How to write a book

Diary of an audiobook: lessons learned in the making of Ever Rest

Last month I released the audiobook version of Ever Rest, my third novel. All creative collaborations bring surprises. There are things we’re glad we did and things we’re glad we didn’t do. Here are the lessons learned.

Meet my narrator

My narrator was Sandy Spangler, who narrated my other two novels. Sandy is a longtime friend who designs computer games, but I didn’t know she had also trained as a voice actor. In 2014 I had a generous sponsorship offer from Amazon’s ACX to make the audiobooks of my first two novels, but couldn’t find a suitable narrator – you can read about the struggles here. Then I discovered Sandy could do it. She was perfect.

We recorded My Memories of a Future Life and Lifeform Three in 2014. You can read about our adventures making those audiobooks here.

We talked about doing Ever Rest if I could get another sponsor. And, late last year, I did. (You know who you are and… thank you.)

So if you want to know how an audiobook is made, here’s the complete guide.

Sales platform

How do you get the audiobook to customers? You need a sales platform or distributor. I chose Findaway Voices. It reaches a huge range of retailers and libraries and has a good rating from the Alliance of Independent Authors watchdog. When my ACX contract expired for My Memories of a Future Life and Lifeform Three, I moved the books to Findaway.

If you don’t already have a narrator, Findaway has a process for putting narrators and authors together.  

Working method

There are two possible ways to make a book with Findaway. You can upload a finished set of files, or you can use Findaway’s author/narrator portal to share and approve files. There’s a small fee for this. There’s no fee if you upload a finished book.

Sandy and I decided that, as we’d already worked together well and were used to working in creative teams anyway, we could figure out our own sharing and approval system. I have a Dropbox subscription so we used that for sharing files.

Here’s one we made earlier.

Set up your system and stick to it obsessively

You need to keep track of where each file is in the recording and approvals process. Is it waiting to be reviewed? Does it need corrections?  Is it passed for final production?

I created a set of folders in Dropbox.

  • For Roz to approve
  • Corrections for Sandy
  • Pickups for Roz to approve
  • Passed by Roz
  • Final mastered files ready for upload

This is very bossy, but Ever Rest has 86 chapters. We needed to stay completely in control of where everything was. Also, I thrive on control and detail.

So many weird names

All audiobooks need pronunciation guides. Ever Rest has a really chewy vocabulary. There are Nepalese place names. Argentinian place names. Mountaineering equipment (karabiner: where does the stress go? KaraBEEner? What vowel sound? KAHRaBINer?)

Powder snow avalanche in the Himalayas – pic by Chagai at English Wikipedia

None of this mattered when the words were shapes in a reader’s mind. Now, there were pitfalls galore. Sniggers or scorn if we got it wrong.

I wrote a pronunciation guide, using phonetic spellings and links to YouTube videos.  

Closer to home, there was another troublesome name: the Long Mynd, a hill in Shropshire. The local pronunciation is ‘minned’, not ‘mind’. On the page, it looks like it could be ‘mind’ and I wanted the reader to feel both versions. At times, this was an important resonance, so how would we handle it in the narration? I added some lines to the text to make the reader aware. Then Sandy, in her narration, judged perfectly when to say ‘mind’ and when to say ‘Mynd’. It’s a joy when your actor is so in tune with the book.

Style of narration

Our next task was to establish the style of narration. How fast should it be read? How emotional should the reading be?

Sandy sent samples at different speeds and emotional registers. I’m glad we did this, instead of plunging straight in, because the book didn’t work as we expected.

At first we kept it low key because the prose has a lot of emotion. That approach is good for some audiobooks and lets the writing do the work, but Ever Rest sounded flat. So Sandy went the other way and it came alive – a rich, expressive reading that we could dial up and down for the different characters and tones.

Once we had the energy and emotion, we thought about the space between the paragraphs. When I was writing, I used paragraph spacing very precisely as a poetic device. I realised as I listened that we sometimes needed a longer gap between paragraphs to let a moment settle. Sandy gave me several versions and we found the perfect paragraph pause.

It’s essential to tackle these questions at the start. Even if you’re not used to directing actors or listening critically to audio, you’ll have a feeling when something is off. You might not know what it is, but discuss it with your narrator. If you make a suggestion that’s dumb, the narrator might have a good solution to the problem. Listen to your gut and discuss any concerns as early as possible, before you’ve got a heap of chapters you’re not happy with.

