Section 5: Synthetic Voice and Audio Direction · Lesson 3 of 5 (23 of 45) · 60 minutes · Evidence: Three-take performance comparison
Say the words ‘I’m fine’ three ways: as a joke, as a lie, and as a plea for help. The words did not change; the meaning did. A voice performance lives in how something is said. When you direct a synthetic voice you do what a film director does with an actor: you decide what the character wants, then you ask for it clearly. Today you practice giving direction that works, and judging whether it did.
Production journal: write your answer below.
What must a filmmaker understand before they can direct multiple voice takes through pace emphasis tone pauses and emotional intention?
Five ingredients shape how a line sounds, and you can name and control each one. Pace is the speed of delivery: slow can feel calm, grave or tired; fast can feel eager, anxious or comic. Tone is the color of the voice: warm, cold, dry, tender, tense. Emphasis is the word that gets weight, and it changes the meaning of a line: ‘I didn’t say he took it’ means something different from ‘I didn’t say he took it’. Pauses give the listener time to feel. And intention is what the character is trying to do to the other person with the line: reassure, warn, tease, hide, beg. Intention is the most powerful ingredient, because it explains the other four.
A direction note is a short instruction a performer, human or synthetic, can follow. The best notes name an intention first and one adjustment second: ‘Mara is hiding how scared she is. Slower, steady, almost flat, and pause before the last word.’ Notes like ‘make it better’ or ‘more emotional’ fail because they give nothing to act on. Avoid vague adjectives alone; tie each one to a reason in the story.
A take is one complete attempt at a line. Professionals record several takes and compare them. To learn from the comparison, change one variable at a time. If take 2 is slower, warmer and has a different emphasis, you cannot tell which change helped. Change the pace only, listen, then change the emphasis only. Keep notes on every take, including the ones you will not use.
Listen with the picture, if you have it, and listen for the intention, not just the sound. Ask three questions. Does it fit what the character wants? Does it fit the shot’s length and mood? Does it sound like a person, or like a machine reading? A take that sounds polished can still be wrong for the scene.
Mara’s key line in Last Light is ‘The light stays on.’ The student writes three takes. Take 1: calm, steady, normal pace; it reads as routine. Take 2: slower, with weight on ‘stays’; it reads as a promise. Take 3: slower still, with a pause after ‘light’; it reads as grief. Compared against the scene, in which Tomas has just said the boat will not return, take 2 is chosen: Mara reassures him, and herself. The student records that take 3 was beautiful but belonged to a later scene.
Voice directors and animation teams use a repeatable routine when they want a line to land. Study it as if you were sitting in the control room, and pause at each decision to ask who is making the call: the person, or the tool.
The director starts with the story, not the sound: what has just happened, what does the character want? Only then do they write a direction note, and they test it by generating or recording a first take. They listen, compare it with the intention they hold in their head, and decide what single thing to change. Then they ask for another take. After three or four attempts they choose, and they write down why.
The tool can produce a take in seconds, which tempts people to generate dozens. Strong directors resist that. A small number of well-planned takes teaches more than a pile of random ones, and the written reason for the final choice is what lets a teammate understand the decision later. Notice also that the tool follows the words of a note, not its spirit; the human must judge whether the delivery is believable.
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 |
|---|---|---|---|
| Define the intention | Reads the scene and writes what the character wants | Nothing; this is a story decision | Whether the intention fits the whole film and not just the line |
| Write the direction note | Names the intention and one change in pace, tone, emphasis or pause | Interprets the note into a delivery | Whether the note was specific enough to act on |
| Generate or record a take | Requests a take and labels it | Produces the audio | Whether it sounds like a person and fits the shot’s length |
| Compare and decide | Changes one variable, compares, and records the reason for the choice | Nothing | Whether the chosen take serves the scene, not just sounds impressive |
Studio task: Give precise direction and compare how one variable changes meaning.
Work with a partner. Choose one short line, five to twelve words, from your film. You will explore how one variable changes meaning. If you need a line, use Tomas’s: ‘I can still make it back before dark.’
Coaching tips. If all three takes sound alike, your change was too small; exaggerate it, then pull it back. If a take sounds unnatural, check whether the line has too many breath groups, or marks that cut against each other. Record every note exactly as you gave it. And avoid asking for an accent, or a manner of speaking tied to a group of people, to make a character feel a certain way; ask for the emotion and the pace instead.
Line: ______ Speaker: ______ What the speaker wants: ______ Take | Direction note given (exact words) | Variable changed | What it seems to mean | Fits the scene? (1-3) 1 | ______ | (baseline) | ______ | ___ 2 | ______ | ______ | ______ | ___ 3 | ______ | ______ | ______ | ___ Chosen take and the evidence: ______ One thing I would try next: ______
Studio task: Produce three takes of a key line and select the strongest with evidence.
Direct a key line from your film, one the story depends on, and prove your choice with evidence.
Checklist before you submit:
Write one direction note that a performer or system can follow.
Deliverable: Three-take performance comparison.
Submit the three-take performance comparison. 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-L03_three-take-performance-comparison_v01_YOURINITIALS
| Criterion | Advanced | Proficient | Developing | Beginning |
|---|---|---|---|---|
| The work addresses the stated story or production purpose. | The line, intention and takes are tied tightly to the scene, and the chosen take clearly serves the story. | The takes serve the scene and the intention is stated clearly. | The takes loosely relate to the scene. | The takes do not connect to the story. |
| The student applies the required method accurately and produces usable capstone evidence. | Direction notes are specific, exactly recorded and each take changes a single named variable; scoring is consistent. | Notes are clear, takes are labeled and mostly change one variable. | Notes are vague or several variables change at once. | Notes or takes are missing. |
| The student identifies at least one quality issue and records a meaningful revision decision. | Identifies a precise weakness in the chosen take and records a concrete next revision. | Identifies a quality issue and records a meaningful revision decision. | Notes a flaw without a clear next step. | No issue or revision recorded. |
| The work complies with consent rights attribution privacy disclosure and school policy. | Uses only the authorized voice, avoids stereotyped delivery, and documents how the choice respects the casting record. | Uses the authorized voice and avoids caricature. | Uses the authorized voice but delivery leans on a stereotype in places. | Uses an unauthorized voice or relies on caricatured delivery. |
Requirement: Avoid caricatured accents or identity-based vocal stereotypes.
Delivery choices carry meaning about people. Do not ask a voice to sound ‘foreign’, ‘ghetto’, ‘dumb’, ‘old’ or otherwise use an accent or manner as shorthand for a type of person. Ask instead for the emotion, the pace and the intention you want, and let the character’s words and actions do the rest. If a role calls for a particular regional or cultural voice, find an authorized voice or a consenting speaker from that community, and ask them how the line should sound. Keep using only the voice you documented in lesson 5.1.