You finished your manuscript. Months of writing, revising, and polishing. Now you want an audiobook — but hiring a narrator costs thousands, and studio time adds up fast. AI voices have closed the quality gap dramatically, yet most authors still struggle with the actual workflow: how do you go from a 60,000-word document to a set of polished, chapter-by-chapter audio files ready for distribution?
This guide walks you through a repeatable process. You'll learn how to segment your manuscript intelligently, apply voice and pacing settings at scale, fine-tune narration for dialogue and emphasis, and export files that meet distributor specifications from platforms like ACX, Findaway Voices, and Authors Republic.
The entire workflow assumes you're working in EchoLive's Studio editor — a segment-based timeline designed for exactly this kind of long-form project.
Preparing Your Manuscript for Import
Before you touch any audio tool, your manuscript needs structure. Distributors require separate files per chapter, so your source document should clearly delineate chapter breaks.
Clean Up Your Source File
Strip out front matter that won't appear in audio — table of contents, dedication pages (unless you want them narrated), and any formatting artifacts from your word processor. Keep chapter headings consistent. "Chapter 1: The Arrival" works better than inconsistent styles like "ONE" followed by "Chapter Two."
Save your manuscript as a .docx, .txt, or .pdf. EchoLive's Smart Import accepts all three formats and uses AI-assisted segmentation to detect chapter boundaries, paragraph breaks, and structural elements automatically.
Import and Verify Segments
Once imported, your manuscript appears as a series of segments in the Studio timeline. Each segment represents a logical block — typically a paragraph or scene break. Review the segmentation before proceeding. The AI does a strong job detecting structure, but you'll occasionally want to merge short segments (a single line of dialogue that got separated) or split long ones (a dense paragraph that needs a breath point mid-way).
This verification step takes 10-15 minutes for a typical novel-length manuscript. It's worth the time — clean segmentation makes every subsequent step faster.
Choosing and Locking Your Narrator Voice
Consistency is everything in audiobook narration. Listeners notice when a voice shifts tone or timbre between chapters. You need to select a primary narrator voice and lock it across your entire project.
Audition with Real Text
Don't pick a voice based on a single demo sentence. Copy a paragraph from your manuscript — ideally one with both narration and dialogue — and audition three to five voices from EchoLive's catalog of 650+ neural voices. Listen for clarity, warmth, and how the voice handles punctuation pauses.
Use Voice DNA recommendations to discover voices that match your genre. A literary fiction novel needs a different texture than a thriller or a children's book. Save your top candidates as favorites, then do a longer test: generate a full chapter with each finalist and listen on headphones, in the car, and through a phone speaker. Your readers will use all three.
Set Project-Level Defaults
Once you've chosen your narrator, set it as the project default. Every new segment inherits this voice automatically. You can still override individual segments later — useful for dialogue or chapter epigraphs read in a different style — but the default ensures consistency without manual repetition.
Audiobooks continue to be a growing format in the U.S. media market, and indie authors who capture even a small share of that demand can benefit enormously from professional-quality narration at scale.
Batch Editing for Pacing and Style
A novel-length manuscript might contain 300-500 segments. Editing each one individually would take days. Batch operations let you apply settings across your entire project — or across selected chapters — in seconds.
Apply Consistent Pacing
Select all segments in a chapter (or the entire project) and set a base speaking rate. For most fiction, a slightly slower pace — around 0.9x to 0.95x — sounds more natural than the default speed. Non-fiction and self-help titles often work better at 1.0x with slightly longer inter-segment pauses.
Use EchoLive's batch settings panel to apply pacing globally, then adjust individual segments that need different treatment. Action sequences might benefit from a slightly faster rate. Reflective passages or emotional beats often land better with more deliberate pacing.
Handle Dialogue and Scene Breaks
For dialogue-heavy chapters, you have two options. The simpler approach: use SSML emphasis and prosody controls to add slight pitch variation and pacing changes within a single narrator voice. This keeps the listening experience cohesive while signaling dialogue shifts.
