Sound Advice
Recording an audio book was… a lot more work than I thought it would be
When I finally published Androids: The Team That Built the Android Operating System after 4 years of interviewing, researching, writing, rewriting, and editing, I was able to relax and stop thinking about the book.
For about two weeks.
Then I thought, “I wonder if people would like an audiobook version? I bet I could do that!”
This thought led to another seven+ month project as I recorded, re-recorded, and edited the audiobook, which was published by Tantor Media last July. Along the way, I learned many things about recording and editing an audiobook, such as the fact that I am not, and never will be, an audio engineer.
Here is some of what I learned.
History
First, some context on my original project idea: I wanted to create a rich audio experience that would mix my narration with original, source recordings of the people I interviews.
I realized that since most of my research existed in recorded interviews, I had a unique opportunity to produce an interesting audio experience that would mix the text I had in the book with the original recordings of the actual people on the team whom I interviewed. How cool would that be?
I had listened to Malcolm Gladwell’s The Bomber Mafia, and had a proof point that it could be done.* In particular, I thought the way he did it would work for this book. The problem with the recordings, of course, was that they were not made with an audio book (or anything more than my private notes) in mind. They were filled with external noises (including chewing during lunch interviews) and the results varied from reasonable-not-great H4n microphone/recorder files to fairly awful recordings of videoconference calls.
* Well, it could be done if you’re Malcolm Gladwell, anyway.
Gladwell had to solve that same problem, where many of his source recordings were from pre-WWII, stored on bad and decaying media. But I realized that if I just sprinkled in small segments of these recordings (like Gladwell did), then they would enliven the overall narrative with a taste of the original conversations, without detracting from the book by spending too much time in recordings that weren’t up to the task.
At least, that was the original idea. But then reality set in.
For one thing, when I polled a small sample of my colleagues about this idea, most of them said “NO!” I hadn’t expected this response, but I should have seen it coming. Most people do not enjoy listening to recordings of themselves, so of course they’d hate the idea of other people hearing what they don’t like to begin with.
Also, by this time I was working with an audiobook publisher (Tantor Media). They have a policy of having a single narrator for their books. So unless I wanted to go looking for another publisher (a project for which I had neither time nor patience), I was back to the traditional approach: a single narrator.
But at least, I thought, I can do it. After years of recording presentations, videos, and podcasts, I figured it’d be pretty straightforward. Silly me.
Home Studio
First, I had to create a “studio,” or at least enough of a quiet space where I could spend hours recording the chapters. The picture above shows most of the equipment I used to create that space:
- Microphones: I initially started (and recorded most of the book) with an Audio-Technica AT2020 USB microphone. This is what I use when recording the Android Developers Backstage podcast from home, and it works well for that. But it does not work well enough for an audiobook, which needs a *lot* higher quality. After recording and editing many hours, I realized that I was dealing with far more audio artifacts than I should, so I asked an audio engineer family member for some advice. She pointed out that I needed dedicated recording equipment and an XLR microphone, rather than USB to my laptop. So I picked up the Electro-Voice RE20 XLR mic and Zoom recorder shown above, and started it all over again.
- The manual for my Zoom recorder (lower left, above) was useful until my dog got hungry. I should be thankful that he didn’t eat the entire thing, since he did exactly that with 2 different rubber coasters and a rubber watch band. Our dog is always very hungry (and never very bright).
- The water bottle was useful for minimizing (though not eliminating) the pops and clicks that naturally come from a dry mouth. I didn’t realize how much my mouth did that until I listened to many, many hours of myself in the recordings. I was already using a pop filter by this time. I’m sure it helped, but there were still plenty of pops and clicks left over. Oh, so much editing.
- I built a Faraday cage in a desperate attempt to eliminate what I assume was electronic interference. I picked up an EMF detector to get some read on what was happening (though it wasn’t terribly useful in figuring out where those signals were coming from or what to do about them). When the cage wasn’t fixing the problem, I grounded it with the homemade grounding wire shown in the picture. When all of this still didn’t fix it, I just carried on recording, opting to fix the artifacts by sheer editing force instead.
