Capturing and Reproducing Dynamic Music

Just how loud is too loud? If you’re the bassoonist in a symphony orchestra and you’re seated right in front of the brass section, you’re going to have to get used to some very loud passages in selected pieces of the symphonic literature. In a recent concert at Disney hall I actually saw musician’s head partially surrounded by a semicircular baffle. It doesn’t get killer loud very often but it does get loud. Very loud. The end of Mussorgsky’s “Great Gates of Kiev” or Stravinsky’s “Rite of Spring” can peak to 120 dB SPL plus. That means that the molecules in the air are very activated and your eardrums are moving very aggressively in sympathetic response. It’s exciting in the context of a great piece of music but it’s not dangerous for the audience (or for the players) unless it’s constant and close up. Loud passages in music are part of what make it interesting, emotional, and compelling.

And the same goes for the quiet passages. The pianissimo tremolo strings near the end of “Firebird” suite that introduce the Horn solo (French Horn…but you never refer to it as the French Horn…it’s just the Horn) is a perfect example or the snare drum the starts the famous Ravel “Bolero” are just above the level of the auditorium. The difference between the two is the dynamic range of the musical composition. In a great hall the potential for very wide dynamics is possible. A very quiet room has an ambient level of about 25 dB SPL. If fortissimo music is played, it can reach momentary peaks above 120 dB SPL…that’s real world dynamics in an ideal setting.

As a recording engineer, I want to capture the dynamics that existed in the hall during the performance. In the past, we were stuck using amplitude modification devices that ensured that our recording devices didn’t crash into the red…and distort. We couldn’t capture real world dynamics. And this applies equally well to commercial music regardless of genre. A jazz or rock drummer can hit a rim shot and blow right past the red zone easily. And singers are the worst! A trained soprano can belt out a note that registers on the Richter scale…I know I’ve recorded many opera singers and it never ceases to amaze me how quickly they can ratchet things up.

Real world dynamics cannot be captured using traditional analog recording equipment without amplitude modification…the use of compressors. However, our commercial recording industry isn’t interested in capturing or delivering real world dynamics except in the rarefied world of classical and jazz. It’s been shown over and over again that louder sells more records. As a result, compressors are used on each individual instrument as a multichannel recording is tracked, compression is used during the mixing process to get that “punchy” sound, heavy compression is used during mastering, AND compression and limiting (a harsher very of compression) is applied at the point of distribution…radio stations, satellite and processing by online channels.

It’s conceivable that consumers don’t need more than 16-bits worth of dynamic range on playback. That’s CD spec and translates to 93 (a few dB taken off of the theoretical 96 thanks for dithering). There are very few recordings that have this level of dynamic range (I know a lot of mine do…but many don’t). However, it’s critical for recording engineers to know that they can run “hotter” without worrying about overages or clipped samples. The primary benefit of using 24-bits or even 32-bits applies to production stage of music recording and means almost nothing to the delivery of “high-resolution” music. When people say they can hear the difference between 16 and 24-bits, they’re probably curing off of the noise floor and not the dynamics of the recordings.