Glossary of common terms 📖

Agentic Dubbing: Dubbing a video by chatting with an AI agent: you describe what you want and the agent sets up and runs the project. Learn more

Audio Separation: Splitting an audio file into its parts, for example the voice and the background music or noise. Learn more

Audiobook: The tool that turns a written manuscript into a narrated audiobook (formerly called Stories). Learn more

Background audio: Everything in your video except the speech: music, sound effects and ambient noise. CAMB.AI keeps it and mixes it under the new voices.

Broadcast: Live stream dubbing: translating and voicing a live event into other languages in real time. Learn more

Credits: The units used to pay for generation (transcription, translation, voiceover and so on). See your usage in Insights. Learn more

Dialogue: One line of speech in the editor, with its speaker, start and end time, and text.

Emotion tags: Cues in square brackets, such as [happy] or [nervous], that tell MARS8-Instruct how to deliver a line. Learn more

Final mix: The finished output: the new voices mixed with the background audio, ready to export. Learn more

Gaps: Silent parts where the voiceover ends before the speaker does, or between dialogues. Learn more

Instant Dubbing: The classic dubbing workflow where you set the options yourself and review each step in the editor. Learn more

Keywords: Lists of names, brands and terms that tell CAMB.AI how to transcribe, translate and pronounce them. Learn more

MARS models: CAMB.AI’s speech models: MARS8.1-Beta (default), MARS8-Pro, MARS8-Instruct, MARS-X Instruct (Beta) and MARS7. Learn more

Merge / Split: Combining two dialogues into one, or dividing one dialogue into two. Learn more

Muted: A dialogue that has no voiceover yet. Learn more

Number of Speakers: How many people speak in your video. Setting it helps CAMB.AI separate their voices. Learn more

Re-generate: Creating a dialogue’s voiceover again after you change its text, voice or settings. Uses credits. Learn more

Re-translate / Spot re-translate: Translating the whole language again, or just one dialogue. Learn more

Source language: The language spoken in your original video or written in your original file.

Speaker: A person (or voice) in your video. Each speaker has a name, gender and voice. Learn more

Speaker Settings: Where you set each speaker’s name, voice and audio settings. Learn more

Speed-ups: Dialogues whose voiceover was sped up to fit the time available. Shown with a speed badge such as 1.26x. Learn more

Stale: A dialogue you edited after its voiceover was generated. Re-generate it to update the audio. Learn more

Subtitles mode: An editor mode for working only on subtitles in a dubbing project. Learn more

Target language(s): The language(s) you translate or dub into.

Transcript: The written text of what is said in your video, split into dialogues. Learn more

Voice cloning: Creating a digital copy of a real voice from a recording. Only clone a voice with the person’s permission. Learn more

Voice Library: Where you find, save and create AI voices. Learn more

Voice reference: The voice sample CAMB.AI uses to generate a speaker’s voice: the original speaker from your video, or a voice from the Voice Library. Learn more

Workspace: A shared space where your team works on projects together. Learn more