have to pass everything.

I see you’re already careful about not including punctuation marks other than -dash- and underscore. You should also be freely Exporting WAV files as safety backups. For example, Export a WAV file of a raw chapter reading, fluffs and all. If Audacity goes into the mud while you’re editing, it’s good not to have to drag out the microphone and read it all again. Just open a copy of the protection backup.

Same when you’re done editing and processing. Export a WAV (Microsoft)16-bit of the final chapter and only then, make the 192 quality MP3 for ACX.

You can send a short test to ACX.

They won’t analyze your studio, so don’t expect a spirited dialog with them.

As a side editorial opinion, I would try to read so it doesn’t sound like you have a gun to your head. You are a performer and you have to be able to be different people on cue. There are readers who go down the book and make up voices for each character. That depends on the book, but it’s a good idea.

Also remember ACX loves consistency. Once you start a book and it’s going well Do Not Change Anything. The start and end of each chapter and the start and end of the book all have to match.


Fantastic! I was able to install the new plug-ins and apply the recommended settings to the suite to get the same results.

I see in the Export>MP3 dialogue where to set the CBR to 192, and the force it to mono checkbox, but the export to MP3 doesn’t have a setting for the sample rate. I’ll do my research on changing that 48000 Hz to 44100.

I notice you don’t apply noise reduction in the process chain - is there a reason? It’s so vewy vewy quiet (as Elmer Fudd would say) with that applied. (It’s NOT applied to the file in that screenshot)

Thank you!

The sample rate for exports is the “Project Rate” (lower left corner of the main Audacity window).
Ideally you would set that as required before you start recording (the factory default is 44100), but you can change later (before you export) if necessary.

Ah, thanks! The Blue Yeti microphone documentation says it’s supposed to be recording at 48000, and I don’t want to make the Yeti mad! :mrgreen:

Thank you Trebor!

Unless you can acoustically reduce the sound of the computer further, I think you’ll have to apply some Noise reduction.
I used Audacity’s noise reduction at 6-6-6, & Couture plugin on these settings to squash down noise even further.

The Blue Yeti microphone documentation says it’s supposed to be recording at 48000

48000 is the video sampling rate. 44100 is the one used with Audio CDs. Either one should work although I believe ACX is expecting 44100. Many people use the Yeti with video and YouTube production where 48000 is a natural. When I shoot for the video editors, everything is 48000.

I think you’ll have to apply some Noise reduction.

Right there we part company. You are a business and as a business, you should be producing work that satisfies the posted requirements and not much more. Every tool, filter, effect and correction you apply over the requirements is a business loss to you. Your noise volume as measured is quieter than -66dB. In English, your noise is half what it needs to be to pass ACX. Taking that upside down, you could double your background noise and still pass (although I wouldn’t do that).

Keep this simple. You still have to pass all the other tasks and specifications to get published and retire rich to a nice cottage on the Mississippi.


I measure the noise floor on that at -64dB,
after the 6dB of amplification required to bring the RMS up to -20dB, that raises the noise-floor to -58dB. (+2dB over ACX).
So it’s a borderline-case for Noise Reduction.

I’m obviously not qualified to weigh in on the technicalities, but IF I were to apply noise reduction, should it come before or after the ACX suite, and at what settings for Step 2? (I know how to get a noise profile and then apply it to the entire project.)

Speaking of projects, I see the advice to keep a WAV file as a backup. Why wouldn’t I just keep a copy of the original project as a backup? Doesn’t it contain more information than a rendered WAV?

Doesn’t it contain more information than a rendered WAV?

It does, but it’s also more brittle. Scan the forum for people having a WAV file fail versus people having trouble opening a Project. Also, people think they’re saving an AUP sound file. Nobody understands the complex interaction between the files and folders in an Audacity Project.

Simple, robust WAV files are good thanks. Nobody wrote you can’t do both.

WAV files won’t do everything. They won’t do multi-channel (over 2). There is a filesize limit, and they won’t save many, many hours of performance. It could be argued that hours and hours isn’t a performance (Ring Cycle notwithstanding), it’s surveillance and Audacity will not do surveillance.

should it come before or after the ACX suite

If you apply it before the suite and ACX Check, you won’t know if you needed it or not. There is a body of users that just blanket apply noise reduction—always—as part of the suite. That can get messy because if you choose a stiff enough reduction setting in Step Two, you can hear it working. It creates sound damage.

Noise Reduction isn’t a gift from the angels. Nobody can hear Noise Reduction of the Beast (6, 6, 6), and 9, 6, 6 is pretty safe, but by the time you get to 12, 6, 6 the performance may fail critical evaluation. That’s why you have to get close with your original performance. You can’t create garbage and depend on the tools to pull you out.

Also see: People desperate to automate Noise Reduction or fold it into a Macro. Not so far.


Probably the best format for backups is to “Save Lossless Copy of Project…”
This provides the benefits of both worlds. The data is rendered as robust and reliable WAV files (“perfect” 32-bit float) and the additional project information.
The only downside is that lossless copies of the project are big - the same size as a normal saved project.

Thanks guys, the ACX suite has already made a big improvement in sound quality. I’m writing some sci-fi, and I’m considering the possibility of narrating my own stories for audiobooks. Meanwhile, I’m making some animations for social media. If it’s okay, I’ll post a YouTube link to one that uses the improved audio when I have one ready.