Does Phenom VidFixr's normalising squash the music?

Applies to Phenom VidFixr 3.0 Updated

No. Phenom VidFixr uses linear normalisation: it applies one even volume change to the whole file, so the difference between the loud and soft parts of a song, its dynamics, is kept rather than compressed. That matters far more for music than for speech.

Linear versus compressed

There are two ways to make a file hit a loudness target:

  • Compressing turns the loud parts down and the soft parts up. Everything ends up at a similar level, and the music loses its punch: the quiet verse and the big chorus sound alike.
  • Linear turns the whole file up or down by the same amount, like moving a volume knob once. The song sounds exactly as it was recorded, just at the right overall level.

Phenom VidFixr always aims for linear.

How does it stay linear?

It works in two passes: first it measures the file, then it levels it using what it measured. Standard loudness tools can quietly switch to compressing in two situations, and Phenom VidFixr guards against both:

  • A song with a very wide range between quiet and loud. Phenom VidFixr allows for that song's own range instead of squeezing it.
  • A quiet song with sharp peaks, where raising it to −16 LUFS would push the peaks past the true peak setting. Phenom VidFixr lowers the loudness target for that file instead of compressing it.

The second case means a few songs can come out slightly below −16 LUFS. That is the price of not squashing them. See Why are some songs still quieter or louder?

More about audio