Normalizing files in batch mode

opensuse 42.1
linux 4.1.15-8-default x86_64
audacity 2.1.1 - distribution release

I would like to have a job (chain?) that reads whatever files are in a directory, adjust the volume, write it back with the same name. I have about 5000 audio files I would like to normalize. They are a mix of MP3 and M4A files.

I started to create a chain but discovered there is no export command for m4a files. The File::Export Audio method recommended in the manual did not produce an export command for m4a.

What would you recommend?

adjust the volume… I have about 5000 audio files I would like to normalize.

You said two different things there. Effect > Normalize doesn’t set volume of the work. It works on sound peaks, such as the tips of the blue waves. Not loudness.

That’s why the ACX AudioBook people have two different specifications, one for peaks and a different one for volume.

I know there’s a struggle with the latest version of pre-built data libraries that Audacity uses to import and export advanced compressed file types. I don’t think all the normal settings are available.

We should wait for a more powerful forum elf.


Normalize can be used in a Chain to make the peak amplitude the same in all files. (As koz mentioned, just because audio files have the same peak (maximum) amplitude does not necessarily mean they will sound the same “loudness”).
Exporting M4A is not supported in Chains, but if you export as WAV you can probably use another program to batch convert them the M4A.

Note that M4A is not actually a “format”. It is just the “file extension” for certain types of files, usually AAC encoded. (

For converting WAV to AAC, “WinFF” may be a suitable option.

Well, foo!

I’d’ve thought such a process would be well established by now what with there being millions of music listeners wanting a relatively even listening level for their collections.

a process would be well established by now

We can put a man on the moon…

There is one way to do this. Chris’s Compressor was designed so he could listen to opera in the car. He couldn’t deal with the volume variations between the acts and segments of the show. Turn the volume up…turn the volume down. The default settings do a reasonable job of leveling out music going through. I increase the first value, compression, from the default 0.5 to 0.77 and the resulting shows have about the same sound and evenness as the local radio station.

There is another process you can use, but it works the same way that SoundCheck does on an iPod. It reduces everything to the quietest song. I don’t remember the tools…

You are, in effect, remixing the music when you do this and many listeners object strongly to anyone doing that.


a process would be well established by now

We can put a man on the moon…

It would seem I will need to modify some shell scripts I have collected, mostly using ffmpeg, to do this.

Which is the part that you are frustrated by? Achieving the same loudness or exporting as M4A?