Section 6: Lip Sync and Performance · Lesson 2 of 5 (27 of 45) · 60 minutes · Evidence: Baseline lip sync take
The first time you press generate, you are not making your final shot. You are asking a question: what does this tool do with my face and my voice? A good filmmaker treats the first answer as information, saves it, and judges it honestly. Today you create that first answer and learn to score it without flinching.
Production journal: write your answer below.
What must a filmmaker understand before they can create a baseline performance and evaluate timing face stability eyes and emotion?
A baseline is the first complete result from your inputs with no tricks and no fixes. Its job is to be a fair starting point you can measure every later version against. If you tweak settings before you ever see the plain result, you will not know what helped. Generate once with sensible defaults, save it, and only then start changing things.
Watch it three times. First, with sound off, looking only at the face. Second, with sound on and your eyes on the mouth. Third, at the size it will appear in your film. Score each of the four features from 1 to 4 and write one specific note, such as “jaw widens too much on the word never“. A score without a note is a feeling, and feelings cannot be fixed.
Every generation is a version. Name each one so you can find it later: project, shot, version number, and a few words about what changed, for example shot12_v01_baseline. Save the input files and any setting you changed next to each version. If you cannot say how a take was made, you cannot repeat it.
Tomas, the supply-boat pilot, says, “Storm’s coming in tonight.” The student generates the baseline. Timing scores 3: the mouth closes on “storm’s” and opens on “coming” almost on time. Stability scores 2: his red knit cap slides slightly on the sentence’s last words. Eyes score 1: he never blinks across eight seconds. Emotion scores 2: the line is a warning, but his face is mild. The student writes those four notes, saves the file as shot03_v01_baseline, and does not hide the eye problem. That note will guide the next version.
Animators and video artists treat a first test as a diagnostic, not a product. Study how they handle that first run, and at each step decide whether the person or the tool carries the decision.
The creator selects a face and a clip, runs the tool with its default options, and plays the result straight through before touching anything. Some tools ask where the face is in the picture or which voice belongs to which face; the creator confirms those choices because a wrong guess here ruins the output. The output is then judged against a short list of features, and the creator keeps the file even when it is unusable, because a failed take explains what to change.
The habit worth stealing is restraint. A hurried creator regenerates immediately and loses track of what changed. A careful one writes the observation, changes one thing, and generates again. That discipline turns a lucky result into a repeatable method.
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 |
|---|---|---|---|
| Confirm face and voice selection | Marks which face speaks and attaches the right audio | May detect faces automatically | That the tool picked the intended face, especially in a group image |
| Generate with defaults | Runs one clean take and saves it as version one | Produces the animation | Whether the settings are school-approved and the output saved with its inputs |
| Review the take | Watches with sound off, then on, then at final size | Nothing; viewing is human | Timing, stability, eyes, and emotion, each against the scene’s purpose |
| Record scores and notes | Writes a score and a specific note per feature | Nothing | That notes name what is wrong precisely enough to act on |
Studio task: Apply the class performance checklist to a demonstration take.
Your class uses one demonstration take (supplied by your teacher or produced together). Everyone scores the same clip so you can compare judgments.
Coaching tips. Point to a moment, not a general impression: “at four seconds the cheek bulges” beats “it looks weird.” If you give a perfect score, say what would make it worse. If you give the lowest score, say what would improve it. Finish by agreeing with your group on the single weakest feature, since that is where your own first revision will probably begin.
Version name: ______ Date: ______ Inputs used (face file / audio file): ______ Settings changed from default: none / ______ Feature | Score 1-4 | Specific note (moment + what you saw) Mouth timing | | Face stability | | Eye movement | | Emotion | | Strongest feature: ______ Weakest feature: ______
Studio task: Generate the first project take and score it without hiding defects.
Using the inputs from your lip sync input package, generate and judge your baseline.
Checklist before you submit:
Record the strongest and weakest feature of the take.
Deliverable: Baseline lip sync take.
Submit the baseline lip sync take. 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_S06-L02_baseline-lip-sync-take_v01_YOURINITIALS
| Criterion | Advanced | Proficient | Developing | Beginning |
|---|---|---|---|---|
| The work addresses the stated story or production purpose. | The take is judged against the scene’s purpose, and the notes connect each defect to its effect on the story. | The take is judged against the scene’s purpose. | The take is scored but only loosely connected to the scene. | No judgment against any story purpose. |
| The student applies the required method accurately and produces usable capstone evidence. | Clean default baseline, saved with inputs and a precise name; all four features scored with moment-specific notes. | Baseline saved and named; all four features scored with notes. | Baseline saved; some features unscored or notes vague. | No usable baseline or no scoring. |
| The student identifies at least one quality issue and records a meaningful revision decision. | Names the weakest feature honestly, picks one change, and explains why it comes first. | Identifies a quality issue and records a meaningful revision decision. | Notes a weakness without a clear next step. | No revision decision recorded. |
| The work complies with consent rights attribution privacy disclosure and school policy. | Authorization of face and voice documented; version records complete; defects not concealed. | Authorization documented; version records kept. | Authorization or version records are incomplete. | Unauthorized face or voice, or defects deliberately hidden. |
Requirement: Do not publish an unreviewed synthetic performance.
A baseline take is a draft, and a draft is not ready for an audience. Do not post or screen a synthetic performance until you have reviewed it closely and your teacher has approved it. Keep the inputs, settings, and tool record for every version, because you may need to show how a performance was made and that its face and voice were authorized.