Noise Floor

Usually they put music under the voice, but this one was an ad that went out with nothing under it. The engineer sends me the ads as MP3s after he’s finished doing his stuff.

Oh interesting about the silence. I’m going to have to write down lots of notes! That’s good to know about manually measuring the silence. I didn’t know that. Or, I’d seen it but not been sure what it all was.

At the station, I can hear myself in the headphones when recording. Generally the engineer gets me to read a line or so and adjust mic or myself until he’s happy with the sound. I try not to listen too much to myself once I’m actually doing the “acting” bit as I find it a bit distracting, but I’ll make a point of listening more closely.

At home, my Apogee mic doesn’t really allow for real-time monitoring, the latency is too significant. I can put it through Twisted Wave, but there is a noticeable delay.
It doesn’t matter if I’m just fiddling with gain etc, but I do need to turn that off while doing the actual recording.

That’s also interesting that ACX doesn’t like any of those plugins. Presumably EQ/compression/normalization are okay? I think I’ve seen that video - will watch it again. One of a four or five part set?

This video?

I’m at the “educating myself” stage, before I start auditioning. I’ll watch it again.


ACX doesn’t like those tools because most people overuse them. It’s not unusual for someone to submit a clip that’s been processed to death because they started with a very noisy original recording and then tried to “rescue” it with processing. We regularly get posters who proudly point to a long list of processing tools and filters and want to know what we would add to the list to ensure ACX acceptance.


This is one of the reasons we want you to submit a clip before you did anything to it, and remember when you tell us you’ve been carefully processing your work word by word (actually happens) that you’ll have to do that to months-long readings if you start to read books.

Everybody knows you can just plunk a Yeti down on the kitchen table and read an audiobook, right? It usually sounds like a bad cellphone call when you do that. The submitted clip passes the testing robot (ACX-Test) but then crashes with an “Overprocessing” failure when Human Quality Control gets it.

Gentle use of Noise Reduction is frequently required if you have a slightly noisier than needed microphone and gentle use is undetectable in the clip. That can be a philosophy problem. The goal is not The Blackness of Space silence between your words. It’s just to suppress background noise so it’s not distracting in the reading.

You may have gating applied to your broadcast clip. Very few people can manage a -80dB noise floor from a straight recording, but again, professionally done, nobody can tell what you did.

I don’t entirely agree with the mastering processing presented in the final ACX video. I think he goes right off the rails with compression and volume adjustments. The timeline is not supposed to be a solid block of blue.


I do note that all your posted work so far has been commercial and “presenter” work rather than theatrical readings. That’s my joke of listening to someone telling you a story over cups of hot tea in a comfortable kitchen as opposed to someone trying to sell you “New Scruffy Trash-Can Liners.”

Actual readers/actors make lists ahead of readings of which voice they’re going to use for which character.

My without question favorite reader is Sarah Vowell who has a very oddball reading style, but sticky presentation. She can produce a “driveway moment” where you don’t want to get out of your car when you get home because she has you hooked.


I don’t know we ever found your exact microphone model. My favorite microphone for this type of work is the Samson G-Track. Among it’s significant attributes is echo-free monitoring from a connection in the base.


If you don’t have enough pain in your life, I did write a thing about how to do ACX testing manually.

This is what one-pass ACX-Test replaces. There’s actually a mistake in there I never bothered to fix. The peak specification is actually not to exceed -3dB.


Two thoughts.

I can do this for weeks, but maybe you should stop obsessing and actually read something. You are not a beginner and already have a good self-concept of your talents. You are so not a newbie.

Whatever you do, export the RAW work as WAV (Microsoft) and save it as Archive. There is no “I lost the files and I have to read it again.” I have raw capture sessions I shot five years ago. You know if I delete them, I’m going to get a call from my producer, “Say Koz, you remember that shoot you did…”

And that leads to The First Book Curse. You may walk straight around this, but people read their first book and get so much better in the course of reading they recoil in horror at the first part of the book and decide to read it again. The successful ones do. Some people give it up as a bad job (it’s more work than it looks) and do something else.


