Skip to walkthrough
A hand holds an envelope in the revised AI-generated storyboard, where the original showed a key.

CueTether

Picture-change review for audio description

When picture changes, audio description needs another look.

Watch the walkthrough

The review in
52 seconds.

Load a clip, then use its playback controls. Each video is about 3 MB and loads fully before playback.

Silent walkthrough. A text summary follows.

Load the clip to enable playback and seeking.

The silent walkthrough introduces four manual review examples: a new action without a cue, a timing shift, a changed object, and a removed action. The review cases below provide the findings in text with frames and human decision fields.

What changed?

Four review examples from three picture edits. Choose a finding to inspect its frames, cue text and timings.

Choose a review finding

New action

No existing cue

An inserted storyboard image depicts an envelope above a tray. No original cue covers it.

Original reference frame: a reception desk with no envelope above a tray.
Original: Reference frame at 00:12
Revised storyboard frame: an envelope above a tray.
Revised: 00:10 to 00:14

The insertion boundary is original 00:10. The original reference frame at 00:12 is shown for context, not as a matched replacement for the inserted shot.

Human decision

Decide whether the new action merits audio description, then inspect available speech space in a real programme.

Timing shift

C02: "He crosses the hall."

The storyboard picture is retained. The insertion moves the cue four seconds later.

Original storyboard frame: a courier crossing the hall at 00:20.
Original: 00:18 to 00:22
The same courier storyboard frame in the revised cut at 00:24.
Revised: 00:22 to 00:26

The same picture appears at a different time. The wording may still fit.

Human decision

Review the proposed retiming and recorded-audio placement. No script or recording has been automatically changed.

Changed object

C03: "She picks up a key."

The revised storyboard shows an envelope where the original picture showed a key. The cue text no longer describes that object.

Original storyboard frame: a hand holding a brass key.
Original: 00:30 to 00:34
Matched revised storyboard frame: the hand holding an envelope instead.
Revised: 00:34 to 00:38

This is a hand-prepared still-image replacement, not evidence of motion understanding.

Human decision

Review the wording, timing and recording required for the revised picture.

Removed action

C04: "He closes the door."

The four-second door-closing storyboard shot has been removed. The existing cue needs review.

Original storyboard frame: the courier at the blue door in the door-closing shot.
Original: 01:00 to 01:04
Revised reference frame after the deleted shot: an empty reception desk.
Revised: Removed at 01:04

The revised reference frame shows the material after the deletion. It is not a replacement for the removed action.

Human decision

Decide whether to remove or revise the cue and its recording. Check the surrounding sequence before making the change.

The insertion shifts retained material four seconds later until the deletion restores the original timing. Wording, retiming, recording, dialogue space, and mixing remain human decisions.

Compare the cuts.

Both timelines are two minutes at 24 frames per second. They are silent sequences of storyboard stills. Their equal length does not mean the cues still line up.

Original cut

Load the clip to enable playback and seeking.

The reference timeline for six authored AD cues.

Revised cut

Load the clip to enable playback and seeking.

A four-second insertion, an object replacement, and a four-second deletion.

These are hand-prepared edits. Inspect the case viewer for the exact changes; the clips do not demonstrate automatic detection.

Review the evidence.

A frame. A cue.
A decision for a person.

The examples above pair frames with cue text, original and revised timings, and a decision for a human reviewer. This report excerpt shows how those findings can be presented for review.

First page of the existing CueTether report, pairing original and revised frames with cue timings and human decision fields.
A page from the concept review report. Explore the four findings above.