how to get it between -23db and -18db for ACX standard

Ok so I got hired to do an audiobook and I need to turn the first 15 soon. I finish recording and editing, but when I went to make sure they were up to ACX requirements they were not. The noise floor was ok about -63db, but when it came the over all measuring between -23db and 18db I was about 30. So I looked up how to fix it and bring it up I ended up bringing the noise floor to high or my earphones. Also I think it made it sound bad when I changed it and I really liked how it sounded before, but that might be because of the noise floor. I then spent most of yesterday trying to figure out what to do and coming to nothing. By the end I decided to try starting over (I did not erase anything) and started experimenting on what to do. After some playing with it and backing the mike up and speaking a little louder I got it to about -24db. I taped were the mike was and everything, but it was late and only did a page. Today when I started recording it is back to the beginning even though I have not changed anything. I have no idea what to do and I am getting really frustrated. How to I record at at least a -23bd!? Please Help

Actually, you have to meet all three. Background noise (hiss (fffff) usually), blue wave peaks at -3 and then overall loudness (RMS) between -18 and -23. Noise doesn’t have to be hiss. It can be street traffic as you read.

Are you on Windows? Do you have any of the Windows Enhanced Services running on your voice? It doesn’t have to actually damage anything like in the wiki entry, but that can mess up your performance.

Do you run Skype, games, or any other program that also uses sound? That can change sound settings and not tell you. Skype is famous for taking over your computer while it’s working…and sometimes not giving it back.

Record a 10 second mono WAV test clip in your normal announcing style with the red sound meters peaking about -6. The first two seconds should be holding your breath and not moving at all (noise test).


Quick note. Do Not apply any filters or effects to your posted clip. We need to hear it before you help it.

Ok some of what you said I did not get what you meant, but I will reply what I did. I know I have to meet all 3, but the overall loudness is my biggest problem. Yes, I have window, but I have no idea what Windows Enhanced Services and the link you left just confused me more. As for skype, in fact I did download it just the other day, but I was recording before that. That might be the reason it worked well that time, but the next day it did not though. I will take it off my computer. I don’t think I have anything else that will do that. As for the audio I will, but it is raining right now and my mic will pick it. It should not last very long though and just to make sure when you say normal announcing style do you mean how I normally would talk in the mic or in a loud announcing voice? :confused:

I don’t know if what I posted earlier did not go through or I’m not sure if these comments just take a while to submit, but anyway… I was going to wait for you to comment back just in case, but o well. Just in case it did not go through I downloaded skype the other day just took it off a few hours ago. In that time I played around with recording different things. Sometimes it would be good and then it would be bad. More bad then good. Then I just decided to go ahead in send in the audio you asked whether it was exactly what you wanted or not. The first try it was really bad I could not even understand what I was saying, it was going in and out. I thought that was really weird and decided to try again and it came out great. The best I had seen all day. Then I tried again and it was also good, so I don’t know what is going on if it will stay this good or go bad again. Was not sure if you wanted to hear the good audio? I also noticed my original audition for the book was also good. I did not know to check it when I turned it in.
I guess if it is good it is good. I don’t know if it will suddenly go bad again or if you still want to hear the audio or if you have any more suggestions? I don’t know what to do, but take advantage of it working well at the moment. Thanks for the help.

ok so it went bad again. I started recording and it seemed good, but in the middle of the recording it went down. I didn’t do anything I did not change positions nothing. I took skype off and restarted the computer. I don’t know if anything else could affect it I don’t have very much. I try to attach what I recorded so you can hear. It is really noticeable at the end.

You are still in Forum Moderation. A Forum Elf has to look over and accept all your posts until we find whether you are going to be well behaved or not. Then your work will post immediately. I’m reading your posts. The forum elves are sprayed over nine time zones and we may not get to you immediately


You may make it past ACX compliance (noise, peaks, etc), but you’re never going to make it past Quality Control with that odd voice tone and fading.

This clip is the poster child for sound processing that Windows has in Version 7 and 8 to help you with conferencing. Conferencing voices don’t have to be pretty, they just have to work.

ACX has to be pretty, too.

Which Windows do you have and which Audacity? What does the Windows screen say when you start the computer?

Audacity > Help > About. We’re expecting three numbers like 2.0.5 or 2.0.6.


I have window 8 and Audacity 2.0.6

I have window 8 and Audacity 2.0.6

It’s almost certain that Windows is interfering with your microphone. Look at these settings or get someone to help you. We can’t take out hollow voice distortion that Windows puts in.

