Supplemental Terms for Dubbing Services

These Supplemental Terms for Dubbing Services apply to the extent that Customer uses dubbing Services.

  1. Service Description. Verbit offers a suite of dubbing solutions delivered via human talent and/or artificial intelligence technologies.
    Human-based dubbing Services are provided using human talent and voice casting. Supplier shall cast voice talent using commercially reasonable professional judgment to match the original performances as closely as practicable, taking into account target language requirements, accent and cultural sensitivity, age range and gender, vocal texture, energy and emotional range, character consistency, and acting and performance skills. Customer shall be provided reasonable opportunity to approve or reject voice talent prior to delivery of Services. Where Customer declines sign off rights, subsequent dissatisfaction with the voice casting selection shall not be a valid basis for rejection.
    AI-based dubbing Services employ AI-generated voices with human quality review and post-production, and may be delivered either through (i) voice cloning, where the voice is synthesized from the source media or Customer-provided samples, or (ii) a voice bank, where the voice is selected from a library of pre-recorded, licensed voices.
  2. Review and Acceptance. Customer must provide written rejection of Output within 14 calendar days after delivery (or re-delivery), with specific and detailed explanation for the rejection. Rejection must be based on nonconformity to the Technical Requirements or Quality Requirements (together, “Dubbing Requirements”) below. Failure to provide a timely and detailed written rejection shall be deemed acceptance.

    Technical Requirements

    1. Speakers are correctly allocated.
    2. All dialogue that is reasonably understandable has been dubbed.
    3. Dialogue is aligned to onscreen lip movement (not applicable for voiceover) and facial expressions of the on-screen actor.
    4. Timing reasonably aligns with original speech and pacing including pauses, overlaps etc.
    5. Overlaps with music or key sound effects avoided, except to the extent original voice overlaps in which case dialogue will be mixed as seamlessly as is reasonably possible with music and sound effects. Supplier recommends that Customer provide a music & effects file where available, which is expected to improve quality.
    6. No introduction of background noise, echo or static.
    7. No overlaps or dropouts; smooth transitions between speakers.
    8. No introduction of clipping, distortion, or mic pops.
    9. Reasonable minimization of breath sounds, mouth clicks, and plosives.
    10. Export in correct format as per Customer specifications (sample rate, framerate, bit depth, codec, loudness).
    11. Filename follows Customer’s convention.
    12. Dialogue should be clear and audible, with consistent volume and clarity across all episodes. Overall dubbed dialogue and music & effects balance meets technical specifications supplied by the Customer and in required format.
    13. Original media frame rate, timecoding, and codecs matched in deliverable.
    14. (AI dubbing only) Where Customer has opted for cloned voice, voice characteristics of the cloned voice shall be substantially identical to the voice characteristics of the applicable source file or samples. Where Customer has opted for voice bank, there should not be substantial miscasting, e.g., apparently male voice for female, apparently adult voice for child.

    Quality Requirements

    1. The dub must use the correct target language, dialect, and pronunciation as specified by Customer. Local idioms and cultural references should be adapted appropriately with appropriate emotion, tone, and energy to match the original performance.
    2. All dubbed dialogue should sound natural, clear and coherent. Consistency must be maintained across all files and episodes.
    3. No incorrect translations or literal translation errors (i.e., the translation should be semantically equivalent to the original source language).
    4. With respect to both translations and pronunciation, apply key names and phrases where guidance is provided by Customer prior to submission of job. Otherwise, pronunciations and word usage applied based on best judgment and/or customary usage where applicable; and retroactive corrections will incur costs.
    5. Censorship compliance.

    Supplier shall review rejections with reasonable promptness and, as Customer’s sole and exclusive remedy, corrects nonconformities to the Dubbing Requirements. Corrections based on rejections other than for nonconformities to the Dubbing Requirements, or rejections not in accordance with the timing requirements above, shall be made in Supplier’s sole discretion and may be subject to additional charges.

  3. Storage. Following accepted delivery of the Output, Supplier will make commercially reasonable efforts to maintain Output for (a) 30 days following delivery of Output in SD or HD resolution; and (b) 10 days following delivery of Output in UHD (4K or 8K) resolution.
  4. Voice Cloning. Where Customer has selected voice cloning for the Output, Customer authorizes Supplier to analyze, process, and create synthetic or cloned voice models that reproduce the vocal characteristics of individuals appearing in the Files (“Voice Clones”) solely as necessary to provide the Services to Customer. Customer represents and warrants that it has obtained all rights, permissions, and consents required under applicable law to permit the foregoing, including consent from each individual whose voice is cloned. Customer shall defend, indemnify, and hold harmless Supplier from any Losses arising out of or relating to any claim that Supplier’s creation or use of the Voice Clones as authorized above infringes, misappropriates, or otherwise violates any third party rights.
  5. Restrictions on Use of Output. Customer shall not, and shall not authorize any third party to, use the Output to train, improve, fine-tune, or benchmark any artificial intelligence or machine learning model.