Some mixing engineers like to have compression or even limiting on their mix-buss. This is fine unless you want to export your mix into stems, especially when you decided for stem-mastering of your song. (If you don’t know what stem mastering is, please read this first.)

The problem is as follows:

Let’s say you have your mixdown (brown) split into two stems, a so-called TV-stem (without vocals, here: blue), and all the vocals (yellowish).

When your mixing engineer exports these two stems (blue, yellow), your mix-buss compressor (or limiter) behaves differently than during mixdown (where it is fed by the whole mix), because some tracks are muted. You see that in the vocal stem, there are parts of the song with no vocals at all, meaning that the compressor is not going to get triggered (surely not in the way as it is triggered by the brown mixdown), when you’re exporting the vocal stem. And since the idea of stems is that they sum up to the original mix, this is not going to happen here as long as there’s a compressor (or limiter) on the mix-buss.

As a consequence, there are two options in this scenario:

1. Using no mix-buss compression at all, or
2. the following workaround:

Whether you have a plug-in (virtual) or real analog outboard compressor (it’s up to you …), the key thing is that it should have an external sidechain! (There are even limiter plug-ins that do, I won’t tell which, so try to find some on your own …) So that when you’re exporting stems, the compressor is triggered by the whole mix (i.e. you have to route the mixdown file to the compressors external sidechain) – even when you’re only exporting the vocal stem, for example. The compressor “reacts” only to the sidechain signal, but processes the stem you want it to process. This  is the closest you can get to ensure that all your stems sum up to your mixdown after you exported all of them correctly. All’s well for mastering, then!

Hope this workaround is helpful! More tips coming soon …