Last Updated: September 14, 2026
Quick Answer
Good voiceover direction gives a performer something to play, not an adjective to hit. Cover four things: tone, pace, emphasis, and pronunciation.
- Describe a situation or feeling, never just "sound excited"
- Mark emphasis on exact words, in the script itself
- Spell tricky names phonetically to remove the most common cause of a retake
- Keep direction notes separate from your script so they do not count toward word count
Often times reads that come back wrong are not talent problems, but brief problems. A performer can only act on what you gave them, and "professional but friendly" gives them almost nothing.
Why Direction Matters As Much (Or More) Than Voice Selection
Buyers spend an hour listening to demos and two minutes writing the brief. That ratio is backwards.
Demos tell you what a performer sounds like at their best on copy they chose. Your direction tells them what to do with copy they have never seen, for an audience they do not know, in a tone you have not described. The second input has more influence on the delivered file. Buyers favor conversational, authentic reads over the polished announcer sound, and that quality comes from direction as much as casting.
Casting directors say the same thing about performers: what separates a good read is acting ability and knowing how to make choices (Backstage, 2020). Your direction is what those choices get made against. Vague direction produces a safe, generic read, because that is the lowest-risk interpretation available. Casting professionals describe the underlying skill as knowing how to analyze copy and give yourself direction, on top of whatever written notes arrive with the script.
The Four Things Every Brief Needs
Tone. The emotion behind the read. Not "upbeat" but "like you just heard good news you can't wait to share."
Pace. Fast, natural, or slow, plus any spots that need to shift. Phone numbers and legal lines slow down. Lists move.
Emphasis. The exact words carrying meaning. Mark them in the copy rather than describing them afterward.
Pronunciation. Brand names, acronyms, place names, and anything a stranger might say wrong. This single habit removes the most common reason audio comes back needing a redo.
Cover those four and you have out-briefed most buyers. Our practical notes on writing a voiceover script go deeper on word count math and markup, and pair well with this page.
Replace Adjectives With Situations
The single biggest upgrade to a brief is trading adjectives for scenarios. Adjectives describe a result. Scenarios give a performer something to do.
|
Weak direction |
Direction that works |
|---|---|
|
Sound excited |
Like you are telling a friend news you can't hold in |
|
Be professional |
Like a doctor explaining something clearly and kindly |
|
Warm and friendly |
Like you are greeting a regular customer by name |
|
Authoritative |
Like you already know this is true and are not arguing |
|
Conversational |
Like you are talking to one person across a table |
|
Urgent |
Like the sale really does end tonight and you like them |
Notice each version implies a connection between speaker and listener. That is what a performer acts. Tone follows from the situation rather than being applied on top of it.
Mark the Script, Not Just the Notes Field
Direction that lives only in a separate notes box asks the performer to hold your intentions in their head while reading. Put the cues where the decisions happen.
Practical markup that works:
- Bold or caps the emphasis word in each line that has one. One per line at most, or you have emphasized nothing.
- Add pause markers where you want breathing room, written as (pause) or a line break.
- Put phonetic spellings inline, right next to the word, like "quinoa (KEEN-wah)."
- Flag tonal shifts in brackets at the line where they happen, like [warmer here] or [drop the energy].
- Note the ending you want, since sign-offs and CTAs are where reads most often land wrong.
One caution on formatting: if you paste a script into an order, keep your direction notes in the direction field rather than inline in the copy, so they do not get counted in your word count. Mark emphasis and pronunciation in the script itself, and put broader tonal notes in the notes field.
Give Reference, But Give It Carefully
A reference read is useful shorthand. Pointing at an existing ad and saying "this energy, but slower" communicates faster than a paragraph.
Two rules make references work. First, say what specifically you want from it, since a performer can't tell whether you mean the pacing, the warmth, or the accent. Second, do not ask for imitation of a recognizable performer, which is both a poor use of the talent you hired and a bad habit given how voice likeness rights are tightening (Holland & Knight, 2026).
References work best alongside your own description, not instead of it. "Like this spot, but less announcer and more like she means it" is a real note. Sending a link with no comment is not.
Direct the Revision, Not Just the Original
The read comes back and something is off. How you describe that determines whether the second pass fixes it.
Weak revision notes describe the problem: "this sounds flat." Strong revision notes describe the target: "can this line feel like you are relieved rather than reporting?" A performer can act the second one.
Be specific about location too. Name the line, the timestamp, or the exact phrase rather than saying the middle section drags. And separate performance notes from technical notes, since "too fast" and "there is a breath before the phone number" are different fixes.
This is where real people beat synthetic audio outright. Listeners are more than twice as likely to trust a human voice (55%) over AI-generated content (23%) (Audacy, 2024), and only a person can act on a note. Directability means you can change one word in one line and get an interpretation back rather than a regenerated take. Our comparison of AI voice generators and human voice actors covers that difference in detail. VoiceJungle includes one free revision on every order, and a clear brief usually means you only need the one.
Listen the Way Your Audience Will
Before you write revision notes, listen in the environment the audio will actually live in. Phone systems get heard through a phone speaker, often by someone already annoyed. Radio spots get heard in a car. Training modules get heard on laptop speakers in an office.
Studio headphones in a quiet room flatter everything and hide the problems your audience will notice. If the read is for phone messaging and IVR, call a phone. If it is an audio ad, play it in a car. If it is eLearning narration, listen at the speed learners will.
The Bottom Line
Direction is the single biggest thing a voiceover buyer controls, and most briefs are one line long. Cover tone, pace, emphasis, and pronunciation. Replace adjectives with situations, because a performer can act "telling a friend good news" and can't act "upbeat." Mark emphasis and phonetic spellings in the script itself, keep broader notes in the direction field so they stay out of your word count, and describe the target rather than the problem when you request a revision. Do that and the first take usually lands. Choose your voice and put a real brief behind it.
Frequently Asked Questions
What should I include in voiceover direction?
Four things: tone (the emotion behind the read), pace (fast, natural, or slow, plus where it shifts), emphasis (the exact words carrying meaning), and pronunciation for any brand name, acronym, or unusual term. Covering those four puts your brief ahead of most, and produces a better first take.
How do I describe the tone I want?
Describe a situation rather than an adjective. "Sound warm" gives a performer nothing to act. "Like you are greeting a regular customer by name" implies a whole situation, and the tone follows from it. Reference a feeling, a person, or a moment instead of a descriptive word.
Should direction notes go in the script or a separate field?
Both, split by type. Mark emphasis and put phonetic spellings inline in the script where the performer reads them. Keep broader tonal and pacing notes in the direction field, since text pasted into the script can count toward your word count and change your project total.
How do I write a good revision note?
Describe the target, not the problem. Instead of "this sounds flat," try "can this line feel like you are relieved rather than reporting?" Name the exact line or phrase, and keep performance notes separate from technical notes like breaths or spacing, since those are different fixes.
Can I send a reference recording as direction?
Yes, and it helps when you say what you want from it. A performer can't tell whether you mean the pacing, the warmth, or the energy unless you name it. Avoid asking for imitation of a recognizable performer, and pair any reference with your own description of the target.
How do I mark pronunciation in a script?
Spell it phonetically right next to the word, like "quinoa (KEEN-wah)." Do the same for brand names, place names, foreign words, and acronyms you want spoken as letters. This one habit removes the most common reason a delivered voiceover needs a second pass.
