Student Short Film Audio Pickup Decision Log

Decide each damaged or missing sound from evidence: identify the exact picture moment, compare the production track and existing alternatives, name the smallest replacement that could work, and assign a testable result or a hold before anyone records again.

1. Anchor the problem to picture

   Record the locked edit version, scene, shot, visible action, in and out timecodes, speaker or sound source, and the exact words or event that must be heard. Play two seconds before and after the problem so the team judges a moment in context rather than a detached waveform. Mark whether the camera shows a close mouth, a turned head, an off-screen speaker, hands touching an object, or no visible source. Those picture facts determine how tightly a replacement must synchronize. Name the editor who placed the marker and the reviewer who confirmed it on the same export.

   People2: editor and picture reviewer

   PlaceLocked-cut review workstation

   Duration3-minute context playback

   Skill checkedConnecting a sound fault to visible evidence

2. Describe the audible failure

   Listen on the approved headphones and ordinary classroom speakers, then describe only what can be observed: covered word, sudden level drop, wind burst, clothing rub, background voice, clipped syllable, missing effect, changing room sound, or timing mismatch. Do not label a track bad without naming the failure and where it begins and ends. Record whether the meaning is understandable, whether the performance remains usable, and whether the problem continues under another speaker. A reviewer repeats the affected words or event from the track; if two listeners disagree, hold the row for another playback instead of choosing a repair from memory.

   People3: editor and two listeners

   PlaceHeadphone and speaker check

   Duration5-minute two-system listen

   Skill checkedSeparating observed faults from opinions

3. Test the production track first

   Open the original production audio that belongs to the picture take and identify every recorded channel by filename, take, and microphone label if one exists. Listen to each channel alone before deciding that the location recording is unusable. Mark the cleanest complete phrase, the damaged region, the natural room sound around it, and any channel that contains a better balance. Try a reversible level match and simple edit boundary without overwriting the source. The production route passes only when the chosen section keeps the intended words or event, stays synchronized, and survives playback through the two review systems.

   People2: dialogue editor and observer

   PlaceOriginal-audio bin and timeline

   Duration7-minute channel comparison

   Skill checkedTesting source channels before replacement

4. Search alternate picture takes

   Use the slate log and shoot records to find up to three alternate takes of the same line or action. Record each filename, take number, framing difference, performance difference, background condition, and whether the full phrase exists. Place an alternate under the locked picture on a separate working track and test the beginning, middle, and end of the phrase. A wide shot or hidden mouth may tolerate a different rhythm; a close visible mouth may not. Keep the chosen alternate linked to its source take, and mark no match when none survives the synchronization and performance check rather than forcing the nearest file into the cut.

   People3: editor, slate logger, director

   PlaceAlternate-take review timeline

   Duration10-minute three-take test

   Skill checkedComparing performance and sync together

5. Match an existing wild line

   Search the wild-track log for the same speaker, words, scene, microphone setup, location, and recording period. Open the actual file and record its duration, surrounding handle, background character, and stated use; a matching filename alone is not enough. Align the line beneath the picture and test consonant timing, sentence pace, vocal energy, distance, and transition into the neighboring production sound. If the wild line covers only part of the problem, identify the exact usable words rather than approving the whole file. Pass it only when a second listener can follow the picture without being told where the replacement begins.

   People3: editor, performer, second listener

   PlaceWild-line bin and picture timeline

   Duration8-minute alignment test

   Skill checkedMatching separately recorded dialogue

6. Separate room tone from dialogue

   Decide whether the missing element is speech, continuous background, or both. Locate room tone recorded at the scene and compare microphone identity, location, time, traffic or appliance pattern, and the noise immediately before and after the edit. Use it only to bridge an exposed gap or smooth a transition; room tone cannot replace a missing word or performance. Build a short reversible test with handles on both sides, then listen for a sudden change in space, repeated noise, or an unnatural silent pocket. Mark the route as support when it improves continuity but leaves the dialogue decision unresolved.

