I love the binaural touch, but does anyone else think that the Chesky albums are all produced with too low of gain? The volume, compared to everything else in my library, is extremely low. I understand that contemporary music can be produced louder due to the loudness war, but the Chesky works are exceptionally low.
Perhaps Chesky could comment on why he chose the output / input levels of the albums?
While I can't say what Dr. Chesky took-on as his approach, as someone who has done a fair amount of live binaural recording, I think I may have a little bit of insight on this one (though arguably, I could be completely wrong - I'm only surmising here).
If you really wanted to know in an absolute sense (about Chesky's tracks), you would have to look at the files in a DAW or editing software that would show you the dBFS levels versus time for each track; if then all of Chesky's tracks were something like - 10 to - 20 dBFS (regardless of the track number) then this would indicate that, for whatever reason, the mastering was done this way deliberately (i.e. comparatively quiet). On the other hand, if certain passages / movements are something like - 2 dBFS then you know that the track was mastered to be as hot as possible while avoiding any clipping. My guess is that a) much of the music lends itself well to dynamics (unlike a lot of 'popular' music), and or b) that
popular music would, by contrast, have a near 0 dBFS maximum value (or worse, 'live its life' at 0 dBFS), but also, a very low crest factor (i.e. a rather low dynamic range), thus raising its apparent loudness (but at the expense of dynamics).
It's entirely possible that Chesky has chosen to maintain relative loudness in an ensemble of pieces.
For example, in a lot of the live binaural recording work that I have done, several pieces were performed, either as a stand-alone, or as movements of a larger work (Choir, Choral, Kirtan, multi-movement symphonies etc). In some cases, I have recorded performances where both a cappella and band / symphony performances have also taken place. So, in terms of dynamics, even absent any compression whatsoever, one must choose whether the tracks will be normalized (so that each is just a few dB down from full-scale) or kept at their respective level. Certain pieces (or movements) are meant to be
pianissimo in order to bookend other parts of the composition, or performance, whether in mood, loudness, or both. Not coincidentally, this is why a lot of my recorded tracks up on soundcloud vary so much in terms of the dBFS levels; pretty much any of the choir stuff has been mastered such that the relative loudness of each song performed has been maintained (because they were performed in one venue, one night, track after track, as a complete show, I felt that this is how they should be presented - as a suite if you will). Likewise, for any of the Band or Orchestral pieces that I have posted, I have kept the movement-to-movement levels correct so that the dynamics (which incidentally are governed by the Composer as well as Conductor / Director) are maintained.
On the other hand, and once again, speaking of tracks that I have engineered and mastered, there are some tracks that were performed along side of others (i.e. at a show), but because they are not released as part of an ensemble of tracks, their dBFS levels can vary wildly, or just the converse is true. That is, for certain things I have released from a given show, perhaps only a handful of tracks are posted - in this case, I may have viewed them sort of as 'singles', and so I wasn't worried about maintaining relative dynamics. In such an instance, raising the overall gain (of a quiet piece) to get its loudness closer to that of a loud piece makes a bit more sense, because the tracks are stand-alone in nature.
Think about it...
From an ensemble perspective (several tracks played in one setting), if you keep the relative loudness (in terms of dBFS) as they were, track to track, then the listener hears all of the relative dynamics. However, should someone choose to listen to one of the quieter performances, the next semi-loud performance will seem thunderous by comparison.
So, it's possible that he chose to keep the relative loudness of some tracks intact, thus making them seem way quiet in an absolute or relative sense. Remember though, for the overwhelming majority of the listening public, 99% do not know what natural dynamics really sound like, as almost all commercially released (read "popular") music is at least somewhat level-compressed, and not much of it is acoustic either.
Because one of the things that draws most people - certainly those whom choose to record using binaural mannequins and such - is the realism of binaural, realism remains at its core, so to ask someone who has engineered a binaural recording to employ level compression is a bit like asking an atheist to become fervently religious (or vice versa). Those who choose to record in binaural (using mannequin head microphones and the like) are already 'vectored' a bit differently than most people in the music industry in that they really want to produce an enveloping experience for the listener.
Thus, my guess is that Dr. Chesky was going for his vision of accuracy and fidelity, maintaining relative dynamics, and so on. At least, this would be my guess...though I could be completely wrong. It wouldn't be the first time, and it certainly would not be the last either.
Again...these are just thoughts and opinions. I am not acquainted with Dr. Chesky (regrettably) nor those who may have participated in the mixing and mastering, so all that I can do is relate things from my binaural-centric perspective and guess at what approach Dr. Chesky chose to use.
Mark