Short definitions for the jargon this tutorial uses. In the lessons, dotted underlines are tooltips—hover or focus them for a quick reminder, or jump here for the full entry.
Agent harness
Also: AI coding environment, agent IDE
An AI coding environment that can help read, edit, and run a project’s files. This can support a code-based production workflow, but the course activities need only a browser and a place to write notes.
amix
An FFmpeg filter for combining audio tracks. Ellie used it to mix dialogue and music. The general editing task is to keep the words clear while balancing the other sounds.
BGM
Background music
Music that sits under dialogue. Good BGM supports mood without competing with words—especially important for kids’ narration and phone speakers.
CC BY
Creative Commons Attribution
Creative Commons Attribution, a license with attribution requirements. Ellie’s music uses CC BY 4.0. The credits page includes the exact music attribution, source, and link to the license terms.
Composition
A video arrangement with a defined size, duration, and frame rate. In Remotion it is represented by a React component. A full story and a teaser can use different arrangements of the same material.
Contact sheet
A single image that shows several shots in a grid (Ellie uses 2×2). Useful for review before you crop each panel into production frames.
Cursor
An AI coding editor, mentioned as background to code-based production workflows. It is not required for the lessons or final assignment.
Antigravity
An AI coding environment mentioned in the Ellie production context. The planning exercises in this course do not depend on it.
FFmpeg
A toolkit for processing audio and video. Ellie used it for tasks such as trimming, fading, and mixing music beneath dialogue. You can plan those operations without running commands.
Frame (video)
One still image in a sequence. At 30 fps, one second of video is 30 frames. Remotion schedules pictures and captions in frame units.
FPS
Frames per second
The number of video frames shown per second. At the 30 fps rate used in the course example, four seconds occupy 120 frames.
Ken Burns
A slow zoom or pan across a still photo so it feels alive. Named after the documentary filmmaker. Ellie uses gentle scale changes on reel stills instead of true video footage.
Kokoro
The neural text-to-speech system used for Ellie’s generated voices. The course focuses on choosing delivery and checking the resulting speech; installation is not part of the activities.
Phonemizer / espeak-ng
A tool that converts written text into phonemes, the sounds of speech. This can be one stage in a text-to-speech system. espeak-ng was part of Ellie’s speech toolchain.
Pipeline
A sequence of stages with clear outputs and handoffs. In this course: story, speech, sound, pictures, assembly, and review. A change at one stage can require checking later stages again.
Pitch shift
Raising or lowering the pitch of a recording. Some methods preserve duration; simply changing playback speed also changes timing. Pitch alone does not determine the apparent age or emotion of a voice.
Reel frame
A vertical picture prepared for a short video. Ellie uses a 1080×1920 canvas, a 9:16 aspect ratio. Your storyboard can use tall boxes in the same proportion.
Remotion
A framework for making videos with React components and a frame-based timeline. Lesson 06 explains the scheduling idea using a table and simple arithmetic; no application setup is needed for that exercise.
Semitone
A musical pitch step, such as the distance between adjacent piano keys. Twelve semitones make an octave. A pitch adjustment changes how high or low a voice sounds.
Shorts / Reels
Short-form video formats such as YouTube Shorts and Instagram Reels. This course plans a vertical 9:16 video; publishing to a platform is not required for the assignment.
Style lock
A reference picture and written rules that guide consistent character design and art style. The image lesson includes visual examples and a sample character brief. Each new picture still needs review.
Subtitle cue
A caption segment with a start and end time. It should match the spoken words and remain visible long enough to read.
Timeline
A plan of what appears or plays at each time. The worked example lists pictures and spoken lines across four intervals totalling 20 seconds.
TTS
Text-to-speech
Software that reads text aloud. Modern neural TTS (like Kokoro) sounds far more natural than old robotic systems, but still needs human listening for pacing and character.
Virtual environment (.venv)
An isolated environment for a Python project’s packages. It is background vocabulary for code-based workflows, not something you need to create for this course.
WAV
An audio file format commonly used to store uncompressed speech during editing. Ellie used separate recordings for spoken beats so individual lines could be revised.
Word reveal
Showing caption words gradually. Spacing the words evenly across a line’s duration is an approximation; exact word timing requires word-level timestamps. Full-line captions are another option.
Tip: in lessons, terms with a dotted underline are tooltips. Hover or Tab to them for a quick definition, or follow the link to this glossary.