I am nearing the end of editing my first audiobook (finally! and getting ready to submit it to ACX.
On ACX, I need to submit each chapter separately. As I recorded the whole book in one Audacity file, I planned to make a new file for each individual chapter to submit it to ACX.
My question is about ACX check. My understanding is that this measures the average levels of your recording to make sure they meets the submission requirements. So, should I use ACX check on my whole book file, or on each individual file once I split up the chapters? (since the average could potentially be different depending on the section of book i check) How does ACX check that your chapters meet the requirements?
I hope that makes sense! I’d appreciate any advice or tips about this process. Thank you!
ACX Check measures three different things: Peaks, RMS (volume) and Noise. It was written by Flynwill by collecting several already existing tools. ACX Check is the collected works with a pretty face.
It’s designed to be very similar to the “robot check” that ACX uses. As a fuzzy rule, nobody is going to listen to your work for theater, expression, tone and metre until you get the basics down.
ACX Check is an add-on but it appears under Analyze, not Effects.
I would check each chapter. If you try to check the whole book, you may run into troubles with maximum allowable show size and length. I’ll play Sad Suzy for a minute and say if you fail a full book test, you’re going to have to dig through the chapters anyway to find the problems.
Fair warning both ACX and ACX Check need some hold your breath and don’t move “Room Tone” somewhere in the performance. ACX Check can’t measure background noise if you don’t stop talking long enough to find any. We need at least a half-second. I think they need more. You should check their instructions.
ACX Technical limits:
— Peaks no louder than -3dB.
— RMS (performance loudness) between -18dB and -23dB.
— Noise no louder than -60dB.
Those are the top three values in the ACX Check panel
Let us know how it comes out.
I planned to make a new file for each individual chapter to submit it to ACX.
And while you’re there, also Export a WAV (Microsoft) 16-bit sound file and save that as your chapter perfect quality protection backup. Copy them to a separate thumb drive, disk or cloud.
If for some reason your Master Project fails to open, you’ll need those to put the book back together.
ACX requires submission in MP3. That was a corporate decision. MP3 quality isn’t as good as WAV. Did you install the Lame software? Audacity will not create an MP3 by itself.
When you say check each chapter, do you mean highlight each chapter within the whole recording and ACX Check it, or save each chapter in a separate document and check them then? Does it make a difference? I will also go ahead and ACX Check the whole book, like you suggested.
Thanks for the advice, just wanted to make sure I understand what I should do!
I would Export a WAV copy of each chapter now and then copy them to a Safe Place.
The military would call what you have now a “Single Point Of Failure.” Just the right mistake or a computer not feeling well this morning and your book may not open and may not be recoverable. How long did it take you to read it?
Good computer hygiene says you should be able to point to two different places that contain your valuable work.
Then after you do that, choose one chapter and check it for compliance.
Or, as a quick alternative, you can record a simple, short sound test and check that. Do it in your normal reading environment and conditions and see how that comes out. If it fails, post it on the forum and we’ll see what’s there. Doesn’t matter what the words are, read the milk carton, but do it in your normal style.
This is an example. I went into my quiet third bedroom just now with a simple sound recorder and cut this. I’m no performer. I subjected it to the AudioBook Mastering tools and then tested it with ACX Check. It passed.
I did cheat a little bit. I cut off knocking over the coffee cup and I did it between metrobuses going by.
The FFFF noise just barely passes. If I was going for client production, I would suppress the background noise more and wonder why anybody in their right mind would hire me to do a voice job.
So yes it is possible for a home performer to do this with minimum equipment. I used an Olympus personal recorder at $120 USD. I processed it on a Mac Mini with normal 2.2.2 Audacity.
Since ACX requires you to submit each chapter individually, you should check each chapter file individually. You need to do this after your editing and any post production.
I use a Razer chroma keyboard (cheaper than Pro Tools keycaps )and have the 1, 2, and 3 keys mapped for my standard post:
Compression,
Limiter,
ACX Check.
When I have my final edit, I simply tap 1 and accept my preferred default, 2 and accept, then 3 to make sure I’m good. When I sit with my artist for a final review, if there are any snips or tweaks in the submission version, it’s just tap 3 again to make sure I didn’t break anything.
By the way, I would strongly suggest that you record each chapter as its own Audacity project. Audacity is a great tool - probably my favorite of the several I have available for long format spoken word. However, it’s not perfect. It is only a matter of time before you have a project get corrupted. If you find yourself in a position of having to do any sort of re-work, you really only want to have to deal with that single chapter project. All the folk suggesting safety backups are telling you that for a reason.
It is only a matter of time before you have a project get corrupted.
What he said.
Compression,
Limiter,
ACX Check.
And you can get away with that if you have a pro microphone system or microphone that doesn’t introduce rumble distortion and you’re recording in a dead-quiet, echo-free room…and you can read aloud.
Provision for some of those corrections is built into Mastering 4.
Uhh… well. Yes. There is that. Yamaha analog mixer has the EQ. I’m using an ART tube preamp and a Peavey CEL 2 in the insert tuned to the talent for primary compressor/expander. It’s one of the most neutral compressors I’ve found yet. The Mbox is pretty much used only for the digital I/O. Everything is proper gain structured.
