My Audiobook Workflow

As you may recall from my last posting, I am working on getting my two books Hooray for Pain! and With It or in It into audio book format. I promised a look at my workflow, so here it is!

So, here are the steps I've been using. Since this is a book of poetry, each file is really very short and it makes more sense to me to do each one, one at a time.

1. Record the audio using the Røde VideoMic Me attached to my iPad or iPhone (using GarageBand with the Monitor setting enabled)
1. The Audio Genie II is plugged into the monitor jack in the back of the microphone.
2. Using Sound Studio's visual VU meters, adjust the gain on the Audio Genie to get mostly green levels.
3. Start recording in Sound Studio, switch over to the Kindle app on my mac (I read directly from the version of the book I uploaded to KDP), and sit quietly for a moment
4. Read the poem. I position the Kindle app and Sound Studio so that I can see the VU meters below, and it will alert me if I'm getting into the yellow levels.
5. Finish recording, sit quietly for a count of three, and then switch back to Sound Studio and stop recording.
2. Listen to the captured audio as-is. If I don't like it, re-record.
3. Click "Normalize" on the toolbar in Sound Studio, with a peak setting of -3 dB (this will give it a little headroom)
4. Listen to the Normalized file. If it sounds good, save the file.
1. I save my files at this point as AAC files with a 320 kbps bit rate. For a really long file I might choose a slower bit rate, but for these files (which are < 1 minute, mostly) it's not a big deal.
2. When I save it, I learned a trick from Izzy Hyman (of http://www.izzyvideo.com/ ) to make it easy to import into Final Cut. In Final Cut Pro X, you can drag a folder into the Music and Sound Browser and it will add it to the sources you can use directly in Final Cut. So I save the files from Sound Studio into a subfolder I created for this project, under the folder I dragged into Final Cut.
5. Make the edits in Final Cut Pro
1. Add the saved file to my working library in Final Cut, create a new Event for it, and add the sound file to the new event in the Timeline
2. After much experimentation, I saved an Effects Preset that combines my starting point for a Noise Gate, Noise Reduction, Doubler, and Volume. I saved the preset and made it my default audio effect through the Final Cut interface, and then I drag the effect onto my audio track. Here are what the settings are/mean:
1. Noise Gate will essentially drop all sound below a decibel threshold. I start with a default of -48 dB and adjust it by listening to the clip all the way through. You'll need to find a decibel setting that doesn't clip your softest speaking parts.
2. Final Cut has outstanding Audio Analysis, and I use their Noise Reduction on top of the Noise Gate. By default, I use a setting of 50%, but occasionally need to bump it up a little higher … especially in Tucson, during the Summer, when it's 110° F and the A/C kicks on every few minutes.
1. There is also an option to de-hum, to adjust for AC electrical current, so if you are using a mic that is connected to electrical (this one is not) you might consider fiddling with that setting as well. 
3. I leave the EQ flat, and I don't add any of the reverb settings from Final Cut, but instead use the Doubler effect with about a 1.05 setting. In essence, this fills out the voice sound by simply repeating it, 0.05 seconds later. It's kind of a cheater's reverb but without some of the obviousness of reverb. If that makes any sense. You could get the same effect by duplicating the audio track, turning off the snap-to-magnet setting on the Timeline, and play around with exactly where the two tracks line up, but I think Doubler is just simpler.
4. Then I listen to the clip in its entirety, watching the VU meter to make sure I don't have any peaks above -3 dB (which is what ACX, Audible's version of KDP, requires) and if I do, adjusting the clip volume down until all peaks are -3 dB or lower. I also listen for clipping of my voice, and if there is any I adjust the Noise Gate appropriately. If there is no clipping, but there is still any background noise, I adjust the Noise Reduction setting under Audio Enhancements.
5. I also listen for "mouth noise" like lip smacking, obvious breaths, and so on. If I find any that sneak through the Noise Gate (which happens), I use the Range Selection tool of Final Cut, zoom way in on the Timeline (down to a few frames, usually), and then put a range around the noise. Once selected, I can pull the volume on that range down to -∞ dB, which eliminates the noise without affecting the timing. If you're careful about how you breath when you are reading, you should be able to isolate these without affecting any actual speaking. It will help to read very deliberately. My prior experience as an actor and singer (when I was younger) helps!
6. Finally, I find the beginning of my actual narration, backtrack two frames less than 1 second, and use the blade tool to remove everything before that. I then repeat this at the end, but with a 2-ish second tail. ACX requires a 0.5–1-second head, and a 1–5-second tail on every file.
6. Once I have the edited sound the way I want it, I do a Share->Master File on the current Project. I set it to be Audio Only, and output as an AAC file. You can output it as an MP3 directly from here, but I take one additional step because Final Cut doesn't give you any options to set the bit rate mode or quality settings.
7. Open the AAC file in Audacity. Using Audacity to convert the AAC file to an MP3 gives more options for the saved MP3 file.
1. Once opened, I Export Audio… and use the following settings:
1. Format: MP3
2. Options: Bit Rate Mode: Constant, Quality: 192 kbps, Channel Mode: Stereo
3. Then I fill out Metadata tags as appropriate

And that's my workflow. I haven't finished yet, so I can't guarantee that this will get approved by ACX. If it does, I'll post a blog entry confirming that; if not, I'll post a blog with the changes I needed to make. Also, I expect I may need to make changes when I start recording With It or in It, since the chapters can be their own files (and will be considerably longer than 1 minute, I suspect!). For doing one poem at a time (which is what ACX wants), this workflow ensures that I get each file done to the standards as best I can right now, and once done I can move to the next poem.

A couple of points: I use Final Cut Pro because my entire original plan was (and remains) to do filming, film/video editing, and it's unquestionably an excellent tool for that. I learned how to use it quite a while ago, when I first started the process of learning video & film editing, and I am very familiar with how to use it. It's expensive ($299 on the Mac App Store). I can certainly dig if you don't want to shell out scratch like that to edit audio; I didn't buy it for the audio. It just so happens that it is as excellent at audio editing as it is at video editing, and since I already had it, and knew how to use it, I adapted it to my workflow. If you're using something else, comment below so everyone can benefit from your experience as well!