We had found the novel’s style and voice.

Leave room for the reader: 1 Regional accents

One of the characters, Elza, has an Australian accent. We wondered: should Sandy do that?

My instinct was no. And anyway, if we’re getting picky, all the characters have accents. Elza’s is the most striking, as she’s Australian, but another has an East London twang. Another has Shropshire, probably with layers of Nepal, where he’s lived. Another is French.

But accents can be intrusive. And in an audiobook, the narrator doesn’t need to ‘do’ the characters’ voices to make us feel their reality. Their presence is so much more than accent. Their natures are expressed through word choice, thoughts, feelings and reactions, especially when they’re pushed out of their comfort zones. What makes them real is their inner life.

We wondered, though, if Elza might need different treatment because characters occasionally mention her Australian intonation. Her boyfriend, Elliot, talks about the ‘amused lilt of her accent’. What should we do about that? We certainly don’t want Crocodile Dundee. Perhaps an upward inflection, just on the preceeding line, which is a very Australian characteristic?

I’ll leave it to you, I said to Sandy.

When I listened to that chapter, Sandy read those lines exactly the same as the rest of the text. They worked perfectly.

Leave room for the reader: 2 Song lyrics

Ever Rest is a novel about a rock band, Ashbirds. Where there are bands, there are songs. So how should we read the lyrics?

Did they need a melody?

I did have ideas of some of the melodies. They came as I wrote the lines. But I never intended to share them with the reader.

As usual, Sandy made a sound judgement. She read them with a light sense of rhythm, as though they were lines of poetry, to let the listener imagine as much or as little melody as they liked. 

Once you’ve made a decision that feels right, you realise what the wrong decision would have done. To add melodies might make the songs feel less serious. It would break the spell of the reader’s personal version, what they were imagining.

I had an example of this at a recent book event. Readers asked me who was the real-life original of the band, Ashbirds. Absolutely nobody agreed with my version. Some thought Ashbirds were heavy rock. That wouldn’t work for me because I mostly dislike guitars – cries of horror from some readers. My Ashbirds would be a bit Peter Gabriel, with his intricate samples. A bit Massive Attack with their soulful synths and beats. A bit Pink Floyd with their introspective darkness. A bit late Police, seething with anger. A bit Simon and Garfunkel, with their voices that sounded like one. 

Howls of disagreement from everyone, who all had their own Ashbirds.

That’s the value of not saying too much, not adding too much.

Big lesson: in voicing the book, you must leave room for the reader.

Oh dear I forgot to mention… words I invented

As we worked, we realised I left some important info out of my pronunciation guide.

Some characters had unusual names that could be pronounced in several ways. Paul Wavell… was he WAYvell or WaVELL? said Sandy. It wasn’t important, but it held Sandy up while she wondered which I wanted.

Also, band names. Ever Rest is a complete ecosystem of rock bands. I had such fun inventing names that looked great on the page. Sandy had to figure out how to say them.

‘Roz, about Vidalvine. Veedal Vine or Vyedal Vine?’

‘The rap star Hobemian. Hoe-BEM-ee-an? Hoe-beh-MY-en? Hee-beh-MEE-en? Rhymed with ‘bohemian’?

Big lesson: add your made-up names to the pronunciation guide. I’m looking at you, fantasy and sci-fi authors, and absolutely everyone else too. We all invent names. Your narrator doesn’t need to play guessing games. They already have enough to do (see below).

Oh dear I also forgot to mention…

Some English place-names.

Sandy is American, so names that were familiar to me were not familiar to her. The London region of Holborn. The Yorkshire town of Scarborough, which is not Scarburrow.

Even if your narrator speaks the same variety of English as you, you might want to add pronunciation notes for any name that has right and wrong options, or options that indicate the character is local. For instance, Shrewsbury – locals call it ‘Shroo’ and everyone else calls it ‘Shreau’.   

Clicks and repeats

The chapters read beautifully, and occasionally I glimpsed the hard work under the polish. In one chapter, Sandy accidentally left an outtake where she was surprised by a line of adjectives. I could hear her faltering as her brain said, blimey, when does this sentence end? Then I heard a couple of hard clicks, her code for a retake. A breath, she repeated the line, back in the flow, perfectly inflected.

