Mexico Medical Student |

One of my recent projects was coming up with some audio/music to use for my daughter’s sleep. White noise is ideal to block out unwanted din, especially living here in Mexico in a closed off gated community (coto) where everybody insists on living their lives outside. You can’t really tell the neighborhood to “shut up,” when it’s 8:00pm. In addition, most houses have no insulation of any kind, so just general noise from the outside such as cars driving by, a short honk, a delivery truck, etc. all have the potential to rouse our little Energizer bunny from the sleep that she naturally fights.

“No problem,” I thought to myself. I just needed a tool to “normalize” the audio so that the mostly classical music doesn’t have the all the dynamic range of the natural louds and softs. Right now, a median volume would mean that soft passages would be unhearable and louder passages (though musically appropriate) would potentially wake her up. Then, I thought, “Wait–I can’t be the only person to have needed this. Let me Google this…”

I could have sworn what I was going to use was a tool/plugin that does “normalization.” After beating my head against that wall for a few days with no good results, I found that what I was looking for is what’s called dynamic compression. “Compression” seemed totally the wrong word to me, because to compress is to remove redundant data to fit a smaller space (and “decompression” would fill it back in). But NOOOO–some crazy group of audio engineers decided to call a limiting of dynamic range–that narrowing the gap between the difference of louds and the softs–”compression.” That’s what radio stations have been doing for years so you can hear everything in car w/o having to ride the volume control, as well as giving extra “punch” to certain audio frequencies so you think one station sounds better even though they play the same recordings.

Jacking with the dynamic range of the latest “Arcade Fire” track is one thing, doing it to a symphony or a string quartet is something different. Classical music is the domain of serious audiophile engineers analyzing acoustics in concert halls, obsessing over the minutiae of different polymer tiles, their placement, etc. all to provide the best aural experience for a live performance. When recording, mics are placed strategically to capture the ambience of the hall, but also throughout the stage. You want to hear the crispness of a freshly rosined bow on a string, the collective breath before a horn section’s opening, the higher frequencies in the “blat” of a low-brass instrument. Trying to dampen these subtleties is blasphemy.

“Baby Einstien” sells 22 minutes of toybox-synthesized classical lullaby crap for $10 on iTunes Music Store. That’s a serious cash-cow; even pop CDs are at least 50 minutes or so. Even if I bought both volumes, that’s not even 45 minutes of non-repetitive music. It is, however, homogenous in terms of timbre and volume, which, along with clever marketing, is why it sells. I’d rather have my daughter listen to the real thing, but I also have a practical need here as well. So imagine my own self-loathing when I subjected many wonderful pieces to a transmogrifying audio laboratory like a Maestro Mengele, removing the artistry, audio quality, and individuality to make a homogenous, similar, and ultimately inferior final product.

I’ll give an example: the “Adagietto” movement from Gustav Mahler’s Fifth Symphony. People in the know will gasp, “You’re putting her to sleep with Mahler!??” because Mahler is known for wide variations in dynamic expression, extremely long symphonies (>100 minutes), very busy and thick scoring (over 120 members in the orchestra, plus a choir at times), and did I mention long symphonies? Relax–this is a slow movement that is scored simply for strings and harp. I even have the relevant excerpt from the score here to follow along for those that want to for the musical sections below:

Note all the instructions Gus has provided above; the score is full of details, all about to be lost. This is the final 18 or so measures of the movement before heading into the buckle-your-seatbelts finale that comes immediately after. In this excerpt, 8 minutes into the movement, we go from quiet yearning to a final gushing emotional outpouring, to a complete fade to nothing. Gorgeous, heart-tugging, and although sublime, completely unworkable to put a child to sleep. Here is a picture of the original waveform in the audio editor (Apple Soundtrack Pro) and the playable excerpt below it:

Now, the bastardized, compressed audio. Note that there is a lot of artifact bringing up the volume from the low-amplitude sections, which you can visually see in the following graphic. This is the price one pays:

Upon finishing compressing 2h of music selections, they were transferred to a minidisc set on “Repeat,” and applied that night. I happy to report it was an unqualified success. I met the dual objective of both blocking out more external noise and giving daughter something to listen to that she can carry with her, even if only subconsciously, as she grows older to appreciate it more. Part of me feels bad for blasphemy I’ve wrought (ok, I’m being a bit facetious here), but Claudia and I certainly get more quality time as daughter gets more sleep, so too bad.

---

P.S. Once unleashed, the compression monster can level anything in its path, rendering even unthinkably inappropriate selections even and unwavering. I’ve always joked to everyone I was going to inflict Soviet composers, such as Shostakovich, on her as soon as possible. I may do some personality damage if I start too early with things like this. LOL!