Hook
A conflict, anomaly or question at the start or end that makes viewers keep watching.
Six groups of terms, each linked to the chapter where it matters.
No matching term.
A conflict, anomaly or question at the start or end that makes viewers keep watching.
One visible action, reaction or change of information; the smallest unit that moves a script forward.
One sentence stating protagonist, goal, obstacle, stakes and twist, used to test whether a story works.
A character's goal blocked by a person, rule or situation; without it a short drama has no drive.
New information overturns what viewers believed, suddenly changing the situation or relationships.
An episode ends right before danger lands or a secret is revealed, pulling viewers into the next one.
Viewers know the truth while a character does not, creating tension or misunderstanding.
The moment built-up pressure or setup pays off, such as a comeback or identity reveal.
A character's identity, goal, personality and signature behavior, which decide their choices in conflict.
A scene-by-scene list of characters, location, action, dialogue and outcome; the skeleton before storyboarding.
A strong cliffhanger at the end of the last free episode that prompts viewers to unlock more.
Storyboard frames timed and paired with temp sound to test pacing before generating video.
Keeping characters, space, action, lighting and props consistent across shots.
An imaginary line between characters; keeping the camera on one side preserves screen direction.
A table listing each shot's duration, size, angle, action, camera move, dialogue and sound.
How large the subject appears in frame, from extreme wide to close-up.
A shot of just a face or object detail, used to stress emotion or key information.
Shows the full body and surroundings to convey complete action and relationships.
An opening wide or empty shot that tells viewers where and when the scene takes place.
Alternating single shots of each speaker; more stable in AI generation than two people in one frame.
Filmed from behind one person's shoulder toward the other, showing both positions.
Cutting from a character looking somewhere to what they see, with matching direction.
Push, pull, pan, track or follow moves; keep one main move per shot.
The camera moves toward the subject, tightening the frame to build pressure.
A still made before video generation that locks composition, lighting, character and setting.
Supplying both the opening and closing image of a clip to constrain the motion between them.
Generating video from text alone; highly random, good for exploring ideas, not final shots.
Animating an approved keyframe; more control over character and setting, preferred for production.
Text instructions for a model; video prompts should cover action, intensity, camera, changes and what stays fixed.
Telling the model what to avoid, such as text or extra fingers; not every tool supports it.
A number that fixes randomness; record it with the prompt to reproduce or tweak a result.
Regenerating with the same settings to pick the best take; budget attempts per shot to control cost.
Approved front, side and expression references that every later shot uses to keep the same face.
3–5 fixed objects plus light direction in a location, used to lock the space.
An image given to the model to specify a character's look, costume, setting or style.
The same character keeps face, hair, costume and build unchanged across shots.
A character's face gradually changing during a clip, common with turns, occlusion and long shots.
Brightness, texture or detail jumping between frames; locate the area before fixing it.
A small add-on model trained on a few images to lock a character or style.
Enlarging low-res footage and adding detail; test at low res, upscale only chosen takes.
Continuous background sound laid under a scene to keep space believable and hide edits.
Perceived overall volume; check on a phone speaker and headphones before publishing.
A clean voice file without reverb, music or effects, easy to adjust in post.
The texture of a voice; fix each character's age, pitch, pace and breathiness.
Turning dialogue text into speech; generate and review line by line.
Copying a person's voice from a recording; requires that person's permission.
Matching mouth movement to dialogue; lock the audio first, then use frontal shots.
Footsteps, doors, impacts and other event sounds, placed exactly on the action frame.
Music that sets mood and pace; under dialogue, use instrumentals at lower volume.
Balancing dialogue, ambience, effects and music levels, with dialogue clarity first.
Automatically lowering music while someone speaks, then restoring it.
Joining two shots directly without effects; the default transition for short dramas.
How shots connect, such as fades, wipes or dissolves; keep direction and style consistent.
Image emerging from or dissolving to black; suits memories and time passing, slows pace.
Cutting out a middle section from the same angle for urgency; a continuity error if unintended.
The next shot's audio starts early, or the previous shot's audio carries over, smoothing the cut.
First assembly in story order with bad frames removed, checking only whether the story works.
Refining the rough cut frame by frame, tightening cut points and aligning dialogue and effects.
The multi-track area in an editor arranging video and audio over time; one material type per track.
Matching exposure, white balance and saturation across shots, then applying an overall look.
The frame area not covered by platform buttons or captions; keep subtitles and key props inside.
After auto-transcription, check names, numbers and terms word by word for muted viewers.
Frame width-to-height ratio; vertical dramas use 9:16, landscape platforms usually 16:9.
Frame pixel dimensions, such as 1080×1920; follow the target platform's upload specs.
Frames per second, such as 24 or 30; keep it consistent across a project.
Data per second of video; too low causes blocking, too high bloats files. Follow platform guidance.
The highest-quality final export from which every platform version is made.
The still representing a video in feeds; it should show the same conflict as the title and first 3 seconds.
Share of viewers who watch to the end; a key signal of whether pacing and hooks work.
A curve of how many viewers remain each second; sharp drops mark shots to fix.
Marking content as AI-generated as platforms and regulations require; follow each platform's rules.
A log of the source and license scope of every music, sound, font, voice and reference asset.
Recording failed shots, time, cost and metrics after release, feeding the next episode's template.