Big lesson: when listening back, stay alert for moments that have been accidentally left in. Every hour of finished recording takes several hours of preparation, recording and editing. Mistakes can happen.

I discovered only one such instance in the entire book, which is a tribute to Sandy’s careful work, but you don’t realise what can go wrong until it does. Watch for outtakes!

When you write a big book, it’s big work

This audiobook took a long time. Eighty-six chapters, about 110,000 words. It was a huge undertaking. Sandy remarked that whereas she’d usually take four to six hours per finished hour of work, she was taking six to eight. 

We discussed why this was. The amount of dialogue? The shifting tonality of the different scenes? The changing narrators? All of these gave Sandy a huge range of emotions to express, sometimes all in one chapter. The book took me, as the writer, seven years and 23 drafts (here’s a post about the seven steps of a long-haul novel). Sandy had to digest that complexity in just a few passes.

Sandy tells me she often recorded several versions of a line, then decided later which to use.

‘I did this very often with dialog. As a reader you can get to the end of a character’s sentence before you realise that you have read it with the wrong tone, or sometimes even the wrong character voice. Even if you aren’t doing strong voice differences such as accents, a well-written character has their own communication style and when reading aloud you want to keep that consistent. Or sometimes I would get to the end of a conversation and decide the energy or pacing of the exchange felt wrong, so I would do it again. The majority of the time the later version was the better one, like a tiny rehearsal followed by a performance. This book had a lot of intense dialog, which made it especially challenging.

‘I also recorded multiple versions of the more dramatic story passages whenever it felt necessary in order to get the emotion across – or at the end of a chapter. For those I always recorded multiple takes to sum up the energy of the moment. I feel words followed by silence need to hang in the air just right.’ 

In case you’re wondering what takes all the time.

The preparation showed. I remember when Sandy presented the final chapters. ‘That argument really took it out of me,’ she said. ‘I kept wanting to scream at her.’ This was exactly what I needed the reader to feel. And the pent-up pressure comes through in the performance. 

Endmatter– to record it or not?

Ever Rest has a discography after the final chapter. We wondered: should we record it? Several readers told me they’d enjoyed it as an unexpected bonus. But it was devised for the page, not for the ear.

Sandy recorded it. It didn’t work. We didn’t include it.

Ever Rest also has the usual endmatter. Acknowledgements, a brief piece about the author, teasers for my other books. We decided not to record them. On the printed page they’re nice as a leave-taking. But read out, in the voice of the story, which is so intimate, they would be jarring.

That’s just my opinion. You may think differently about your book.

Titles and other bits

Audiobooks need various official bits of start and end matter. The title, a copyright notice etc. To make this simple, Findaway Voices suggests a format that’s accepted by all the retailers. This is the recommendation for endmatter:

This has been [Title], 

Written by [Author Name],

Narrated by [Narrator Name], 

Copyright [Year of Manuscript and Name of Rights Holder],
Production Copyright [Year of Audiobook Production] by [Rights Holder]


The audiobook cover has to be square, so I pulled up my design file for the paperback and prepared to tweak. I saw this, the record sleeve visible in its entirety, before I cropped it to book proportions. (See the blue outline.)

Gasp. When I was designing the cover, I dearly wanted to use all the record sleeve. If only the book could be square. Now it could be. Did I dare? It would be quite a departure from the actual cover. Did that matter? And the titling would be small. Did that matter? It might not. The sales platforms would show the title and author anyway.

I put it to my Facebook friends.

Yes, yes, yes, yes, yes, yes, they said. It’s unusual. It’s got impact. Do it.

So I did. I also added the rosette for the Eric Hoffer Grand Prize.

Choosing a sample for the sales page

The sales platforms give readers a sample, as they do with ebooks. For ebooks it’s automatic – the beginning of the book. For audiobooks, you get to choose an excerpt.

With choice comes… dithering.

We deliberated about the sample for all the time we were making the book.

There are seven POV characters. Who gets to be on the sample? Will listeners think the book is only about that person? There were many chapters that seemed to transmit the book’s themes and dilemmas, or so I thought, as the author. What would someone think if this was their first taste of the book? I found it impossible to tell, I was too close.

