I want to talk about the acoustics of live music and recordings. As I write this I’m back in Boston for a week, re-calibrating my ears with (excuse the expression) the “absolute sound” of live music in various concert halls. On Friday the Boston Symphony played symphonies by Mozart and Shostakovich, producing (as always) magnificent sound with the aid of Symphony Hall’s near-legendary acoustics. On Saturday and Sunday the Boston Philharmonic played music of Kodály, Bartók, and Dvorák in two mid-size concert halls, Jordan Hall in Boston and Sanders Theater in Cambridge. (The latter is part of Memorial Hall, built a century ago to commemorate Harvard’s Civil War dead).
The Philharmonic plays under Benjamin Zander, the brilliant British teacher/conductor. For two decades Zander has been setting an example for Norrington, et al, with thrilling performances of Beethoven and Mahler symphonies that follow the composers’ original tempos and phrasings. Private recordings of these performances have been widely circulated on tape to enthusiasts; his first commercial recordings will be released this year.
Working with Ken Deen, a computer programmer and one of Boston’s many skilled freelance recordists, I recorded both Philharmonic concerts. On Saturday it rained all day, and the sound of the orchestra in Jordan Hall was as brilliant and full-bodied as we’ve come to expect. On Sunday a cold front blasted into Boston; as I walked to Sanders Theater the bitterly cold wind nearly blew my jacket off. We hung the mikes at the same location that has yielded excellent sound in previous recordings. But with the first notes of the concert it became obvious that the recording was too dry, lacking the nice balance of ambience and direct sound that we always aim for and often get.
Hoping to get more “air” in the sound, we raised each mike by about 12″ while a piano was being moved on-stage for the Bartók concerto. During many years of recording in this hall I have found that small changes in microphone height produce dramatic changes in the recorded level of hall ambience. But this time it didn’t work: the recorded ambience improved only marginally, while the sound of the violins became somewhat wiry (footnote 1).
For the second half of the concert, Ken stayed backstage to start the recorders while I joined the audience out front. Since the concert was nearly sold out, I was surprised (and lucky) to find an empty seat in the front row of the balcony, right in the center of the hall. It soon became clear that the lack of high-frequency ambience was not a fault of our miking; it was real. Reflections off the hall’s many old wood surfaces gave the midrange a richly burnished golden glow that made oboe and clarinet solos especially beautiful; but there was no high-frequency “air” in the live sound.
In part, this can be blamed on the audience; the partially reflective wooden seats were covered by 1500 absorptive bodies and wool jackets. But in part it was due to the cold, dry air. When dry winter air is brought indoors and heated to room temperature, its relative humidity (expressed as a percentage) becomes very low indeed.
When soundwaves travel through air, some of their energy is absorbed by vibrating oxygen molecules, and this absorption increases steeply with frequency. For the path-lengths found in concert halls, the absorption is negligible at low audio frequencies but becomes significant in the highs. And when the relative humidity is low the absorption curve rises most steeply, soaking up all of the highs. (The humidity has been so low this week that my lips are painfully chapped and cracked.)
As a sound reverberates back and forth between walls 60′ apart, its highs are progressively absorbed while the lows remain strong. Therefore the reverberation time of a concert hall varies with frequency: bass energy may take two seconds to die away to inaudibility, while mid-frequencies take 1.5 seconds and highs only half a second. In a nutshell, this selective absorption—combined with the reflective character of the walls and ceiling—is the core of what a concert hall does to enrich the sound of a symphony orchestra, strengthening its bass and contouring its midrange with a golden warmth that is your reward for buying a $30 ticket. (A great performance, if you encounter one, is a bonus.)
But as I suggested above, when the humidity is very low the highs are absorbed too rapidly, robbing the sound of “air” and liveliness. Water vapor ameliorates the absorption, improving the transmission of sound at high frequencies. The October 1990 issue of the Journal of the Acoustical Society of America contains newly revised equations and graphs on pp.2019–2020 that describe how the air’s absorption of sound varies with temperature, humidity, and the frequency of the sound.
The dependence on humidity is a particularly complicated function. In the normal range from 20 to 99% relative humidity (RH), acoustic energy at midrange frequencies is absorbed at the same low rate (0.1dB per 100′ of air-path length) regardless of the value of RH. But at high audio frequencies the transmission of sound through the air improves dramatically as the amount of water vapor increases. The following table shows how rapidly the sound is absorbed per 100′ of air-path length, at a frequency of 10kHz:
Relative Humidity Absorption:
20% 9dB
50% 6dB
80% 3dB
This is not a small effect. (A 9dB loss at 10kHz is about what you get when you turn the treble all the way down on an amplifier that has conventional tone controls.) I noticed it in live orchestral sound many years ago, before I learned anything about acoustics. During the winter concert season, when New England air is often cold and dry, the Boston Symphony had a characteristically warm sound with rather soft highs; even the cymbals seemed muted. But when the same orchestra performed as the Boston Pops during the rainy months of April and May, its sound had much more brilliance and punch, and the cymbals acquired a satisfying sheen.
