Acquiring the skill to mix a piece of music takes time and practice, and like playing an instrument, you’re never done as there’s always something new to learn. Along the way we all make mistakes, which is how we learn — as they say, the person who never made a mistake never made anything.
One of the first things we learn in mixing is that we need some kind of reliable monitoring system, whether it be loudspeakers in a treated room or suitable headphones, that allows us to hear what we’re doing with some degree of accuracy. Once we can hear what we are actually doing, mixing a song becomes a balance between art and science. The science aspect relates to things like levels, headroom, final loudness and, to an extent, frequency masking. The finer degrees of balance, panning and the application of effects are more art.
Today we have software that can balance mixes and set up a chain of plug-ins to automatically fine-tune individual tracks and polish whole mixes. We have online services as well as software that can master tracks for us, and more processing plug-ins with libraries of presets than we can shake a stick at. For the less experienced mix engineer, this kind of software assistance certainly helps get listenable results, but it can be a double-edged sword unless used thoughtfully. In my opinion, the more we allow software to take the place of experience, the less experience we gain. How can we learn how to better mix and master if the job is always being done for us by plug-in presets and auto-mix algorithms? Such automation also takes the art out of mixing because the results are always based on algorithms, which, even with the benefit of machine learning, can only apply a set of rules. They can’t appreciate the emotion of lyrics, the importance of light instrumental touches or the vision of the composer. So while they might produce a perfectly acceptable mix, it won’t necessarily be the best it can be.
...by examining what the software has actually done, rather than simply accepting the results, you can learn from it and gain the confidence you need to make your own mix decisions.
Of course you can take a machine-guided mix simply as a starting point and then fine-tune the various aspects yourself but, again, that requires experience. Earlier I mentioned using such products thoughtfully, and that’s where their positive side shines through. If you can enhance your experience by examining what the software has actually done, rather than simply accepting the results, you can learn from it and gain the confidence you need to make your own mix decisions. For example, if a particular processing chain appears to work well on vocals, dive in and look at what the EQ is doing, how much gain reduction the compressor has applied and so on. If you like a plug-in preset, look at the settings and try to figure out why it works so well. Then try to set up something similar from scratch so that you can learn what works and what doesn’t. In other words, allow yourself to make those all-important mistakes.
Paul White Editor In Chief