Effects > Normalization Q's

First, let me say that I’m an amateur audio editor. I mainly import audio for my own personal library, and do some very minor edits (like joining songs that were meant to be played together - one leading into the other).

My questions here are in regards to normalizing volume on the entire library, as I have some older CD’s that have a much lower volume than some of the newer stuff. I’m getting tired of driving along on a road trip chatting with the wife and have to constantly adjust volume so we can maintain a conversation. The music app doesn’t do a good job of leveling the volume.

My music library is stored on my network on a Synology NAS box, and is currently 365.4Gb - 13,619 files. They are all Apple Lossless.

  1. When you normalize songs in one folder to say -1dB, will those songs be the same volume as those in another folder that was processed at -1dB, or would I have to process all songs together?

  2. When you process files and normalize them do you have to output them from Audacity to a new file to save the effect, or will Audacity apply the normalization to the files in their stored location?

  3. Is Audacity the best way to process the entire library, or is there another (better) way of doing it?

The problem here is that “loudness” is not directly related to “amplitude”.

A simple demonstration - play this short file that has two sounds, one after the other.

As shown here, the first sound has a higher “amplitude” (the "peak level"is greater), but when you play the file I expect that you will find that the second sounds “louder”.

First Track001.png

If your media player supports “ReplayGain” (Apple’s equivalent is called “soundcheck”), then that is the best way to get tracks to play at the same “loudness”.

If your media player supports “ReplayGain” (Apple’s equivalent is called “soundcheck”), then that is the best way to get tracks to play at the same “loudness”.

Yes, it does as I only use the Music app (formerly iTunes). The problem is syncing playlists to my phone (iPhone) and playing the songs in the car (or to a Bluetooth/Sonos AirPlay speaker). When songs are synced the volumes are not equalized, and the iPhone Music app “soundcheck” setting really doesn’t work that well from my experience.

What I’m experiencing are songs that are way louder than others. If you adjust the volume for the loudest songs then the lower volume ones are pretty much inaudible when they play. The best way to for me to show you is to take two excerpts from songs & post them here, but I’n not sure how to do that (can you please help me with that too?).

In that case you will need to apply “Loudness Normalization” to the tracks. See: https://manual.audacityteam.org/man/loudness_normalization.html

Got the file sharing thing…


Here’s a “before and after” screenshot.
The upper track is your original sample. The first section is selected.
I applied “Loudness Normalization” set to “-17 LUF” to the first section, then applied it to the second section (they have to be processed individually).
The lower track shows the result.

Here is the result as an audio file:

In this case, I think the first section sounds a bit louder than the second, but it’s much closer to the same “loudness” than the original.

Thank you Steve! That is what I want to achieve. I’ve also just verified that on my MacBook “Soundcheck” seems to be working (somewhat, but not that well), but not at all on my iPhone.

My next questions are this… 1. Can I perform the “Loudness Normalization” on the files with lower volume and bring them up to the level of the other files, or do I have to process everything? 2. I know nothing about the Luf scale - which will achieve a greater “loudness” during processing - a larger number, or smaller (i.e. -17 Luf v.s. -5 Luf)?

You will need to process all tracks.
If you look at the screenshot that I posted, the first “song” was increased in loudness and the second was decreased. It would not be possible to make the first song as loud as the second song was originally without causing severe distortion.
The trick is to normalize all songs to the same “LUF” level.

Note that the numbers are negative, so “-5” is higher than “-10”.
The closer to zero, the louder it will be, but if too high there will be distortion. You need to pick a level that allows all of the tracks to be normalized to the same level without any of them distorting.

For most modern music, -17 LUF is a good choice, but if you listen to classical music, or any other music that has a lot of dynamic range, then -23 LUF would be better.

Thank you again Steve! Processing all tracks will be another big project for me.

“Macros” can help: https://manual.audacityteam.org/man/macros.html