I’m not a Windows forum elf, so I don’t have access to all these settings myself.

– Right-click over the Speaker icon by the system clock then choose Recording Devices to open the Recording tab of “Sound”

– Right-click over Microphone (or whatever device you are recording from) and choose Properties

– There will probably be an Enhancements tab where you can disable all or selected “Sound Effects” - if needs be, also look in the Levels or Custom tabs

–You may also need to change or disable environment settings. To check settings, click the Playback tab in the main window of “Sound”, right-click over Speakers, choose Properties then click the Enhancements tab.


Okay, I got my mic, I have my closet studio set up (although I have plans for improving it).

I processed this in Audacity with a small amount of noise reduction (about 5db) and light compression and normalization. I left breaths in because it sounds more natural to me, but I did reduce the amplification on the breathing (manually). I exported with my default settings to mp3 - which I think put it at 128 kbps.

Can I get some feedback on whether I seem to be hitting decent audiobook standards?


You seem to be, yes, after a little push.

Effect > Normalize: [X]Normalize to -3 [X]Remove DC

That’s a minor volume shift. That brings the blue wave peaks up to -3. After that, I measured the silent area just after “Part One.” That comes in about -62 or so. The overall RMS (volume) is measured at -18.3.

All pass. ACX is expecting real people to read these things and real people have to gasp and take in air. That bothers nobody. Objectionable noise is the TV running in the background, traffic noises on Maple Blvd or a constant rain-in-the-trees hiss (fffffff). Those are deadly.

You can’t do production on MP3s, so you have to go back to the original WAV files or Audacity Projects (or both) and patch them. WAV files are good for archive. You can make MP3 client deliverables from them easily and you can do WAV post production in the event your publisher wants the work in a different format or venue.

It’s a pleasant reading voice and it will be interesting to see if you can keep that up through a novel. And yes, hand-tuning each take is going to get very tired in a hurry. If they’re natural noises associated with you performing, I’d leave them there.

I understand that ACX allows you to post a 15 minute sample of work before you read all six volumes in a way they don’t like.


Another note, peaks may shift around. The MP3 process rejiggers the composition of your sound and it’s not shocking if the peaks move. You can open up your exported MP3 and Effect > Amplify (don’t apply it, just look at the numbers). That will tell you where the peaks ended up on the client deliverable. If we’re being obsessive, I believe they should not go over -3. That may be what happened to your test clip.



Yes, I believe I actually normalized down lower than -3 because I’d read something about the MP3 shifting it. (Also, I normalized on the whole file, and I think there were some louder parts later.)

I don’t record in Audacity: I have a Sony PCM-M10 digital recorder so I can get away from the computer. (Using a Shure SM58 with it.) I always keep the WAV files it creates. I only ditch them if I feel it’s something that I’d rather re-record than retrieve the old file.


Oh, and re the question of being able to keep a voice up through a novel: that’s why I’m doing a weekly podcast. I’ve been doing a novella for the past six weeks, and I find that I’ve been glad of a technical excuse to redo it now and then, just because I felt the performance aspect went flat. That usually seems to happen because I’m tired, or because I’m concentrating on something else.


It seems to be working fine now. Thank you so much :smiley:

I have a Sony PCM-M10 digital recorder so I can get away from the computer. (Using a Shure SM58 with it.) I always keep the WAV files it creates.

We’ve been known to recommend a technique like that. I wondered how you got past the Windows bubbling voice.


Just a quick note: acx want 192kbps on most files.


Just a quick note: acx want 192kbps on most files.

Yes, but you should be making those from your WAV masters. Too many people think it’s good to make the MP3 files as masters and then are horrified when the compression damage goes up on a second or third edit.

I think one or two of our posters got stuck with that problem. They submitted lower bitrate than is required and they also had to edit the work. I’m betting ACX rapidly ran into audible compression damage on those submissions.

ACX will also allow a 15 minute test submission.


The MP3 thing catches everybody. Never do production in MP3.

Production sound is made up of collections of individual tones. This is the sound analysis/Fourier thing. All the tones don’t just add up, however, some of them in the normal process of musical sound subtract. MP3 is in the business of deleting tones that “aren’t important” in order to make files smaller. If MP3 happens to delete one of the subtracting tones, the MP3 sound volume can actually go up slightly.

Check it before you deliver to a client if they care.