If I want to export from Ardour 7.4 an audio with 3 dBFS headroom, in the export window I choose: Normalize Peak: -3 dBFS
OK, audio is exported correctly, analysis says it has a -3 dBFS peak, so 3 dBFS headroom.
It’s correct?
But the problem is when I export the Stems: how do I make sure that if one then uploads those Stems, he gets an overall headroom of 3 dBFS?
In “Stem Export” I tried to choose again as above:
Normalize Peak: -3 dBFS
but this time it doesn’t work, the normalization is applied to all tracks and it’s not good, the balance is lost.
How can I do?
Headroom is an analog concept, not directly relevant in the digital world. Leaving headroom makes sense when recording (to avoid clipping), and when exporting for analog playback (to avoid inter-sample peaks).
When exchanging stems, it is preferable to just send the audio data as-is.
If the remote person is to mix it, then send the raw audio, preferably using the format that you recorded to.
If this is for a re-mix or master or education, then apply track & bus processing (checkbox in Ardour’s export dialog’s “Channels” tab). This way just summing all the individual stems will sound just like the current mix.