How Long to Edit an Audiobook: Hours Per Finished Hour, VoiceEditSuite
← Guides

Workflow

How Long to Edit an Audiobook: Hours Per Finished Hour

Updated October 16, 2026·15 min read
Black and white photograph of an editor working through a long audiobook timeline late at night

Editing an audiobook takes about 3 hours for every finished hour of audio. That is the figure in ACX's own production guidance and the ratio Audible's in-house editors aim for. Add about 2 hours of recording and 1.2 hours of proof listening, and one finished hour costs roughly 6.2 hours of work, so a 10-hour book is a 60-plus hour project for a narrator who does their own post. Experienced narrators who record punch and roll report getting the whole job down to 3 hours per finished hour, and in one documented case under 2.

Audiobook work is quoted per finished hour (PFH): one hour of final, listenable audio, however long it took to make. So the gap between the finished hour and the real hours is where a narrator's actual pay lives. This guide lays out every published ratio we could source, explains why they disagree, walks a 10-hour book through the numbers, and shows what a PFH rate pays per working hour at each ratio. If you are a podcaster, the same mechanical editing work (retakes, breaths, gaps, a proof listen) is what eats your evenings too, so most of what follows about where the time goes applies to an episode as much as a chapter.

The Published Ratios, Source by Source

There is no single industry standard written into a contract, but the numbers that are actually published cluster tightly. Here they are, each with what it counts, because most disagreement between sources is really disagreement about which stages are in the total.

Every figure here is as its source states it. The full-stage estimates sit near 6 hours; the lower ones count fewer stages or describe an optimized workflow.

ACX: 2 + 3 + 1.2 = 6.2 hours

ACX's help article on narrating an audiobook gives the clearest breakdown anywhere. It takes about 2 hours to record one finished hour, an editor (who might be the narrator) about 3 hours to edit each finished hour, and about 1.2 hours to review each hour of finished audio. Its stated total is 6.2 hours per finished hour, with variation for individual factors. The same article warns that new narrators generally take between 5 and 7 hours to record one hour, and that the time should come down as they get fluent with their tools. ACX's own blog says much the same in round numbers: an experienced ACX producer will spend roughly six hours in production for every hour of completed audio, so a five-hour title means at least 30 hours of work.

Audible's editors: 3:1 on the edit, 1.2:1 on QC

When ACX published how Audible Studios reviews its own final audio, it gave the in-house targets: editors aim for 3:1 on the edit pass and 1.2:1 on the QC pass. For a 10-hour book that is roughly 30 hours of editing and 12 hours of QC, and ACX adds that it can take longer depending on the subject matter, the language and the number of errors the narrator made. That last clause matters. Editing time is not a fixed property of the book. It is mostly a function of how clean the raw recording is.

What working narrators and producers report

Narrator Karen Commins published the same 2 / 3 / 1.2 split back in 2011 and told narrators to multiply finished hours by 6, calling it a very conservative estimate, with one important footnote: pre-reading the book and looking up pronunciations is not included. Narrator and producer James Fowler writes that the general rule of thumb for editing is 3:1, that an experienced producer can get below that, and that a single finished hour can mean anything from 3 to 5 hours of studio time. Brian Wiggins' narrator rate guide describes the whole job (recording, editing out bad takes and mouth noise, listening against the manuscript, mastering) as generally a 3:1 time ratio, so a 10-hour book takes about 30 hours.

ACX's "Money Talks" article for rights holders sits in the middle. It says the narrator is usually spending two hours reading in the booth, then another two to three hours editing, mixing and mastering each finished hour, and sums it up as about four hours of work per finished hour. Notice what that version leaves out: a separate proof listen. Put the proof pass back in and you are near ACX's 6.2 again.

Why the numbers disagree

The 6-hour figures count every stage: record, edit, proof and master, done by one person working the traditional way (record everything, edit afterwards). The 3-hour figures assume punch and roll, where most of the editing happens during the recording, and often a narrator with years of practice. Neither is wrong. Quote from the one that matches how you actually work, and measure yourself on a chapter before you trust either.

