跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
视频与音频

dialogue-editing-adr

Provider-independent dialogue editing and ADR direction for AI agents producing generated videos, films, ads, avatar clips, explainers, localization/dubbing, podcasts, recuts, and social content. Use for dialogue prep, repair, room tone, take comping, sync, ADR cueing, dubbing direction, pronunciation, de-essing, plosives, mouth clicks, voice continuity, synthetic voice consent boundaries, captions/transcripts handoff, mix handoff, loudness, intelligibility, accessibility, and QA.

calesthio · 已有 AI 审核记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add calesthio/generative-media-skills --skill dialogue-editing-adr

技能说明


name: dialogue-editing-adr description: "Provider-independent dialogue editing and ADR direction for AI agents producing generated videos, films, ads, avatar clips, explainers, localization/dubbing, podcasts, recuts, and social content. Use for dialogue prep, repair, room tone, take comping, sync, ADR cueing, dubbing direction, pronunciation, de-essing, plosives, mouth clicks, voice continuity, synthetic voice consent boundaries, captions/transcripts handoff, mix handoff, loudness, intelligibility, accessibility, and QA."

Dialogue editing and ADR direction

Use this skill when spoken audio must carry story, brand, instruction, character, or accessibility. Treat dialogue as a performance and an editorial contract, not just a waveform. The job is to preserve intelligibility, intent, continuity, and rights while leaving the mixer, captioner, localization team, or downstream video agent with unambiguous material.

Evidence posture

Documented facts below come from professional standards, official platform guidance, official tool documentation, or rights organizations. Volatile platform/tool/legal facts were checked on 2026-07-10. Empirical observations are limited to listening/QC behaviors an agent should perform on the actual project; no universal audio-processing result is assumed without auditioning. Production heuristics are practical defaults to test against picture and delivery specs, not laws of audio.

First pass: decide what kind of dialogue problem this is

Classify each line or section before touching it:

  • production-edit: usable source audio needs cleanup, smoothing, fill, conform, or comping.
  • performance-fix: the words are technically usable but the read, emotion, pronunciation, emphasis, or timing fails the scene.
  • ADR-replace: production sound is not intelligible, has unrecoverable noise/clipping, contains changed script, needs a legal/brand correction, or must match a new cut.
  • localization-dub: new language or regional version must preserve meaning, performance, timing, and picture sync.
  • VO-style-dub: original voice remains partly perceptible or intentionally dipped under a localized voice-over.
  • synthetic-rerender: a licensed synthetic voice or generic TTS voice must be regenerated, edited, or directed.
  • caption/transcript-handoff: text assets must represent the final heard audio accurately and accessibly.

Do not choose ADR or synthetic replacement merely because repair is inconvenient. First determine whether the original performance can be made intelligible without audible artifacts, whether the actor/voice rights permit replacement, and whether the replacement will match picture and continuity.

Documented facts to respect

  • ITU-R BS.1770-5 specifies algorithms for measuring audio programme loudness and true-peak level; it is a measurement method, not a creative mix recipe. Source: ITU-R BS.1770-5, verified 2026-07-10.
  • EBU R 128 version 5.0 recommends programme loudness normalization at -23 LUFS and uses loudness range and true peak descriptors. Source: EBU R 128, verified 2026-07-10.
  • FCC caption quality guidance for U.S. television uses four quality pillars: accuracy, synchronicity, completeness, and placement. The FCC guide says prerecorded programming is expected to have full compliance because offline review is possible. Source: FCC DA 21-469, verified 2026-07-10.
  • W3C WCAG 2.2 SC 1.2.2 requires captions for prerecorded audio in synchronized media, with captions covering dialogue, speaker identification, and meaningful non-speech information. Source: W3C WAI Understanding SC 1.2.2, verified 2026-07-10.
  • Netflix timed-text general requirements use 5/6 second minimum and 7 second maximum subtitle-event durations, a two-line maximum, and positioning that avoids overlap with relevant on-screen text. These are Netflix requirements, not universal specs. Source: Netflix Partner Help Center, verified 2026-07-10.
  • Netflix dubbing guidance emphasizes that lip sync should be judged primarily against picture, not original-version waveforms; key moments should account for mouth flaps, labials, and start/end mouth shapes. Source: Netflix Dubbing Guiding Principles, verified 2026-07-10.
  • Netflix dubbing/mixing guidance asks for dialogue to be embedded naturally in the music-and-effects bed, with realistic level, EQ, reverbs, perspective, futzes, and continuity across languages. Source: Netflix Mixing Style Guide for Dubbed Content, verified 2026-07-10.
  • Netflix nonfiction English dubbing guidance recommends recording clean dialogue without dynamic processing except a low-cut no higher than 100 Hz, removing tongue clicks/saliva unless performance-driven, using fades on dialogue clips, editing against the audio reference, and ending dubbed dialogue at the same time as the original version where that workflow applies. Source: Netflix Partner Help Center, verified 2026-07-10.
  • iZotope RX documentation describes Mouth De-click as detecting and reducing mouth noises such as clicks and lip smacks, while warning that high sensitivity can affect plosives or damage the original signal. RX De-plosive is documented as identifying and reducing plosives while preserving speech fundamentals and harmonics, but higher strength can harm useful speech. These are tool capabilities, not guarantees. Source: iZotope RX docs, verified 2026-07-10.
  • The U.S. Copyright Office's AI report page states that Part 1, published July 31, 2024, addresses digital replicas, while SAG-AFTRA's AI framework states clear consent, fair compensation, and control over performances as guardrails. NAVA recommends performer contracts include consent, limits on AI/synthetic use and training, opt-out or term limits, payment, exclusivity clarity, and safe storage/tracking. These are not a complete legal opinion. Sources: U.S. Copyright Office, SAG-AFTRA, NAVA, verified 2026-07-10.

Empirical checks to perform on the actual media

Listen, do not infer from meters alone. Fo

阅读完整来源 ↗