   People2: sound editor and continuity listener

   PlaceDialogue-background test track

   Duration6-minute three-boundary loop

   Skill checkedDistinguishing ambience from speech repair

7. Choose an isolated effect

   For a door, prop, footstep, impact, switch, or other visible event, first check whether an approved isolated effect already exists in the asset or wild-track log. Record the effect filename, source description, visible action, surface or object, duration, perspective, and permission label. Align the transient to the frame where the event visibly occurs, then test its level and distance against the surrounding scene. Do not use an effect to hide unclear dialogue or imply an action the picture does not show. The effect route passes when the sound identifies the event, fits the perspective, and has a documented source the project may use.

   People3: effects editor, picture checker, asset owner

   PlaceEffects bin and frame display

   Duration7-minute event-sync test

   Skill checkedMatching sound identity to visible action

8. Define a new dialogue pickup

   If production, alternate takes, and existing wild lines fail, write the smallest dialogue unit that needs a new recording. Include the exact approved words, speaker, cue number, picture timecode, two surrounding lines, visible mouth condition, target duration, performance note, microphone reference, and room reference. Name who can approve the wording and who can schedule the performer; the log itself grants neither authority. Preserve the original production track beneath the work copy. Mark whether the pickup needs tight mouth synchronization, loose timing for an obscured face, or no visible sync for off-screen speech so the session has a measurable target.

   People4: performer, director, recordist, editor

   PlaceApproved quiet recording space

   Duration12-minute targeted pickup block

   Skill checkedWriting a bounded rerecording cue

9. Define a new effects pickup

   When an existing isolated effect cannot match, describe the specific object, surface, action, pace, perspective, and picture cue that a new recording must reproduce. State whether the frame needs one impact, a sequence, a continuous movement, or a clean tail. Choose a safe location and name the student responsible for the object and the student responsible for recording; stop if the action, access, or supervision is not approved. Capture a slate, three controlled variations, and room tone without deleting rejected takes. The decision row stays open until one take is aligned to picture and checked at normal playback level.

   People3: performer, recordist, supervisor

   PlaceApproved prop-recording area

   Duration10-minute three-variation capture

   Skill checkedTurning a picture cue into a recordable effect

10. Compare the candidate routes

    Place only the viable candidates on labeled working tracks: production channel, alternate take, wild line, isolated effect, new pickup test, or silence with room tone. Play the same marked range for the editor, director, and one listener who did not build the repair. Score each candidate as pass, fail, or hold for meaning, synchronization, performance or event identity, background continuity, and source traceability. Keep the labels visible so nobody votes for an unnamed composite. Choose the least invasive route that passes every required check, and record why a more complicated route was rejected rather than treating extra processing as automatic improvement.

    People3: editor, director, fresh listener

    PlaceLabeled comparison timeline

    Duration9-minute blind-order playback

    Skill checkedChoosing from observable route tests

11. Assign a hold when evidence is missing

    Use hold instead of inventing a result when the cut is not stable, the source file is absent, the words are unapproved, the performer is unavailable, permission is unclear, a prop action is unsafe, or the team cannot reproduce the failure. Record the missing evidence, current safe state, named owner, next review time, and what event will reopen the decision. Keep the original audio and every tested candidate; do not export a held replacement as final or fill the gap with an unrelated file. The supervisor signs only the boundary and owner, not a guessed technical or rights conclusion.

    People2: issue owner and supervisor

    PlaceShared hold list beside the edit

    Duration4-minute boundary review

    Skill checkedStopping when required evidence is absent

12. Verify and close the decision

    Replay the chosen route from the current edit export, not only inside a soloed audio track. Check the marked range on headphones and classroom speakers, then compare it with the written cue, source filename, timecode, route label, and reviewer names. Record pass, fail, or hold for sync, meaning, continuity, traceability, and export presence. A second student locates the source and rebuilds the route from the log without verbal coaching. Close the row only when the rebuilt result matches the approved playback; otherwise return it to the failed check with a named owner and next test time.

    People4: editor, director, verifier, supervisor

    PlaceExport review and source archive

    Duration10-minute rebuild-and-play test

    Skill checkedClosing a repair with reproducible evidence