Stage by Stage: Where the Hours Go

Before the booth: prep (not in most totals)

None of the published ratios include preparation, and Commins says so explicitly. Reading the whole book before you record, building a character list, and looking up every name and place are real hours that a PFH rate also has to cover. A pronunciation list made before the first session is one of the cheapest ways to cut editing and pickup time later, because a mispronounced name in chapter 3 that you only catch in chapter 20 means re-recording every instance. Our guide to building a pronunciation glossary for narration covers how to set one up.

Recording: about 2 hours per finished hour

Even a clean narrator re-reads lines, stops for water and rests the voice, and the finished hour is shorter than the session because the mistakes and the pauses come out. ACX's 2:1 is a working professional's pace; its 5 to 7 hours for newcomers is the honest starting point. Character-heavy fiction is slower in the booth as well as in the edit, which we come back to below. A 93,000-word manuscript at ACX's planning figure of about 9,300 words per finished hour is a 10-hour book, and at 2:1 that is 20 hours behind the microphone.

Editing: about 3 hours per finished hour

This is the biggest single stage in every full estimate, and the time does not go where most people guess. The artistic decisions are quick. The clock disappears into mechanical work:

  • Finding the keeper take. You read the sentence three times mid-chapter. Which one stays? Scrubbing back and forth to compare takes is the single biggest time sink when you record straight through.
  • Breath and gap management. Hundreds of breaths per chapter, each one a small decision: keep, soften or cut. Then the pause after every sentence and paragraph has to feel consistent across a whole book.
  • Mouth noise and clicks. The same small repair applied hundreds of times, and much of it cheaper to prevent at the mic than to fix in the edit.
  • Hunting the flub you half remember. You know there was a stumble somewhere in chapter 7. Finding it means listening, not editing.

James Fowler adds a detail that every dialogue-heavy narrator will recognize: jumping between characters creates tiny hesitations that you do not hear while performing but do hear in the edit, and they all have to be tightened. That is why the same narrator can edit a narrative non-fiction title noticeably faster than a novel with a large cast.

Proofing and QC: about 1.2 hours per finished hour

A narrator at a home studio desk marking pickups in pencil on a printed manuscript page, headphones on, with a waveform open on the laptop screen beside it
The proof listen: every finished minute heard against the manuscript, with each error marked for a pickup.

The proof pass is a full listen of the finished audio against the text, which is why it can never be shorter than the book itself. ACX's 1.2:1 leaves just a little room for stopping and noting. According to ACX's write-up of the Audible Studios process, the QC listener checks that the words are voiced exactly as written, listens for mispronunciations, and checks that character names, place names, voices and accents stay consistent across the whole book. Errors are marked as pickups, and the guidance is to mark the sentence before and after each error too, so the re-recorded section can be cut in without a seam.

Pickups themselves are not in any published ratio, and their cost depends on when they are found. A misread caught in your own proof pass is a two-minute fix. The same misread caught by a rights holder after delivery means opening the session again, matching your voice and room from weeks earlier, re-editing, re-mastering and re-delivering. That is the strongest argument for proofing every chapter before you master it, not after the book is finished.

Mastering to spec: folded into the post hours

ACX's "Money Talks" estimate bundles mastering into its two to three hours of editing, mixing and mastering, and none of the sources we found gives mastering its own ratio, so we will not invent one. What is fixed is the target. ACX's audio submission requirements ask for every file to measure between -23 dB and -18 dB RMS, peaks no higher than -3 dB, a noise floor no higher than -60 dB RMS, and 1 to 5 seconds of room tone at the head and tail. ACX's own Audio Lab checks RMS and peaks but says it does not detect noise floor or editing problems, so you still have to check those yourself. Our ACX loudness requirements guide and room tone and noise floor guide cover how to hit each number first time, which matters because a rejected book is re-work across every chapter.

What Changes the Ratio

Punch and roll: the biggest single lever

Punch and roll means that when you flub a line, you stop, roll back a few seconds, listen to the lead-in and re-record over the mistake in place. The edit happens while you record, and the session file comes out close to clean. The best documented effect is Brian Wiggins' own: he writes that his production used to take 4 to 5 hours per finished hour, and that after switching to punch and roll and using an automated tool for QA, his average is 1.8 hours. That is one narrator's figure, not a guarantee, but it shows the size of the lever. The trade-off is that the booth hours can stretch a little, because every stop and roll-back costs a few seconds; what you get back is most of the 3:1 edit. Our punch and roll recording guide explains how to set it up in the common DAWs.

