DOCS/CONTENT & AUDIO CRAFT/Audio Mastering & -14 LUFS

Audio Mastering & -14 LUFS

Voice synthesis, BGM sidechain ducking, SFX placement, and broadcast loudness.

Shotrunner mixes neural speech, ducked background music, and impact sound effects into a broadcast-compliant audio master calibrated to −14.0 LUFS.

3-Track Audio Architecture

The FFmpeg audio filtergraph dynamically mixes three simultaneous audio streams:

•Voice Narration (Center): Spoken audio synthesized by Fish Audio S2.1 Pro (audio.wav).
•Background Music Bed (Stereo): Mood-tagged BGM track configured in plan.yaml with automatic ducking.
•Impact SFX Accents: Synchronized sub drops, risers, and whooshes on scene cuts.

Sidechain BGM Ducking

When the narrator speaks, background music ducks automatically to preserve speech intelligibility. During brief pauses between sentences, the music volume subtly rises to sustain pacing:

plan.yaml

1bgm:

2 track: "shared/background_music/bgm_suspense_covert_affair_investigation_v1.mp3"

3 volume: 0.08

5scenes:

6 - start: 0.00

7 end: 3.45

8 image: "../../images/hero/hero-confident-cu-c-t.jpg"

9 shot_type: "CU"

10 sfx: "sub_drop" # Fires shared/sfx/sub_drop.wav on cut

−14.0 LUFS Mastering & --loudfix

YouTube Shorts applies strict volume normalization. Shotrunner masters audio to exactly −14.0 LUFS integrated loudness with a true peak ceiling of −1.0 dBTP.

If QC flags any loudness deviation, run the automatic loudfix command:

Terminal (Auto Remaster)

1shorts qc --video series/<series>/scripts/<slug>/short.mp4 --plan series/<series>/scripts/<slug>/plan.yaml --loudfix

Was this "Audio Mastering & -14 LUFS" page helpful?

Image Scraping & Vision

Multi-source retrieval, OCR watermark rejection, framing tags, and library grammar.

CLI Reference

Complete command matrix and syntax for all 16 shorts CLI subcommands.