Start with the room, not the microphone
Untreated rooms are the number one reason home audiobook recording gets rejected by publishers and platforms. Before spending on gear, kill reflections: record inside a closet full of clothes, hang moving blankets around your desk, or build a corner of foam panels. A $100 dynamic microphone in a dead room beats a $1,000 condenser in an echoing bedroom every time.
For the microphone itself, a dynamic mic (SM58-class or a broadcast dynamic) forgives an imperfect room; a large-diaphragm condenser rewards a treated one. Record in mono at 44.1 or 48 kHz WAV. Aim for peaks around −6 dB, never touching 0.
Narration pacing: slower than feels natural
First-time narrators read 20 to 30 percent too fast. Audiobook listeners follow at speech pace without the text in front of them, so land near 150 to 160 words per minute and let punctuation breathe. Reading from a teleprompter-style display, with the current sentence centered and large, keeps your eyes from racing ahead, which is what usually drags the pace up.
Fix mistakes with punch-and-roll, not with editing
Professional studios repair narration mistakes with punch-and-roll: when you flub a word, the engineer rolls playback a few seconds before the mistake, you hear your own cadence, and recording takes over at the punch point. You re-say one sentence, matched in tone and rhythm, instead of re-reading the page or spending your evening cutting a timeline in an editor.
This is the core of the Medrak Recorder workflow. Every sentence of your script is a line on screen; a mistake is a click on the waveform and one press of Punch Back. The recorder plays your run-up, drops a marker, and re-records from exactly that word, keeping everything before it. What used to be an editing session becomes part of the read itself.
Record by chapter, organize by line
Treat every chapter as a project and every paragraph or sentence as a script line. Line-based script recording gives you three things a long timeline never will: you always know exactly what has been read, retakes attach to the sentence they replace, and exporting produces files your editor can actually use — separate WAV clips per line, or one continuous WAV with an SRT file that marks where every line begins.
That SRT matters more than it looks: proof-listening against text with timestamps is how you catch the misread sentence in hour seven without scrubbing for it.
Keep a consistency ritual
- Same room, same mic position, same input gain every session; photograph your setup.
- Re-listen to the last minute of the previous session before recording the next one.
- Record a short room-tone sample each session for your editor.
- Watch your levels while you read — a meter you can see beats one you check afterward.
- Stop while your voice is still fresh; fatigue is audible long before it is painful.
The tools, in one place
You can assemble this workflow from a DAW, a script PDF, and discipline. Or use a recorder built around it: Medrak Recorder is a teleprompter recorder that scrolls your manuscript as you speak (speech recognition follows your voice, hands-free), records each line as its own take, repairs mistakes with punch-and-roll, and exports editor-ready WAV. It runs in the browser, and a desktop edition keeps copyright-sensitive manuscripts entirely on your own machine.
New to script-linked recording? Start with what a teleprompter recorder is, or read the script recording workflow guide next.