Section 5: Synthetic Voice and Audio Direction · Lesson 2 of 5 (22 of 45) · 60 minutes · Evidence: Voice recording script
Take any sentence from a textbook and read it aloud, fast, as if you were the narrator of a film. You will almost certainly stumble, run out of air, or lose the meaning. Writing meant for the eye and writing meant for the ear are different crafts. Today you learn to rewrite your script so a voice, human or synthetic, can perform it well on the first try.
Production journal: write your answer below.
What must a filmmaker understand before they can divide script text into short speakable lines with punctuation pauses and pronunciation cues?
A reader can look back at a long sentence. A listener cannot. In a film the audience hears each line once, in real time, usually while looking at something else. So dialogue and narration must be easy to say and easy to follow. Short words beat long ones. Active phrasing beats stacked clauses. And one idea at a time beats three.
A breath group is the stretch of words a speaker says in one breath, usually three to ten words, ending where a natural pause falls. Divide every long sentence into breath groups and put each on its own line. This is the unit a voice performs. A synthetic voice that receives a 40-word sentence has to guess where to breathe; if you break the line, you decide.
Voice tools and human actors both read punctuation as instruction. A comma asks for a short pause. A period asks for a fuller stop and a falling voice. A question mark lifts the end. A dash cuts a thought short. An ellipsis trails off and invites hesitation. Capital letters or italics mark emphasis, the word the listener must not miss. Use them sparingly: if every word is stressed, none is.
Pronunciation problems come from names, places, acronyms, numbers and borrowed words. Add a cue the first time such a word appears: spell it the way it sounds (for example, ‘Caer Dorran’ becomes ‘care DOR-an’), write numbers out as words when the reading is ambiguous, and spell out letters in an acronym when they should be said separately. Keep a pronunciation list for the whole film. Check your tool’s own way of marking a pronunciation, and keep your original spelling in the master script so the subtitles stay correct.
In Last Light, Tomas has this line on the page: ‘I brought the supplies, which took longer than expected because the tide was out, so I had to anchor off the rocks and carry everything, including your mail, which is a lot.’ Spoken, it is breathless and flat. The student breaks it into units, one per line:
The student marks a pause after ‘thought’, stresses ‘All’, and adds a cue for the place name. The meaning is the same, but now a voice can perform it.
Voice-over writers and audiobook directors prepare a script before a single sound is made. Study their process as if you were reading over an editor’s shoulder, and pause at each decision to ask who decides: the person, or the tool.
The editor first reads the script aloud, to themselves, and marks every place they stumble. Each stumble is a sign that the line needs a break or a rewrite. Then they cut long sentences into breath groups and list every name, place and number that a voice could misread. A tool may suggest pronunciations, but only a person who knows the world of the film knows how a character name should truly be said.
Finally the editor adds the marks that will guide delivery, such as pauses and emphasis, then reads the whole script again at natural speed with a timer. If a line runs much longer than the shot it belongs to, it must be trimmed. Notice that the tool never decides how long a line should be. That belongs to the filmmaker, who knows how long the picture lasts.
As you study, record each decision point in your production journal.
| Decision point | What the filmmaker does | What the tool automates | What still needs human review |
|---|---|---|---|
| Read the script aloud | Marks every stumble, breath and tongue-twister | Nothing; this needs a human ear | Whether the stumbles come from the writing or from the reader |
| Split into breath groups | Breaks long sentences into short lines, one idea each | Can suggest line breaks if asked | That the meaning and rhythm survive the break |
| Mark pauses and emphasis | Uses punctuation and emphasis marks sparingly | Interprets the marks when it speaks | Whether the marks produce the intended delivery, checked by listening |
| List pronunciation cues | Respells names, places, numbers and acronyms | May suggest a reading | That every suggested reading is correct for this film |
Studio task: Mark a paragraph into breath groups and performance units.
Work with a partner on one paragraph, about 60 to 100 words, from your film, or use this narration for Last Light: ‘Mara had kept the lamp at Caer Dorran for eleven years, through winters that closed the harbor for weeks, and she had never once let it go dark, until the night the supply boat did not come.’
Coaching tips. If a line still feels long, ask whether it has two ideas; two ideas means two lines. If you cannot decide what to emphasize, ask what the character wants the listener to understand, and stress that word. Remember that the audience hears this once. And before you send any text to a generation service, check that it contains no real names, addresses or other private data, and nothing unsafe.
Line ID: ______ Speaker: ______ Scene/shot: ______ Original text: ______ Performance text (one breath group per line): 1. ______ 2. ______ 3. ______ Emphasis words: ______ Pauses (after which words; short or long): ______ Pronunciation cues (word -> respelled): ______ Private data removed? yes / no Expected length (seconds): ______
Studio task: Prepare every spoken line for generation or recording.
Prepare every spoken line in your film, narration and dialogue, for recording or generation.
Checklist before you submit:
Read one line aloud and mark where it needs a pause.
Deliverable: Voice recording script.
Submit the voice recording script. The file must use the course naming convention and include enough process evidence to show the student's decisions and revisions.
File name (course naming convention): AIF_S05-L02_voice-recording-script_v01_YOURINITIALS
| Criterion | Advanced | Proficient | Developing | Beginning |
|---|---|---|---|---|
| The work addresses the stated story or production purpose. | Every line is shaped to fit its scene and character, with lengths checked against the picture. | Lines suit the story and characters, and timing is checked. | Some lines suit the story; timing is not checked. | Lines are unprepared or disconnected from the film. |
| The student applies the required method accurately and produces usable capstone evidence. | Breath groups, emphasis, pauses and pronunciation cues are accurate, sparing and consistent across the whole script. | Breath groups, pauses, emphasis and cues are applied correctly in most lines. | Marks are present but inconsistent or overused. | Little or no performance marking. |
| The student identifies at least one quality issue and records a meaningful revision decision. | Records a specific before-and-after revision with a listening check and explains why it improved the line. | Identifies a quality issue and records a meaningful revision decision. | Notes a problem without a clear revision. | No revision recorded. |
| The work complies with consent rights attribution privacy disclosure and school policy. | Private data and unsafe content are removed from every line; the master script and subtitles stay accurate. | Private data and unsafe content are removed; policy is followed. | Some private data or risky content remains unchecked. | Sends private data or unsafe instructions to a service. |
Requirement: Remove private data and unsafe instructions before sending text to a service.
A script often contains real details by accident: a classmate’s full name, a street, a phone number, a teacher’s nickname. Remove or replace those before sending any text to a service, because you cannot fully control where the text goes. Do not write lines that give unsafe instructions, even as part of a fictional scene. Keep your original script as the master, with a separate performance copy, so credits and captions stay correct and your changes stay traceable.