There were some chapters that worked well as standalones but were far too long – the limit was 5 minutes.

Eventually I chose one, which was an important interior struggle for one of the main characters.

Nooooooooo, said Husband Dave. It won’t be memorable enough because it’s so interior. Can you find something where two characters meet or do something weird?

Eventually we got it. You can listen to it here.

Check everything madly, a lot of times

When all the content was correct, Sandy did her post-production work, then messaged when there was a complete set of files in the Dropbox folder, ready for me to upload.

Then came an orgy of checking. I double-checked I had all the chapter files, and the beginning and end credits. As there were 86 chapter files I checked them twice. If one was missing, now was the time to know, not when a hugely stressful error message came in from Findaway. Or from a reader.

I uploaded the files. Immediately we had a formatting glitch. However, Findaway’s help pages are so clear that I was able to discover the reason, told Sandy, she sorted it out and sent the files back to me.

Comforting note – Findaway’s help pages are really good. Their help team is also very responsive.

I checked again I had all the files. Any time you’ve done a mass upload, download or anything, something could go missing. You can never check too many times. It’s always worth it.

When I uploaded the novel, I saw its length for the first time. A whole 11 hours, 45 minutes. Again, no wonder it took a while.


When you upload the files, Findaway Voices suggests a retail price and a library price according to the book’s length. I saw no reason to argue! You can set discount pricing for launch offers and generate codes for review copies.  

We’re live

A few nailbiting days while Findaway did some technical checks of their own. Then hurrah – we passed. Findaway has a page with all the retail links. There are loads of them, as there are with ebooks. I have a universal link for each of my titles via Books2Read, and hurrah it as a section for audiobooks.

I also updated pages on my website and my newsletter welcome sequence, to let people know the audiobook is available. Now I keep coming across other random mentions on my website and blog that need updating.

Note for next time – make a comprehensive list of links or pages that need updating whenever I have a new release. Try to streamline where possible. We live and learn.   

Those audiobook lessons

  • Set up a workflow process and stick to it obsessively.
  • Create a pronunciation guide – any names and words you’ve made up, technical terms, foreign words and names.
  • Do a trial run to establish pace, level of emotion and pauses between paragraphs.
  • Look out for instances where important resonance might not translate. Perhaps adjust the text – remember mind and Mynd.
  • Discuss a policy about accents, unusual text like song lyrics but remember less is more.
  • Leave room for the reader!
  • Get your cover adjusted – and perhaps explore exciting possibilities that weren’t possible in the paperback or ebook edition.
  • Watch out for repeated sections and other kinds of outtake.
  • Check and recheck everything madly.
  • Make it easy for people to find your new audiobook – update your buy links and any pages or newsletter materials that mention your books.

And here they are, the magnificent three. My Memories of a Future Life. Lifeform Three. Ever Rest.

There’s a lot more about writing in my Nail Your Novel books – find them here. If you’re curious about my own work, find novels here and my travel memoir here. And if you’re curious about what’s going on at my own writing desk, here’s my latest newsletter. You can subscribe to future updates here.

How to write a book

Voices and accents for your audiobook – how to choose the right narrator

5936204848_a498d5ee50_bRight now, my voice actor Sandy is in her recording booth, speaking like a bod. Lifeform Three, my second novel, is currently in production as an audiobook, so this week I went to an event in London hosted by ACX and the Alliance of Independent Authors. Audible president Jason Ojalvo and author-entrepreneur Joanna Penn were speaking, and they had some interesting points on the business of choosing and working with a voice actor. (For an introduction to how to work with ACX, including auditioning actors, look at this post.)

Male or female
So a female author or a male main character must need a certain gender of voice actor, right? (And if you’re crossing the gender divide, how do you choose?)

Actually, it’s less of a cut-and-dried rule than you’d think. Jason said he’d often had authors who’d specified they wanted a female voice, then when a male actor had auditioned it had been the perfect match – even in genres like romance, whose readership are very definite in their expectations.

Jason made the point that the book – or the author’s work in audio – might have a voice that’s independent of the voice of the author or the character; it is its own identity. We’ll come back to this.

