I have been reading some interesting material on dynamic range recently. Well, it was 'Dynamic Range Day' on March 16, so it seems appropriate.
My point of view is that I hate the over-mastered sound as much as anyone else… Except for the people in record label A&R departments who decide what we are allowed to hear. They all seem to think that louder equals better. It does up to a certain point. But many people in the industry feel that current releases go far beyond the limits of acceptability.
But when I say that I hate the over-mastered sound, it doesn't mean that I hate the mastered sound. No, in fact I love to hear mastering tastefully done. It can turn a good mix into a powerful one, improving both frequency balance and the overall impact of the sound.
Does mastering always have to reduce dynamic range though? We are often led to believe that it does, but in fact it doesn't have to.
Dynamic range defined
One common definition of dynamic range is the difference between the peak and the RMS level of a signal. Since the peak level in a commercial release is always at full-scale, then the mastering process will level out the peaks then bring up the RMS level, thus reducing the dynamic range.
If the difference is less than 12 decibels, then the music will start to suffer. Less than 8 dB and the sound will be aggressive and harsh. 14 dB (or DR14) is thought to be a reasonable difference to aim for, to preserve dynamic range.
You could look at dynamic range from a more musical point of view. Suppose for example that the RMS level of a song was -8 dB during a loud section. The peaks would be at 0 dBFS so this would represent DR8 and probably sound rather harsh.
But maybe it's meant to sound harsh – it's a loud section of the song. Maybe the song has a quieter section where the RMS level is around -20 dB (the peaks in this section would probably be lower than 0 dBFS).
Musically speaking, it would be reasonable to say that this song has a dynamic range of 12 dB when comparing the loud section and the quiet section.
So here's a thought…
What about mastering the loud sections and quiet sections of this song separately?
The loud sections would be mastered in the conventional way. The quiet sections could be mastered in a similar way, but the peak levels held down to -12 dBFS and the RMS levels possibly 10 or 12 dB below that.
Both the loud and quiet sections can now have a similar mastered sound, but in musical terms there is indeed dynamic range… 12 or more decibels of it in fact, comparing either the peaks of the loud sections to the peaks of the quiets, or the RMS levels of the loud sections to the RMS levels of the quiets.
Food for thought
There is a little bit of food for thought here. Normally an entire mix is mastered with the same parameters. But if a song varies in level as performed, then there is a case for varying the mastering parameters as the song progresses. It could combine the best features of mastering done well, with the louds and quiets that modern music often so desperately lacks.