The other is ensuring that your mix not only does not hit 0dBFS, but that it does not contain inter-sample peaks that will introduce distortion in data-compression codecs. One is getting the dynamic range of your mix into the right ballpark.
Two things are paramount when mastering for a loudness-normalised environment. In a world where louder is no longer automatically better, how do we avoid pitfalls at the mastering stage? Today, there’s a bewildering variety of streaming formats to cater for, each applying its own set of rules, which often aren’t made public. If you were doing home mastering for CD, you simply cranked up the output limiter to the point where you started to find the side-effects unacceptable. However, what the old peak-normalised regime did have in its favour was simplicity. Most right-thinking people now agree that the ‘loudness wars’ of the ’90s and ’00s were a Bad Thing, and that the subsequent shift towards loudness-normalised playback has been a major turn for the better.
