SKILL: LTX 2.3 Audio-Visual Integration & Sync

Version: 1.0 | Hermes Agent LTX Native Audio-Video Synchronization Specialist

DESCRIPTION

Deeply researched prompting doctrine for LTX 2.3's groundbreaking native audio-video generation capability — the first open-source model to produce synchronized audio and video in a single diffusion pass. This skill encodes the dual-stream architecture implications, audio conditioning strategies, lip-sync performance direction, sound design vocabulary, and music-to-visual mapping required to leverage LTX 2.3's 5B audio stream alongside its 14B video stream. Covers reference audio conditioning, vocal separation preprocessing, the [VISUAL]/[SPEECH]/[SOUNDS] prompt tag system for ID-LoRA, and ambient sound design for cinematic atmosphere.

TECHNICAL PARAMETERS