For spoken word, I’m using either and AKG c414 or a Superlux r102 ribbon mic. Though, it’s not a “booth,” the room can be considered a “treated space” - I have control of how much reflections and/or ambient noise gets into the system. It just looks like an office/library.
It’s “old school” but it gets me pretty good noise floors and minimal post-processing. The talent likes it because yes, they do know how to read, and none of the inflection, dynamics, or nuance is lost or squashed.
That said, if your post production process takes more steps, it’s easy to map the key steps in Audacity. I like the “one step at a time,” 1-2-…-n approach because it let’s me tweak in the odd case that something needs to change for an isolated session.
And if it wasn’t clear, do your post on the individual chapters as well as the ACX Check.
I don’t know if you’ve experienced this, but the home USB microphones, in addition to having “restrained volume,” don’t filter the DC/battery system. The most obvious problem this can cause is the “Yeti Curse” where USB data leaks into the sound.
But that’s not the only problem. These microphones can also produce noise below 20Hz and some of it can be loud enough to keep you from passing ACX noise test “for some reason.” The noise is below audibility, so you can have a performance with dreadful rumble/thunder/earthquake noise nobody can hear.
it’s not a “booth,” the room can be considered a “treated space”
That picture is the Main Conference Room which is fully sound treated including padded air conditioning ducts.
Most posters are trying to do this in the apartment with white walls and polished wood floors. It’s difficult to get a good, clean reading with conditions like that.
ACX knows. They have (as far as I know still there) a video where they describe the walk-in sound booth you’re going to build in your apartment. http://www.acx.com/help/setup/202008260
It’s “old school” but it gets me pretty good noise floors and minimal post-processing.
“New School” is a little rough to watch. There is a British VLOG where the performers shoot outside and a good fraction of the time the voices are inaudible from wind noise. They went to several months of shooting that way. Somebody must have noticed.
But if you want to read for Audiobooks, you still have to make past the ACX Robot, simulated by ACX Check. And then pass Human Quality Control to make sure you don’t sound like a bad cellphone—or a bathroom.
Oddly, I don’t remember anyone ever getting bounced for making “human” noises like breathing, but that’s the first thing many new performers go for.
As previously mentioned, break your chapters apart and check them. You have to break them apart anyway. Besides, unless the entire book is under 30 minutes or so, you can’t check it all at once. Some of my chapters run near 35 minutes and when they’re that long, you cannot check the entire section. (I don’t know what the exact ACX check length limit is). But anyway, if you do run into the limit on any of your chapters, just pick sections of it and do the test. This is what I’ve done in the past and haven’t run into any problems. The ACX check is a nice little tool to have.
It saves a lot of shuffling test files back and forth to ACX. I have a posting here somewhere about how to perform ACX Check manually with the older, existing tools. It is possible but it’s really painful.
I see there’s an unimportant error in there I never went back to correct. All the tools work in Stereo now.
You should know that ACX Check needs a good half-second of silent Room Tone somewhere in the work to get the noise value. If you don’t have that, it will pick the lowest volume background noise it can find and you could easily appear to fail noise in error.
Also please note that everything should be done in Mono. One blue wave. ACX recommends submitting that way and everything just works better in mono. The system runs faster, the tools can check twice as much work before running out of steam and storage goes twice as far.
ACX test limit is about 37 minutes at 44100 sampling rate. Apparently it doesn’t matter stereo or mono because it rips stereo apart and checks each side independently.
In Audacity there is an option to export a file as an MP3. Is it easier to save chapters as MP3 rather than as WAV files so that I don’t have to go through the ACX file conversion process to convert the WAV files to the MP3 format? Or will ACX not accept files that were exported directly as MP3 files?
Yes you can export directly to MP3, but it is highly recommended to also export a backup copy in WAV format.
If you find that you need to make additional edits or other changes after exporting, then you can import the WAV into a new Audacity project, make the changes and re-export. This will be better quality that using the MP3 version, because MP3 encoding is like photocopying in that there’s a small but irreversible loss of quality each time it is encoded.
I would also recommend making a backup of the raw recording, immediately after pressing the Stop button. In the event of a catastrophic failure during editing, you will at least not have to redo the recording from scratch.
Audacity really likes saving Projects. Projects are nice because they will save your preferences, edits, timelines, selections, etc, etc. Projects will not save UNDO. Projects are saved in many individual files (sometimes a great many) and if anything happens to any of them, the show may be trash. Without the WAV backup, you may be reading that chapter again.
It is possible to submit a short test to ACX. This will prove your production pipeline without cold-submitting bushels of sound files and weeks of work.
We should note that ACX has two test levels. There is the automated robot which, in general, tests for the same things that ACX Check does, but then they have Human Quality Control. That’s a human who tests for theater. Are you pleasant to listen to, do you stutter or do you like to spit on the microphone? Did you overprocess your poor work to force it to pass? That’s a common failure.
If you do fail, they will generally explain why and maybe one or two other comments. They won’t troubleshoot your studio and they won’t engage with you.
and getting ready to submit it to ACX.