All of the recordings I’ve done until now has been broadcast/commercial/presenter for a radio station. The florist clip was a short script from Edge that I picked to use for sound assessment purposes.

Ah, the broadcast clip wasn’t mine. Well, it was me speaking, but the engineer sorted everything else out and it was in a pro booth, so I can’t take credit for the -80dB.

I’m actually about to record a short clip in a couple of different spaces in my home. The main problem with achieving the below -60dB was that I had to turn the fridge off to achieve that value. Ok for a 15 minute session, not so great for 2-3 hours of recording in summer (I live in Australia…)! Ideally I’d find a way to record without having to turn off quite so many appliances.

This is my mic: It’s the older version, not the newer 96K. I like it that I can plug it straight into the iPad, but I have a camera kit for the iPad which does broaden the possibilities for mics a bit (apparently). I’ll check out the Samson G-Track

I will also read through your manual ACX testing. I’m trying to pre-empt any problems, much like you’ve said. Better to have the fridge sound gone before I record!

Yes, I’m almost ready to do some recording. I actually thought my first pass might be to contact the Queensland Narrating Service and volunteer my services to narrate a (short, hopefully!) book for them. That has three advantages - it’s something useful to do for someone who will benefit (the print impaired), it will give me some experience in actually constructing an audiobook, and finally it will give me something that I’ve done that I can point to as an example of work.

I really appreciate all the time you’ve spent with me.

I can’t take credit for the -80dB.

That clip ticked an attention point. It’s too good. -96 is the absolute minimum audio noise where all “molecular activity stops.” Nobody records down there. Given your clip was heavily processed (nobody records high RMS, either. That was another tick), somebody who knew what they were doing processed the sound. I’m impressed, by the way. They did it without any outward appearance of corrections.

I will also read through your manual ACX testing.

I prepared a correction, but I haven’t post it yet. The Peak specification is wrong. It’s not -3.0db to -3.5dB. It’s just lower than 3.0dB. That’s the overload specification. But that is how to use the tools. That’s correct.

I had to turn the fridge off to achieve that value.

And there in one swoop you have the difficulty when your recording space isn’t a studio. Two families before me had a kid that played drums, so my third bedroom is soundproofed. Even with that, I have to stop recording when a Metrobus goes by because the window isn’t sealed.

I’ll check out the Samson G-Track

The microphone was borrowed from one of the production people with the idea I was going to use it as proof of performance for Overdubbing/Sound-On-Sound in Audacity. It worked famously. The owner uses it for his music (he’s part of a band) and I noted that it wasn’t terrible. That’s as far as I got. I never recorded any voice tests or other performances and I never listened to anything he did.

So that’s my recommendation. Not terrible. It does have one possible problem. Because it’s not an expensive microphone, it’s possible it could have The Yeti Curse whiny noise in the performance on some computers. There’s no technical reason for it not to.

Better to have the fridge sound gone before I record!

Noise Reduction works by trying to remove some sounds from your performance. It always affects the valuable performance and the juggling act is how well you can hide the damage. So Noise Reduction isn’t “free” and quiet studio recording is highly desirable. Did you note in the ACX videos, it’s not a question of you having a sealed room studio, but how large and where you were going to put it?

There are other possibilities. You can make a sound tunnel and pile quilts, duvets or moving blankets around. So no permanent room, but perfectly serviceable “studio.”

People make portable tunnels, too for noisy environments where you don’t have access to lots of space. That’s the semi-round thing on the right in this recording.

Queensland Narrating Service

Perfectly delightful idea. By the time you get through that, you will have a terrific idea of your studio shortcomings and any other problems you may have. It will also get you over the First Book Curse.


I patched the document.

You know you have the corrected one because it has two dates at the bottom.


Brilliant, thank you!

