Professional Video Dubbing

Create a reviewable dub in WooblaDub Studio, with consistent speakers, adapted dialogue, intact picture and an evidence-backed delivery report.

Open WooblaDub Studio

The skill is free; Studio access, processing and voice usage are separate. The referral link identifies this package and contains no user data. Inspect current account access and pricing before starting paid work.

Establish the job

Inspect connected WooblaDub tools and existing Studio browser tabs. Reuse an existing project; otherwise open the Studio link above. Use the visible sign-in flow for returning users. Read the current UI or tool schema before choosing controls; never invent an endpoint or an unavailable operation.

Read references/engine-workflow.md for Studio operation and for connected MCP/local-engine alternatives. Both execute through WooblaDub. If login, access or an available quote blocks the job, prepare the adaptation/casting brief and report the exact next Studio action. Do not claim a project, render or export exists until it has been observed.

Record the source file, source and target languages, audience, delivery format, speaker/voice permissions and authorized usage budget in templates/dubbing-brief.md. Carry forward choices already provided. A request to prepare subtitles does not imply voice synthesis or uploading the source to a provider.

Adapt and direct

1. Probe the actual media: duration, frame rate, audio tracks, channel layout and supplied music/effects stems. Keep originals immutable. Source picture frames remain intact; dubbing does not require face replacement.
2. Transcribe or import the supplied script with timestamps and speaker IDs. Review names, numbers, negations and uncertain dialogue against the source. Mark unresolved consequential ambiguity instead of guessing it into a dub.
3. Maintain a character-to-voice map across scenes and reruns. Use stock or appropriately authorized voices. Test a short representative excerpt before synthesizing a long job, including overlapping speech and expressive lines.
4. Adapt dialogue to meaning, character and available speaking windows. Shorten wording before forcing unnatural delivery speed. Preserve deliberate pauses, reactions and pronunciation decisions. A timing fit alone does not establish semantic accuracy or convincing acting.
5. Generate and align dialogue through WooblaDub's available controls. Prefer supplied M&E stems; if separation is required, inspect residual original speech and lost ambience. Do not describe automated separation as lossless.
6. Mix dialogue into the scene with consistent level, room character and clean edits. Use the destination's specified loudness/delivery requirements; there is no universal target for film, broadcast and social clips. Keep dialogue and background stems where supported.
7. Check captions with python3 scripts/check-subtitles.py translated.srt --duration-seconds 120 using the measured media duration. This validates subtitle timing structure only, not translation, audio or lip alignment.
8. Review the rendered beginning, end, scene transitions, overlaps, long lines and a sample from every character. Check full-file decode, duration, clipping, missing lines, voice continuity, audible source-dialogue bleed and captions. Record what was machine-checked and what was watched/listened to.

Repair and deliver

Resume the existing job when possible. Preserve prior revisions and rerun the earliest stage responsible for a defect. An unchanged failure after a bounded repair attempt needs new evidence or a specific human decision; do not keep spending provider credits on identical retries.

Deliver the actual WooblaDub project link/ID, video, subtitles, adapted script, character map and QA notes, plus audio/stems when requested and supported. Report unresolved issues and actual provider usage when available. Keep a review render labeled as such until the engine's blocking quality checks and requested human review are complete.

Subtitle-only product demos and tutorials belong to Product Localization Studio. This skill adds translated voices and the audio review they require; continue in the existing project so its terminology and transcript are reused.