When I originally looked for a voice actor, I specified a British accent, but as many of you probably know, the narrator I chose is from the US. Initially I got a lot of US actors auditioning because I was one of the guinea-pig authors when ACX launched in the UK – they hadn’t yet got a bank of UK actors to choose from. So I heard a smorgasbord of attempts to ‘do British’, some convincing and some not. But I soon realised it didn’t matter after a few minutes anyway. The accent was irrelevant. The interpretation of the book went deeper than a voice’s characteristic twang, or lack of it. What was actually important was the voice actor’s understanding of the work.

And Sandy, regardless of the flavour of her English, was the most in tune with the novel. She also liked a lot of books that I liked. I picked her.

ojalvoSame voice for all your books?
Jason said if you have a series, listeners expect the same voice throughout or it breaks the story world. Authors of standalone books, obviously, might search for new narrators each time. I’m happy with Sandy for both my novels even though they are different in tone – because she works well with my style and outlook.

Joanna has two series, so she cast a narrator for each. Funnily enough, we might have ended up with the same one, as the narrator for her dark crime series was one of the auditioners for My Memories of a Future Life! Small world.

Hunting for narrators
You’re not limited to only the voice actors who approach you – and indeed, many authors don’t find an ideal match that way. Jason encouraged authors to hunt around the ACX narrator profiles, listen to their samples and invite them to audition for yours. Or some authors do what I did – if you know a voice actor who’d be perfect, introduce them to the system.

Working with unfamiliar accents
Joanna, like me, is British, and ended up working with an American narrator. Once into the recording process she found there were pronunciations that were alien to her Brit-tuned ears but natural to the US narrator. What to do about them? Tomayto or tomahhto?

Before recordings start, you need to discuss this, and also tricky pronunciations such as character or place names. Sandy and I talked about it. I suspected there would be many more variations than I’d have be able to think of. If I’d decided ‘leisure’ couldn’t be ‘leesure’, I’d have then, for the sake of consistency, had to pull her up on words I never dreamed had a US difference until I heard them.

And the difference goes further than isolated words – sentence emphasis is also radically different. US English stresses the adjective in a phrase like ‘lying on a sticky mat’. UK English stresses the noun (UK: ‘on a sticky mat’, US: ‘on a sticky mat’).

Joanna Penn, author entrepreneur

I didn’t want to stilt my narrator with unnecessary strictures so I asked her to pronounce her usual way. I’m glad, because there were hundreds of differences. Hundreds. It would have been madness. In any case, that didn’t matter. So long as the interpretation of the line was true, the emotion understood, the accent was irrelevant.

Joanna had also come to this conclusion, saying there’s a lot we need to leave to the narrator’s judgement and style. She intervened in place name pronunciations, but allowed everything else to go with the actor’s natural style and emphasis.

Having said that, an audiobook is a creative relationship. The voice actor is expecting you to guide them on  interpretation. Sandy and I spent several emails discussing how the bod characters in Lifeform Three should sound and what their individual characteristics were. I sent her short recordings of how they seemed in my own head as I wrote them, which she turned into polished performances. It was quite a feat for her – sometimes she had four or five characters in one scene and had to inhabit all those minds, as well as switching to thoughtful narration. For me it was easy because I wrote them. For her, it was mind-and-tongue gymnastics.

You can probably see why questions of ‘leesure’ versus ‘lezzure’ cease to be important. Forget them.

Don’t expect a drama performance
Jason pointed out that the audiobook isn’t a stage or film performance. It’s a reading – a quieter, more subtle business. Characters’ accents don’t need to be full-on impersonations, they are a hint. Passages of emotion don’t have to be performed, merely rendered so they bring to life what is already in the prose.

In prose, the writer has already done the job on the page. The voice actor is converting that into sound. It’s intimate; it’s not slaughtering the back row. It’s murmuring in your ears.

The voice that is the best conduit for your work
Ultimately, the best narrator is the right person to inhabit the book and bring it alive, from its lightest moments to its darkest corners. If you’re weighing up possible narrators, be prepared to revise what you imagined. If you thought the narrator should be British or male, but the more true interpretation, the one that gives you goosebumps, is US and female, that actor is the one to choose. The differences will vanish as soon as the listener gets into the story. After a minute or two, they won’t notice.

