Yes, that’s the microphone, but it’s the plus version. I hear it’s better. I put a towel and book under my mic and redid the forum test, which I’ve attached to this email. This raw sample is -55.4 dB.
I imported this waw file and did the mastering on it, then noise reduction, applying 6, 6, 6. The noise is now 65. 7. Now that must be good enough, right?
This raw sample is -55.4 dB.
I get -55.1dB, but yes, the evil, rumble, vibration sounds are gone. That must have been building noises from the floor.
There you lose me. I mastered with the three tools and ACX Check gives me noise of -55.3dB. Mastering usually boosts voice volume, but the first tool, Filter Curve takes out some rumble sounds, so in your case, it came out almost even.
I applied Noise Reduction of the beast (6, 6, 6) and got -61.3dB not -65.7dB, and no, that’s not enough.
It does meet ACX specifications and you could submit that way, but if you do anything wrong anywhere, the reading is going to fall out of compliance.
You could opt for a much stiffer Noise Reduction. 12, 6, 6 will give you an easy pass.
There is another forum poster who failed one of the chapters in the middle of her book. We’re working out the best way to rescue it without damaging her voice.
Koz
Filtering out everything below 70Hz, then applying a Notch at 120Hz,
reduces the noise-floor measurement by over 16dB …
Thanks, Trebor. I appreciate all those specifics 
120Hz is what power systems in the US do when they’re not feeling well. That’s fluorescent light buzz and some electric motors. I have an older CFL (spiral) bulb which makes tones like that and many lamp dimmers can make that sound.
I had one really bad LED lamp which would buzz at that pitch.
If you have any of those things, you might be able to just stop making that sound and not have to worry about filtering it out in software. Do you have a glass monitor (remember those)? They can radiate at that pitch.
How are you reading your script? If you get your laptop, phone or pad screen too close to the microphone you might get that sound. It’s one of the reasons I still read from paper.
You don’t have an analog microphone, but a bad analog cable can make that sound.
Koz
Once you start reading, you can’t change anything. ACX wants their chapters and segments to match. The beginning and ends of the book should match, too. There is a New User effect. You start reading the book an amateur and you finish a seasoned professional. That’s when you realize how bad the beginning of the book is and want to read parts of it again.
Some of the changes and settings we’re using are married to your room. If you change the room, we may need to design new settings.
Now you know what a nice, stable, quiet studio does.
There is a way to send a short voice test to ACX for evaluation. Read through this post.
https://forum.audacityteam.org/t/how-to-send-a-test-to-acx/49588/1
Koz
Filtering out everything below 70Hz, then applying a Notch at 120Hz,
reduces the noise-floor measurement by over 16dB …
You maybe able to just use the notch. Audiobook Mastering takes out everything below 100Hz, so the 70Hz filter isn’t needed. Another reason to use all of Mastering as a matched suite of tools and not try to pick and choose which steps to apply.
Koz
Thanks, Koz, for all those pointers. No led lights in my room. I don’t think I have a glass screen. I read from my Huawei phone—very similar iPhone, but way, way cheaper—but it’s not right next to the mic. At least I don’t think so, lol. It’s about ten inches away. Maybe I can increase the distance.
Another question:
Once you have mastered and applied noise reduction on your files—in other words, they are ready for submission—does it matter if some are -67.5dB, others are -62.3dB? Isn’t it okay to submit all AS LONG AS they don’t cross the -60dB line? If you say it’s not okay, why would they put the 60 number in the guidelines?
Not sure what this means. Why is 65.7dB not enough? And what do you mean if you anything wrong? Isn’t that measure, 65.7, for the entire chapter? What could I do wrong AFTER that?
You can submit anything quieter than -60dB. That’s the ACX Limit. The guideline is -65dB. The problem with consistently using the limit is the variability of measuring. Note that one forum posting where you got a different Noise number than I did on the same reading. That can’t be the difference between -59.5dB and -60dB. only one of those will pass, and it’s possible ACX will reject both.
Audiobook Mastering has guidelines built-in. The limiter setting, for example, is slightly better than the Peak requirement. You should do that with Noise, too.
I posted the method to submit a test to ACX. You can do that and see what they say. Note the requirements are slightly different than the forum test. In their case you can speak up to five minutes and their silent portions are different.
One correction on your silent portions. Yours aren’t silent. I can clearly hear you moving around and breathing. You can’t do that. The goal is to measure the sound the room is making. That’s why another name for it is “room tone.” Freeze and hold your breath for those segments, breath and then speak.
Koz
Just to avoid fuzzy words problems. Noise is everything in the room except your voice.
ACX is sensitive to distractions. Anything that distracts from the enjoyment of your clear voice is forbidden. That’s why they hate too many effects and filters applied to your voice. Nobody is going to pay money to hear bad cellphone voice, for example.
There are exceptions even to their limits. There is a noise that some USB microphones make that sounds like mosquitoes close to your ear. ACX will reject voices with that noise even though the voices pass all the regular noise measurements.
Koz