- There are two pairs of XLR cables above. I bought a higher-quality set in another attempt to eliminate electronic noise interference. The shielded cables seemed better… but still didn’t solve the problems I was hearing.
- Shock mounts are useful for reducing the noise that can come from bumping the mic, stand, table, etc, to avoid passing that noise along to the mic and the recording. But the mic was so sensitive that I found it better to just not bump those things to begin with.
- The Cloudlifter mic activator was necessary to boost the signal of my XLR mic.
Not shown: The USB mic, the boom arm mic stand, the tablet that I read the narration from, and the laptop+software that I used to edit the audio. Also, the towel and blanket used on the desktop and wall to dampen echoes, and the blanket draped over the window to lessen external noise.
Oh, so much equipment….
Random Thoughts and Notes
My final, edited book is about 10.5 hours long. The full audio I recorded was probably 12+ hours (re-recording spots when planes flew overhead or my heat or A/C kicked on or the dogs barked at the mailman). Add in the recordings I made before I switched mics and I was probably at 20 hours or more or reading my book into the microphone(s).
But recording wasn’t the hard, or time-consuming, part. It was pretty easy to record a narration, since the text was right there in front of me. No, the hard part was the editing.
Have you ever listened to an audiobook and heard the narrator breathe? Me either. This means that either the narrator is able to breathe silently (or far enough away from the mic to effectively be silent), or someone edited out every. single. breath. I am not a professional narrator (this being my first and only audiobook recording), so I had to opt for the latter. It made me want to stop breathing entirely, forever.
Add to that the multiple takes I recorded for some passages, and the endless pops, clicks, and artifacts that I had to edit out or minimize, and it took probably 10x the amount of recording time to edit the results. So I spent well over 100 hours of editing by the time I finished.
Random Tips
Software: I used Audacity. I’m sure there are better tools out there that the professionals use. I liked audacity for the price (free), although I would have paid for the right tool if it would have helped. But it was clear that anything I got would require a lot of time to become proficient and I was already way over time-budget on this project, so I went with what I knew. Audacity was (I believe) sufficient, or at least it did the job for what I knew I had to do.
Room tone: Every time I record a video at work, the professionals have me stand still and give them several seconds of “room tone” (remaining silent while they record the silent room). I finally understand why. I needed this in each chapter recording so that I had a “silence” sound that I could paste over noises I didn’t want (such as the breathing and other artifacts mentioned above). I originally thought I’d just delete the sound that was there, but that doesn’t work at all, because the audio goes from narration over that quiet room noise down to complete silence. It’s jarring. Thus, room tone.
Artifacts: Honestly, I used so many techniques for minimizing audio artifacts that I lost track of them. Sometimes you can get away with pasting room tone, but often the artifact will happen over the audio you want, so you have to find a way to reduce those frequencies, or use various effects in the software to find and fix these spots. Sorry, you’re on your own. This was by far the hardest part of editing and it took forever. It made me want to be an audio engineer so I’d know how to do it more efficiently. But it made me want even more to not be an audio engineer so I’d never have to do it again, ever.
Footnotes: One of the unique, or at least noticeable, things about my book is the veritable plethora of footnotes.* A small number of them are citations (sources of sales charts, for example), but for the most part they are part of the story, asides from me with personal anecdotes that I believe help explain (or humorize) the overall narrative. But how to handle them in a serial audiobook format? When you read a book in hardcopy or ebook format, you know when you are choosing to read a footnote, so it is clear what the context is and you do not get confused. But if I want to read a footnote to the listener in the middle of a paragraph of related text, how do I do that without completely confusing the listener?
The approach I took was twofold. First, I started every footnote with the word “footnote.” Second, I applied a filter (sort of an AM-radio effect) to every footnote, so that listeners know immediately that these sections are different from the main text.
* like this.
The Result
Despite the huge amount of time that the project took (especially, as I said, the editing part), I’m pleased with the result, and I hope that the book’s listeners are as well.
In case you are interested to check out the result, the audiobook is available from the publisher, as well as the usual online stores, including Amazon, audiobooks.com, Apple, and more.
The book is also available in paperback and eBook format, from your local bookstore or favorite online destination.