In those days local music critics often suggested that the orchestra played with less subtlety and refinement under Arthur Fiedler’s Pops baton than under the “serious” symphony conductors (Munch, Leinsdorf, William Steinberg). But when a modern airconditioning system was installed during the 1970s, the sound of the orchestra became much more consistent year-round. The Pops today under John Williams doesn’t sound as brilliant as it often did under Fiedler; I think it’s because the airconditioning system prevents the humidity in the hall from ever approaching 100%, even during a spring shower. Conversely, the BSO under Ozawa has a more brilliant sound on the average than his predecessors achieved, partly because the humidity in the hall never gets as low as it used to. As a bonus, stable humidity also helps the players keep their instruments in tune.
The effect of humidity will be proportionately smaller at home, where path-lengths are shorter. But it wouldn’t be surprising if some audiophiles have noticed a small seasonal variation in the sound of their systems.
Incidentally, the preceding comments about the Pops refer to the live sound of the orchestra in the hall, not to recordings. Most of the Philips recordings conducted by John Williams have sounded bright because of accent miking, yielding a sound that has little relation to what audiences hear at Pops concerts.
Even with simpler “purist” techniques, recording microphones normally are located much closer to the stage than most audience members ever get. Mikes are rarely placed more than 10′ behind the conductor—a location where the hall has much less effect on the sound than at a typical audience seat 50′ away. (Symphony Hall in Boston is 128′ long, so at 50′ you would still be in the expensive seats in the front half of the hall.) Result: even the best recordings usually have a brighter tonal balance than what you’d hear in the hall.
The obvious answer is to place microphones where the audience sits, in order to capture the same sound that you hear at a concert. People who have tried it were dismayed to find that the recordings were hopelessly muddled. The obstacle is that microphones don’t have the human ear’s ability to focus on the first-arrival sound from the stage and suppress the first 50 milliseconds’ worth of reflections from other directions. (The tendency of the ear/brain system to lock onto the first arrival and perceive early reflections as timbral modifiers of that sound is called the Haas or precedence effect.) In a stereo recording the direct sound from the stage and the early reflections from the sides of the hall are captured in the same channels and are all played back from speakers in front of you, depriving you of the psychoacoustic cues that your ears provide when you are in the hall.
The bottom line: If you want a clear recording, the mikes must always be placed much closer to the source than most of the audience. Typically they are hung above the first or second row of seats, about 8′ behind and slightly above the conductor. I’m speaking here of purist stereo miking; for multimiked recordings the mikes are placed even closer, among (or above) the instruments of the orchestra.
The close placement of microphones leads both to a brighter tonal balance (relatively unaffected by the hall’s acoustics and absorptive air-path losses) and a magnified soundstage perspective (with greater width and more obvious front-to-back layering in depth), compared to what most of the audience hears in the hall. If you like this exaggerated and hyper-detailed perspective, you may choose to regard recordings as a substitute for an idealized up-front seat, rather than for any actual audience seat that you are likely to occupy at next week’s concert.
Considering only the question of lifelike reproduction of timbre, the logical conclusion is that, while it may be desirable for loudspeakers to produce flat on-axis frequency response when measured at a relatively close distance (1m), it is not wise to equalize speakers to deliver flat response at your chair. Efforts to do so are likely to produce sound that is unpleasantly thin and bright. Speakers that sound convincingly “musical” usually have a response curve that compensates for the brightness of recordings by sloping slightly downward in the treble when measured at a normal distance.
What I’m suggesting for loudspeakers is not a sharp high-frequency rolloff but a response curve that could be drawn as a straight line with a slight downward tilt. In fact, many speakers already provide this. Of course the result depends not only on their on-axis response but also on how their off-axis response interacts with the acoustics of your listening room.
Of course, this suggestion applies mainly to the playback of large-scale music performed in large spaces. If you’re more interested in music performed by small ensembles (classical chamber music, folk ballads, jazz combos, or studio-recorded rock), which you might want to hear from only 10′ away in real life, then a playback system that delivers flat response all the way to your chair could be best.
Since recordings also differ in their degree of excess brightness, what the world needs is a good spectrum-tilt control. Unlike ordinary bass and treble controls, which boost or cut the low bass and high treble by 10dB or more, a tilt control would allow you to make relatively subtle adjustments to the overall brightness or warmth of the sound. I’m talking about making a change of just 1 or 2dB (or, in some cases, 0.5dB) in the relative energy levels at 300 and 3kHz. Such a change is plainly audible, and is enough to serve as a useful corrective for overly bright (or dull) recordings, CD players, amplifiers, et al.
A spectrum-tilt control has long been a feature of Quad preamps, but this good idea has been ignored by most other manufacturers. The omission of conventional bass and treble controls from audiophile preamps is understandable; but I’m puzzled by the near-universal failure to include a tilt control as a more truly useful substitute.
Footnote 1: In most musical instruments, overtones radiate with varying intensities in different directions. In particular, violins produce their most smoothly textured sound for listeners or microphones that are approximately at the same level as the instrument. Imagine a violin as a loudspeaker: strong high-frequency harmonics are projected “on-axis”; ie, upward along a direction perpendicular to the violin’s top plate. The common practice of mounting mikes high above the front of an orchestra, where they can look directly into the violins’ f-holes, contributes to the edgy sound of many orchestra recordings. I once occupied a front-row balcony seat, directly above the left edge of the stage, while Lorin Maazel conducted Mendelssohn’s “Italian” Symphony. The only thing I remember about the performance is that after 20 minutes of hearing the piercing sound of massed violins from above, I had a painful headache.
Click Here: 2020 soccer tracksuit