What problem does it solve? Audio tracks placed in a HyperFrames composition often fight each other — a music bed drowns out a voiceover, a voice sounds boomy or harsh, or levels drift across a take. This Skill mixes already-placed audio by writing effect chains, volume automation, and voiceover carve settings directly onto the composition's audio and video elements. ## Core Features & Use Cases - Effect Chains: Build serial data-fx-chain JSON on audio/video elements using filters, compressor, limiter, gate, saturation, delay, reverb, chorus, phaser, and bitcrush, with the same graph driving preview and render. - Voiceover Carve: Run scripts/carve.mjs to analyze voice tracks and automatically cut the bands they occupy out of the music bed, with dynamic ducking envelopes that follow the speech. - Automation Envelopes: Write data-automation lanes on track volume or automatable effect parameters using clip-local breakpoints. - Diagnosis & Presets: Measure files you cannot hear with ffmpeg-based recipes, then apply named jobs (Tame Boominess, Reduce Mud, Add Clarity) or presets like voice-clean. - Use Case: Given a composition with narration over background music, run the carve script to make the voice intelligible while the bed keeps its low end and top, then add a limiter as the final ceiling. ## Quick Start Run the carve script on my composition index.html so the music bed makes room for the narration track.