Since I released My Memories of a Future Life, some people have asked me why I chose an American, and indeed have mentioned it in reviews. Then they report that they got immersed. Your best narrator is the person who can inhabit the book, who can become its voice in the reader’s head and make them forget everything else.

You can find the finished audiobook here (US) and here (UK). And you can find out more about My Memories of a Future Life here.

mmaudioheads bigger

Thanks for the pic Michael Mol
Any tips or questions to add? Have you made an audiobook? If you listen to a lot of audiobooks, do you have any feedback on what makes a good narrator? Let’s discuss!

How to write a book

How to make an audiobook with ACX – more tips for narrators, producers and authors

A week or so ago I talked about making audio books with ACX, the self-publishing arm of Audible. From the author’s end it’s relatively simple – pitch your book, listen to auditions, guide the narrator and review chapters as they’re posted on ACX. But at the other end of the line, the narrator/producer is spending 4-6 hours on each finished hour you hear. What are they doing? When you listen to the files, what problems should you be alert for? And if you’re narrating and producing your own book, what do you need to know?

sandy n me

audiocloseupI put these questions to my narrator, Sandy Spangler.

Sandy: First I print out the pages of the text. Then I review them to refresh my memory and make notes about pronunciation or content/emphasis. Then I prep the ‘studio’ – which is my closet. I set up the laptop and hook up the mic and headphones.

Each chapter (or chunks if they are long) gets recorded in one go because the laptop needs to be outside of the closet, away from the mic or we can hear the fan. I hit record, then shut myself in with mic, headphones, and a glass of water.

Roz: You need studio-grade equipment to meet the quality standards for an audiobook. Podcasting gear won’t cut it. Equipment notes follow at the end of the piece. Back to Sandy.

Sandy: I monitor the audio as I go via headphones. Any time there is a mistake I do a retake and keep going so I don’t interrupt the flow.

Roz: Watch for these when you’re reviewing the uploaded files. Even with the most meticulous narrator, a repeated phrase or two can slip through. The finished quality is the responsibility of both of you!

Sandy: I usually catch around 98% of the errors. The tough ones are when I read a word incorrectly but it sounds right at the time (like make instead of makes) so I don’t catch it. Sometimes I can fix it while editing but sometimes I have to re-record the word or sentence. That’s a pain because it holds up the workflow.

The only time I come out is when I need to check a pronunciation – Roz has some pretty atypical words! Oedema? Nebulae? Roentgen??

Roz: Sorry about that…

Sandy: I record six or so chapters at a time, until my voice gets tired, then load them onto my main PC for editing. I splice together the chapters if they are in chunks, then compress the audio and equalize so the sound quality is good. Then I listen to each chapter and follow along using the printed manuscript to make sure it is correct. I try to fix any mistakes, and make a note of the ones I can’t. I also adjust the pauses between lines so it flows dramatically.

Roz: See the first post about establishing the perfect pause!

Sandy: I listen for the best takes and remove the bad ones, and cut out extra noises like mouth clicks and breaths. I usually end up listening to each recorded line at least twice, sometimes as many as five or six times. I spend more time on the dramatic passages because those feel important to get right.

Roz: This is like the writing process!

Sandy: Once a chapter is complete I run a range check to make sure it fits within the ACX parameters. I adjust the volume as needed, then export it as an MP3 ready to upload.

Incomplete chapters waiting for pickups get put aside until after my next recording session so I can drop in re-recorded lines. Sometimes the new lines need to be tweaked to get them to match the original recording – different days can sound quite different.

sandy's micEquipment

Sandy: Do your research before spending money on equipment. Get the best setup you can afford because when you are recording a solo speaking voice there isn’t much to hide behind, and there is only so much you can do in post-production.

Most audiophiles recommend a high-end microphone with a pre-amp to convert the analogue sound to a digital signal. The pre-amp is almost as important as the mic, so if you go this route you have to spend quite a bit of money to get a good sound. If you are planning on doing professional recording full time this is probably the way to go.

USB microphones have a built-in pre-amp, but traditionally sound tinny and aren’t warm enough for audiobooks. However, they have come a long way recently because of the huge rise in amateur voiceover work for video blogs and podcasts. The mic I bought is a high-end professional brand (Shure PG42) and tuned for voice recording. Since this voice project is a bit of an experiment for me, I wasn’t prepared to buy a full mic/pre-amp system, so I invested in one of the best USB mics.

