Managing the noise level in mastering for vinyl
The significant problem of noise in vinyl makes it important for the mastering engineer to keep the level of the music well above the level of the noise products of the medium.
This implies compression. Much popular music is compressed heavily so that its dynamic range extends only from very loud down to quite loud. There are few or no quiet sections. This is good because the music can be well above the level of the noise at all times.
Classical music of course is full of quiet sections. And this is where noise will be noticeable. It is necessary to make sure that the quiet sections are no quieter than they need be. For classical music, it is better to compress manually before passing the signal through a compressor.
There is a good example in the original Enigma label recording of a Beethoven symphony where there was one peak that was 6 dB higher in level than anything else in the whole recording.
The section around the peak was copied with some carefully-done fader riding to ensure that the peak was brought down by 6 dB, almost unnoticeably.
The section was then edited back into the master. The result was that on the finished disc the signal level of the entire recording was 6 dB higher than it otherwise would have been, apart from just that one peak..
In other words, the signal to noise ratio had been improved by 6 dB, which in classical music is certainly worthwhile.
There is a trend for music of all kinds to be released in premium vinyl editions, not just heavily compressed popular music, so the art of the vinyl mastering engineer will be in demand for some time to come.