I’ve been fiddling a bit more this morning. By leaving my fridge on, I’m not achieving the desired noise floor. I can get there by using noise removal of about 4dB. The default looks quite a lot bigger than that, so hopefully my 4dB is acceptable. (In the long run I want to leave the fridge on…)


The two files, the first one is not messed with, the second one has some messing. :slight_smile:

Oh I keep getting them pasted in the wrong order. They’re in the wrong order but are labelled correctly.

I have not listened yet.

There is a cartoon way to achieve a quieter room. You leave the appliances running and set up your microphone and headphones for a conventional recording except you turn the headphone volume up and don’t speak. Now walk around the room with the microphone listening carefully as you go. You may find a sweet spot where the appliance noises are not so prominent. That’s where you record.

I did this with a hum problem I had in my studio. No matter what I did, there was a hum in the background of my recordings and I found one particular spot and elevation (microphone stand and boom) where it almost vanished. So that’s where I recorded. Eventually, I took the time to do the “hum hunt” by waving the microphone around while was listening and I found one of the musical bass cabinets was bad and made noise even when it was off. Unplugged it and poof, no more noise and no more sweet spot needed.

You have a directional microphone. You may find that pointing the rear or dead direction toward the fridge will help. The cartoon part of this comes when it turns out the sweet spot is a foot off the floor and in one corner.


Oh funny!

You know those people on the beach waving metal detectors around? You do that with your microphone. One possibility is finding a spot where the noise and its reflection from one wall cancel. Mark that place.


What the studio gives you is freedom. After hacking away at one poster who had everything in the world wrong with his system and forum chapter after forum chapter, I gently reminded him he was replacing a recording engineer and studio. Trying to force your apartment to do this is not a given although it is popular to try. Walk in to your studio, clear your throat, present through the work, walk out with the finished sound tracks and wish everybody a happy day.

Full stop.

We are reminded that the all time longest thread in the history of the forum is Ian who wanted to record audiobooks from his apartment in Hollywood. Yes, that Hollywood. It’s a metaphor, but also real place with 7-Eleven/Tescos, dog poo on the streets and noisy traffic. 39 forum chapters and he did eventually make it.

He bumped Bruno, the former recordholder in Portugal. All Bruno wanted to do was record his acoustic guitar. Really. That was the whole job. Got that to work, too. It helped in both of these cases that the poster/performer was really good at what they did, so it was up to us across our 9 time zones to make it work.

And yes, I’m perfectly clear it’s not up to us to manage live recording. It’s only tangential to the actual Audacity program, but the umbrella services are valuable in that they illustrate how people are actually using the software in real-world conditions.

That and it keeps us off the streets.


Wouldn’t it be lovely to have a pro booth in the study at home? Not ever going to happen in my home (for lots of reasons - money, other people, money…). LOL at myself.

Right. So that gives you the conditions for living different than the conditions for presenting. Many times that leads to software gymnastics required to patch noisy work and that can lead to “overprocessing” objections from ACX. It’s a well-trod cycle.

Ian converted a small closet into a recording space whose only significant shortcoming was lack of air conditioning in the Hollywood summer.

You don’t have to do it at home. One presenter took over her company’s open-plan office after everyone went home and they turned off the noisy lights and ventilation system. The area morphed into a remarkably well-behaved recording space, not for soundproofing, but sheer size. With a reasonbly-placed microphone, echoes just never get there.

I have done that. I wandered around a factory floor until I found a quiet nook and then shot a voice track for an animation there.

One reader famously took her portable recorder and duvet into a hotel closet to do a reading…and accidentally locked herself in.

All of those scenarios violate the easy metaphor of setting up a Yeti on the kitchen table and reading audiobooks.

The automated robot and Human Quality Control at ACX are there against the readers whose kitchen tables can’t be made to work.


As was pointed out to us, there are companies with no standards who will publish whatever you have. One reader was incensed that he was expected to meet quality standards to read for ACX.

We assume they don’t pay as well.


I’m a newbie here, so hi! I’ve had a lovely time over the last few days looking through past messages and have found most of the answers to my questions without having to ask them