There are some great resources out there for mic comparisons, such as this one. The other advice I would offer is to give yourself a crash course in post-production – for compression, equalisation and to remove mistakes and odd noises. There is a ton of great info on the web, such as articles like this …and video tutorials like this. ACX also has some very useful tips.

Roz again:

Oh cripes, two names that sound the same!

This really caused a hiccup, and made me rather unpopular with Sandy. Listening to her chapters, I discovered I had a Gene (the main hypnotist character) and a neighbour Jean. On the page, they are perfectly distinct, but in the ears… they sound the same. This gave Sandy some extra rerecording and messed up our schedule.

I talked in my previous post about the pronunciation guide. When you write this, check you don’t have two names that a listener might get confused!

Ooh, knotty word

Less of a problem, and certainly more amusing, was a word that didn’t translate well to a US accent. There’s a line where Gene, the hypnotist, is described as wearing a ‘knotty’ jumper. In Sandy’s voice it came out as ‘naughty’. I imagine comics letterers have the same problem with ‘flick’.

This might not have mattered, except that the knotty jumper occurs again in tense scenes that would be rendered farcical if the listener thought a character was wearing a ‘naughty’ jumper. We decided to stay the right side of serious and replace it with a less troublesome ‘rough-knit’…


UPDATE: we’re now finished and you can find the finished audiobook here (US) and here (UK). And you can find out more about My Memories of a Future Life here.

mmaudioheads bigger

In the meantime, do you have any questions or tips on working with audio? Do you listen to audiobooks? What do you like or not like about them? Share in the comments!

Stop press! Once your audiobook is finished, you’ll need to market it. Joanna Penn has some great tips in this post.


How to make an audiobook using ACX

Do you want to release your title as an audiobook? If you live in the US, you can go through ACX, the DIY arm of Audible, but ACX wasn’t open to UK authors – until now. For the past month, I’ve had both my novels in production as a test pilot, and now I can tell you what I’ve learned so far about offering a title, choosing a narrator and working with them.


What’s ACX?

Good question. ACX is a network where narrators and producers can meet authors who want their work released as audiobooks. Once you’ve hooked up, you can then use the site as an interface to create the book, keep track of contracts and monitor sales. In short, it’s genius.

Setting up

You know how tedious it is every time you set up an identity on a new site? All that form-filling and profile-making? ACX requires minimal faff. Once you tell them who you are and what book you’d like to offer, it pulls the detail off Amazon.

Getting voices

You can:

  • opt to narrate and produce the audiobook yourself, but to do this you must have professional-quality equipment and experience of sound editing, or the book won’t pass the quality check.
  • pluck a willing narrator/producer out of the ether – this is what I did.

Pitching your book

Next to your book info, you can add notes to make your book more attractive to collaborators – your platform, sales figures and anything else that will convince them you’re worth working with. Which brings me to…


Making an audiobook isn’t cheap. An average novel is about 10 hours of narration (roughly 90,000 words) and is likely to cost $200 or more per finished hour.

You have several options if you’re seeking a narrator/producer on ACX

  • pay up front
  • pay a royalty share (which I did)

All the ins and outs of this are much better explained on the ACX site, so check them out there.

mm acx

My acx journey – mistakes made and lucky discoveries

 Choose an audition passage

When I talked to the ACX crew, they told me many writers put up the first few pages as the audition piece. This can be a mistake, because the beginning may not be typical of the book’s action. I looked for a challenging scene with dramatic dialogue as well as the narrator’s internal thoughts, which I felt would test the reader’s approach more usefully.

I added notes about the context of the scene and the style of the book – and waited for auditions.

And lo, they rolled in. (This was in itself a wonderful surprise.) Once I got over the novelty, I realised I needed to tweak my presentation.

Accent& mistake #1 – ACX lets you specify the age, style and accent of the reader. Age and style were easy enough to choose, but accent caused me more trouble. I assumed this had to be British as, obviously, I’m a Brit, my characters are Brits and I write with British language. However, ACX is predominantly US, so that vastly reduced the available talent pool. Some of the voice actors did very credible Brit accents. Some couldn’t pull it off and sounded Chinese or German. Others ignored my stipulation – quite wisely as it turned out they sounded just fine in their natural accents. So I quickly realised accent was a detail that didn’t matter, and edited my directions. Indeed the narrator I chose is American.

