AI Media
A practical guide to transcripts, speaker diarization, timecodes, translation adaptation, glossaries, voice matching, phonemes, visemes, lip sync, background preservation, subtitles, and captions.
12 matching terms
Source analysis
Transcript
A time-aware written record of the spoken content in the source audio or video.
Use it as the editable source for translation, speaker assignment, captions, and quality review.
transcribe every spoken line with punctuation and retain the original timecodesSource analysis
Speaker diarization
Identifies when different people speak and assigns each segment to a speaker.
Use it before translating or dubbing conversations with multiple speakers.
separate the dialogue into Speaker A and Speaker B, preserve interruptions and overlapsSource analysis
Timecode
A timestamp that identifies the start and end position of a line or event.
Use it to keep translated lines, captions, edits, and review notes aligned.
keep each translated line within its original in and out timecodesTranslation
Translation adaptation
Rewrites a literal translation so it preserves meaning while fitting audience, culture, tone, and available speaking time.
Use it when direct translation sounds unnatural or is too long for the scene.
adapt into natural Korean dialogue, preserve the joke and intent, fit the original line durationTranslation
Pronunciation glossary
A shared list of approved translations and pronunciations for names, brands, places, and specialist terms.
Use it to keep terminology consistent across speakers, episodes, and languages.
apply the approved glossary: Acme stays untranslated and is pronounced 'ACK-mee'Voice
Voice match
Selects or designs a target voice with qualities compatible with the source speaker.
Use it to preserve character age, energy, vocal weight, and narrative role without copying identity unnecessarily.
match the source speaker's calm energy and vocal weight, use an authorized target voiceVoice
Voice continuity
Keeps a dubbed speaker's identity, performance level, accent, and recording character stable across scenes.
Use it for episodic work, pickups, split sessions, and regenerated lines.
use the same approved voice profile and room character for every line from Speaker ASpeech and lip sync
Phoneme
A smallest contrastive unit of spoken sound used to analyze pronunciation and timing.
Use it to diagnose pronunciation errors and align generated speech more precisely.
align the final consonant phoneme with the source word ending, keep natural coarticulationSpeech and lip sync
Viseme
A visible mouth shape associated with one or more speech sounds.
Use it when adjusting facial animation or checking whether mouth shapes support the dubbed speech.
align mouth visemes to the dubbed phrase, preserve natural jaw and lip movementSpeech and lip sync
Lip sync
Aligns visible mouth movement with the timing and sound pattern of replacement speech.
Use it after translation and voice generation, especially for close-up dialogue.
Strong correction can distort teeth, lips, or facial identity; review close-up frames.
match lip movement to the Korean dub, preserve facial identity and natural expressionsAudio mix
Background preservation
Retains music, ambience, and sound effects while replacing or translating dialogue.
Use it when dialogue is not available as an isolated production stem.
replace only the dialogue, preserve the original music, ambience, and sound effectsText output
Subtitle and caption
Timed text that represents translated dialogue or includes dialogue plus meaningful non-speech audio.
Use subtitles for language access and captions when sound cues and speaker information are also required.
create Korean captions with speaker labels, sound cues, readable line breaks, and preserved timecodes