Experience

ACX's 5 to 7 hours of recording for new narrators against 2 hours for experienced ones is the clearest sign that experience compresses everything. Fewer flubs means fewer takes to compare, fewer pickups and a faster proof. Wiggins' rate guide ties pricing tiers to the number of completed books for the same reason, suggesting a moderately experienced narrator should have at least 30 titles behind them. If you are on your first few books, budget for the high end and track your real hours per chapter, because your own number is worth more than anybody's average.

Genre, dialogue and the manuscript

Fowler is direct about it: character-heavy audiobooks take much longer in the edit than narrative non-fiction, and character dialogue is more tiring to perform, which feeds more errors into the edit. ACX's Audible Studios guidance adds subject matter and language as factors. Technical non-fiction full of names, numbers and terms is slow in a different way: the reading is careful and the proof pass has to check every figure.

Whether you do your own post

If you hire an editor and a proofer, your own time drops to roughly the recording and pickups, but the money comes out of the rate. ACX's "Money Talks" article gives one rough split for rights holders: about $200 PFH for narration and another $200 PFH for post-production, putting a fully produced book in the $300 to $400 PFH range it mentions. Whichever way you work, someone spends the editing and proofing hours. The question is only whether they are yours or paid for.

Worked Example: A 10-Hour Book

Take a 93,000-word novel and one narrator doing all of their own post-production the traditional way. ACX's planning figure of about 9,300 words per finished hour makes it a 10-hour book. Then apply ACX's ratios stage by stage.

The same book at the published ratios. Prep and pickups come on top.
  1. Finished length: 93,000 ÷ 9,300 = 10 finished hours.
  2. Recording: 2 × 10 = 20 hours in the booth.
  3. Editing and mastering: 3 × 10 = 30 hours, matching the 30 hours Audible's editors would budget for the same book.
  4. Proof and QC: 1.2 × 10 = 12 hours of listening against the manuscript.
  5. Total: 62 working hours, before pre-reading the book and before any pickups.

How that lands on a calendar depends on how many focused hours a day you can do. As a plain illustration, at 6 hours of work a day, 62 hours is a little over 10 working days; at 4 hours a day it is about 16. Booth time is usually the constraint, because voices tire long before editing fingers do, so many narrators record in the morning and edit later in the day.

Now run the same book at the faster figures. At the 3:1 norm in Wiggins' rate guide it is 30 hours of work. At his 1.8 punch and roll average it is 18 hours. The book did not change; the workflow did. That 44-hour spread between 62 and 18 is the most important number in this article, because it is the difference between a rate that pays and one that does not.

Measure your own ratio on one chapter

Time a single chapter end to end: booth, edit, proof, master. Divide your total minutes by the chapter's finished minutes. That is your ratio, and it will be more accurate for your next quote than any average here. Do it again after ten chapters, because the number moves fast early on.

How Per-Finished-Hour Rates Map to Real Hours

A PFH rate only means something once you divide it by your hours per finished hour. The table below does that division for four reference rates and three published ratios. The arithmetic is ours; the rates and ratios are sourced on the chart.

Same rate, three workflows. Workflow changes your hourly pay as much as the rate does.

Read across a row and the point is obvious. At $100 PFH, the minimum ACX requires on Royalty Share Plus projects for a SAG-AFTRA narrator's health and retirement contributions, the traditional 6.2-hour workflow pays about $16 an hour before tax and equipment. Using the 1.8-hour workflow, the same rate pays about $56. At $250 PFH, the upper end of the "very experienced" non-union tier in Wiggins' guide, the gap is $40 an hour against $139.

Commins made the same point in 2011 with a simple formula: multiply the real hours by the hourly rate you need to live on, and that is the minimum the PFH has to cover. If you are weighing a royalty share deal against a flat rate, the hours are the same either way, which is why our comparison of ACX royalty share versus per finished hour starts with your production ratio. On the union side, SAG-AFTRA's audiobook work is also paid per finished hour, with health and retirement contributions on top. We could not confirm a current published minimum from the union itself, so check the terms for a given publisher with SAG-AFTRA directly rather than relying on figures quoted secondhand.