Another reason to choose an accent other than your own is if the majority of your readers are in another territory. I sell a lot in the US, so an American accent might make them feel more at home.

Accent isn’t the only deciding factor, though.

Suitability for the material – While the narrator might be able to do a good job with the audition scene, you have to be sure they’ll interpret the whole of your book in the right way. A Regency romance needs a completely different approach from literary fiction, and I can imagine it’s a nightmare to realise your narrator simply doesn’t ‘get’ your book. If you have a contender, poke around in their ACX profile and follow up any websites where they demonstrate other books they’ve narrated. Also, ask them what they like to read.

Acting versus reading & mistake #2 – some books benefit from a reader who will do a lot of gutsy acting, including distinct voices for the main characters. But for most fiction, that’s too much. They’re journeys in prose and need a more intimate, subtle treatment, which might even sound flat to some ears. Listeners know they’re being read to. They don’t need rollicking declamation – or music or sound effects. And a good reader can make it clear who’s talking without bursting into different voices, so you actually need less ‘acting’ than you might think.

I made another mistake here in my original guidance notes for the audition. I didn’t ask for different character voices, but I did explain the book had sections in a different tone – the female narrator, and the future incarnation who was a male version of her. Thankfully, before it went live a friend pointed out that this might cause a lot of horrible baritone overacting, and that I should simply let the text do the work.

Ultimately, you choose a narrator on a hunch that they fit your work. One author I spoke to at LBF said he knew when he’d found his because the guy sounded like the ideal voice he’d have chosen – but better. That’s how I found mine too – although it was by a more roundabout route.


Meet Sandy

And here is Sandy Spangler – my narrator!

I had a shortlist of possible voices, including seasoned Broadway actors, but there was one question I couldn’t answer. When I looked into their backgrounds, none of them had narrated a novel like mine, and I was worried they wouldn’t get it. Then I remembered a friend who I’d heard do narration work – on a computer game, of all things. At the time, I noticed how she had an insightful, feisty quality that reminded me of Laurie Anderson. Even better, I knew her reading tastes made her a good fit.

I contacted her. She didn’t know about ACX, but she was keen to give it a go, registered and sent me an audition. Her reading was just right – inhabiting the material with well-judged expression and I knew the book would suit her personality. If you’ve been a subscriber here for a while you might recognise her from some of the goofy photos I’ve used on my posts. But her voice absolutely suits my kind of fiction, and if yours is like mine you might like her too (here’s her ACX page).

Our process

Here’s how we’re working.

Pronunciation guide – All books have peculiar words and names and you need to warn your narrator of these. We set up an editable file on Google docs. As I said, Sandy’s American and I’m a Brit, so we had to decide whether she should pronounce words like ‘leisure’ and ‘z’ in the UK or US version. I decided we could fiddle endlessly with this so I asked her to do whatever was natural. If we tried to anglicise everything there would be certain words we’d miss, or the stresses would still be American. And I didn’t want to get in the way of her doing the most instinctive job. So she says tomayto while I say tomahto. No big deal.

Pace – one of the first tasks is to approve the first 15 minutes of the audiobook. Sandy was afraid I might think she was reading too slow, but I felt the text suited a measured pace. ACX actually advise that you err on the side of slow because listeners can artificially speed the reading up if they want.

Pauses – you need pauses between paragraphs, scene switches, and maybe in other sections too. We spent an email exchange identifying exactly the right kind of pause for each.

Listen to finished chapters – You need to set aside time to listen to chapters as your narrator uploads them to ACX. We have a schedule and a chart where we tick off chapters I’ve approved or asked for modifications (usually these are pronunciations or stresses). ACX gives you a time code so you can pinpoint exactly where an edit is needed.

Next time, I’ll delve into the narrator’s side, including what exactly is involved in creating an audio book.

UPDATE: we’re now finished and you can find the finished audiobook here. And you can find out more about My Memories of a Future Life here.

mmaudioheads bigger

In the meantime, tell me: have you made an audiobook? Are you tempted to? Have you any tips to offer or questions to ask? Share in the comments!