The more advanced approach: assign a secondary voice to dialogue segments. This works well for books with a clear two-character structure (romance, buddy comedies) but can get unwieldy with large casts. Start simple — you can always add complexity in revision.
For scene breaks and chapter transitions, insert break segments with 1-2 seconds of silence. Distributors expect clean separation between chapters, and listeners appreciate the breathing room.
Fine-Tuning Problem Passages
Every manuscript has passages that trip up text-to-speech engines. Unusual names, technical terminology, intentional sentence fragments, and poetry all need attention.
Pronunciation and Phonemes
Character names are the most common issue. If your protagonist is named "Caelum" and the engine defaults to an unexpected pronunciation, use EchoLive's visual SSML tools to set a phoneme override. You define the pronunciation once, and it applies everywhere that name appears.
The same approach works for made-up words in fantasy and science fiction, brand names, or regional dialect spellings. Build a pronunciation guide early in your project — it saves time across subsequent chapters.
Emphasis and Emotional Beats
Italicized words in your manuscript usually signal emphasis. Smart Import preserves this formatting and converts it to SSML emphasis tags automatically. Review these — sometimes italic text is used for internal thoughts or foreign words rather than vocal stress, and you'll want to adjust accordingly.
For critical emotional moments — a revelation, a confession, a climactic line — manually set prosody adjustments. A slight decrease in rate combined with increased volume on key words can transform a flat reading into something genuinely affecting.
Research from Stanford University's Virtual Human Interaction Lab has shown that listeners form emotional connections with synthetic voices when prosody mimics natural human speech patterns — pauses before important words, pitch variation during emotional content, and tempo changes that match narrative tension.
Exporting Distributor-Ready Files
The final step transforms your polished project into files that meet distributor specifications. Different platforms have different requirements, but the common standard is straightforward.
ACX and Audible Requirements
ACX's current upload requirements are for separate MP3 files encoded at CBR (constant bit rate), 192 kbps or higher, 44.1 kHz. Audio levels must have peaks no higher than -3 dB, RMS between -23 dB and -18 dB, and a noise floor below -60 dB. ACX also requires brief room tone at the start and end of each file, but chapter files do not have a blanket 20-minute minimum. Always confirm the latest details before submission: https://help.acx.com/s/article/audio-submission-requirements
EchoLive's production exports handle the format requirements automatically. Export as MP3, select your bitrate, and the platform generates individual files per chapter based on your segment groupings. Name your exports following the distributor's convention — typically the book title followed by the chapter number.
Other Distributors
Findaway Voices and Authors Republic accept similar specifications. WAV exports at 44.1 kHz / 16-bit work universally if you want maximum flexibility. The files are larger, but they give you a lossless master you can convert to any format later.
For a full breakdown of document-to-audio conversion options — including format selection, segment bundling, and timeline exports — EchoLive's use-case guide covers the specifics.
Quality Checking Before Submission
Listen to the first and last 30 seconds of every chapter file. Check for clipped audio, unnatural pauses at segment boundaries, and pronunciation errors you might have missed. Distributors will reject files with technical issues, and re-uploading delays your launch.
Budgeting Your Audiobook Project
A typical 80,000-word novel produces roughly 8-10 hours of audio. With EchoLive's minute packs, you can produce an entire audiobook without a subscription commitment. The Plus pack (1,000 minutes for $50) covers most novel-length projects with room to spare for revisions and re-generations.
Minutes never expire, so you can work at your own pace — one chapter per week or the entire book in a weekend sprint. Every paid account unlocks the full voice catalog, meaning you're not locked into lower-quality voices at entry-level pricing.
Conclusion
Producing a professional audiobook no longer requires a studio budget or months of coordination with a human narrator. By segmenting your manuscript cleanly, choosing a consistent AI voice, batch-editing pacing and style settings, fine-tuning problem passages with SSML, and exporting to distributor specifications, you can go from finished manuscript to published audiobook in days rather than months.
The workflow is repeatable — once you've built your first audiobook, the second one goes twice as fast. If you're ready to turn your manuscript into audio, open EchoLive's Studio and start with a single chapter. You'll hear your words come alive in minutes.
Originally published on EchoLive.
