Skip to content

Delivery specs

Audio targets, per platform

The loudness, true-peak, and codec floor I hit before delivery. Every platform has its own normalization curve; mastering for one is not mastering for all.

Platform Loudness True peak Range Delivery
Social video
TikTok Phone speakers are the venue. Headroom matters more than peak loudness; the platform will pull you down. -14 LUFS-i -1 dBTP 4-6 LU (narrow) AAC-LC 256 kbps · 48 kHz · stereo
YouTube Mix once, deliver once. YouTube negotiates the codec at playback from the same upload. -14 LUFS-i -1 dBTP 6-10 LU AAC-LC 384 kbps · 48 kHz · stereo (Opus accepted)
Instagram (Reels / Stories) Same family as TikTok. Slightly different curve, same posture. -14 LUFS-i -1 dBTP 4-6 LU (narrow) AAC-LC 256 kbps · 44.1 kHz · stereo
Audio-only
Podcast Wider LRA than music streaming. Dialogue should breathe; do not crush voice for loudness parity with music. -16 LUFS-i (stereo) / -19 LUFS-i (mono) -1 dBTP 8-12 LU (wide) AAC-LC 128 kbps mono / 192 kbps stereo · 44.1 kHz
Spotify Mix to one truth and let normalization land it. Submitting a brick master only loses you headroom. -14 LUFS-i (Loud) / -19 LUFS-i (Quiet) -2 dBTP (for AAC) 6-9 LU Submit 16-bit / 44.1 kHz WAV; platform encodes to Ogg Vorbis / AAC
Apple Music Sound Check is on by default for most listeners. Mastered for iTunes metadata adds a hint of headroom; not required. -16 LUFS-i (Sound Check anchor) -1 dBTP 6-9 LU ALAC 24-bit / 48 kHz (lossless); AAC 256 kbps fallback
Tidal Subscriber base skews audiophile; do not over-compress. -14 LUFS-i -1 dBTP 6-9 LU FLAC 16/44.1 (HiFi); MQA / FLAC 24-bit for HiFi Plus
Broadcast
Broadcast (TV / radio) QC will reject for true-peak overshoot before they reject for loudness. Pick EU R128 or US A/85 by delivery territory. -23 LUFS-i (EBU R128) / -24 LKFS (ATSC A/85) -2 dBTP (R128) / -2 dBTP (A/85) 5-15 LU (program-dependent) PCM 24-bit / 48 kHz (BWF); Dolby E for surround chains
Cinema
Theatrical Mix on an X-curve-calibrated stage at reference level. LUFS metering is the wrong tool here; SPL and dialogue intelligibility win. Dialogue at 85 dB SPL (C-weighted, slow) @ -20 dBFS RMS -3 dBTP (headroom for projection chain) Wide (no LUFS target; calibrated to the room) PCM 24-bit / 48 kHz · 5.1 or 7.1 print master
Atmos (home delivery) Renderer-side trim handles binaural / 5.1.2 / 7.1.4 fold-downs. Mix the bed, then audition the renders before locking. -18 LKFS dialogue anchor (Dolby Atmos spec) -1 dBTP 10-15 LU ADM BWF (Dolby Atmos Master) · 24-bit / 48 kHz
Streaming VOD (Netflix / Amazon / Apple TV+) Dialogue-gated loudness, not program-loudness. Netflix Sound Mix Spec v1.5 is the canonical reference. -27 LKFS dialogue anchor (integrated, dialogue-gated) -2 dBTP Up to 18 LU (program-dependent) PCM 24-bit / 48 kHz · 5.1 print master + 2.0 fold-down
General
Safe master (lands anywhere) When you do not know the destination yet. Survives streaming normalization without obvious damage; not optimal for any single platform, but never embarrassing. -16 LUFS-i -1.5 dBTP 8 LU PCM 24-bit / 48 kHz WAV (transcode to AAC 256 as needed)

Specs sourced from EBU R128, ITU-R BS.1770-4, AES streaming guidance, and platform-published normalization docs. Verified against my own delivery logs across film, broadcast, and creator-side mixes.

Get your free assessment