There is a demand side to this as well. The Audio Publishers Association's 2026 Sales Survey put US audiobook sales at $2.43 billion for 2025, up 9 percent, with publishers reporting over 750,000 active titles, 43 percent more than the year before. A growing catalog means more books to narrate, and more books chasing the same pool of rates, which makes your hours per finished hour a competitive advantage as much as a cost.

Where Automation Saves Time, and Where It Does Not

Split the post hours into two kinds of work. Judgment work: is this read right, is this character consistent, does this pause serve the line. Pattern work: which of three takes is the complete one, where are the false starts, which breaths are too loud, are the gaps even. Judgment work is the craft and stays with you. Pattern work is where software can take real hours off the 3:1, and it is the same work that fills a podcaster's edit of a long interview.

↗ Try the tool

Auto Clean Up

Auto Clean Up does a first pass on the pattern work: it transcribes your session, keeps the complete take of a repeated line, drops false starts, quietens breaths and tightens long gaps, then lists every decision so you can reverse any of them with one click. You still listen and approve; it just takes the scrubbing out of the edit.

Open Auto Clean Up →

What automation does not replace is the proof listen. A tool can flag a dropped or swapped word against the manuscript, and that speeds up the QC pass a great deal, but the publisher's QC is listening for the same things Audible's editors listen for: the right word, the right pronunciation, the same character voice in chapter 30 as in chapter 2. Somebody has to hear the book. Automation changes how much of the 1.2 hours goes to hunting and how much goes to fixing. It does not make the listen optional.

The honest summary: punch and roll moves most editing into the booth, automation removes much of the mechanical cleanup that is left, and a disciplined chapter-by-chapter proof stops errors from turning into pickups after delivery. Together those are how working narrators get from the 6-hour estimates toward the 2 to 3 hour ones.

The Per-Chapter Checklist

  • Pre-read done and pronunciations listed before the first session.
  • Recorded punch and roll, so the session file is nearly clean already.
  • Mechanical edit done (best takes, gaps, breaths, clicks), by machine or by hand.
  • Proofed against the manuscript, every chapter, before mastering.
  • Mastered to ACX's numbers (-23 to -18 dB RMS, peaks at -3 dB or lower, noise floor at -60 dB or lower, 1 to 5 seconds of room tone) and checked on the delivered file.
  • Your real hours for the chapter written down, so your next quote uses your ratio.
  • Raw session archived untouched, in case pickups are needed later.

Frequently asked questions

How long does it take to edit one hour of audiobook?

About 3 hours per finished hour. That is ACX's figure and the ratio Audible's in-house editors aim for. Experienced narrators who record punch and roll often edit faster, because much of the editing happened in the booth.

How long does it take to produce a 10-hour audiobook?

At ACX's ratios, about 62 working hours for a narrator doing everything: 20 recording, 30 editing and 12 proofing, plus prep and pickups. At the 3:1 norm cited by working narrators it is about 30 hours.

How many words is one finished hour of audiobook?

ACX uses about 9,300 words per finished hour. Divide the manuscript's word count by 9,300 for a planning estimate; a 93,000-word book is roughly 10 finished hours.

What does per finished hour (PFH) mean?

It is pay per hour of final, delivered audio, not per hour worked. A $250 PFH rate on a 10-hour book is $2,500 whether production took 20 hours or 60, which is why your hours per finished hour decide your real hourly pay.

How long does audiobook proofing take?

ACX puts review at about 1.2 hours per finished hour, matching the 1.2:1 QC ratio Audible's editors use. It can never be shorter than the book, because every minute has to be heard against the text.

What is the fastest way to cut audiobook editing time?

Record punch and roll, so mistakes are fixed as you go. One narrator documented dropping from 4 to 5 hours per finished hour to 1.8 by combining it with automated QA. Prep pronunciations first and proof each chapter before mastering.

Sources and Further Reading

Keep reading

Career

How to Choose a Voice Over Coach: Cost, Formats and Red Flags

13 min read

Podcasting

How to Start a Podcast in 2026: A Complete Beginner's Guide

18 min read

Technique

Self-Directing When Nobody Else Is on the Call

9 min read