The Fifth Element #17 Page 3

One reason I think we tend to play things too loudly (guilty as charged here, at least for Mahler, and Elgar's The Dream of Gerontius) is to make up for the lack of the sense of envelopment we get in a performance space. Another reason is that many home systems can't do justice to bass fundamentals, so we turn up the volume to try to compensate. I've found that with speakers that are full-range or almost full-range, I tend to listen at lower average levels. Really.

Anyway, even in the absence of hard data about SPL levels at the recording session (information I have included in the technical notes of many of my recordings, most of which are about 70-74dBA from a listening position), the doctrine I adhere to (which, I believe, Peter Walker of Quad fame may have been the first to posit) is that, for a recording made with a simple stereo microphone array, the apparent volume of the instruments should correspond to their apparent distance from the listening position.

In other words, a close-in but purist recording of an acoustic guitar should sound no louder than a guitar close up, and a purist recording of a piano that is more distant, and that lets you hear a fair amount of hall ambience, should sound no louder than it would under similar live circumstances. Certainly, we all add a little volume to make up for the usually higher ambient noise of our listening rooms, and please be assured that I am not proposing that people who play things much louder than real should be thrown into jail without trials. By all means, try them first.

Seriously, though, my mania on this particular issue is not just a manifestation of leftover Neoplatonism. What Moulton's lecture on equal loudness curves (in Vol.3) makes much more obvious—with musical examples—than I ever could with mere words on paper, is that, when we depart significantly from the loudness level of the original, we impose a new and foreign EQ curve on the music, owing to our ears' nonlinear perception of bass and treble at different reference loudnesses.

This is so important that I have to say it again: Music is going to sound as it was intended—in terms of its essential tonal balance, and not just loudness—only if it is played back at something like its original level.

I anticipate the objection that the original level of rock music is a rock concert's ear-endangering SPL. Actually, most engineers who do the final EQ and mastering for most commercially successful bands' records do all their work at a reference level of about 85dB. Bob Ludwig will leave the room if people want to hear it much louder.

The point—which, again, Dave Moulton's lecture drives home with carefully prepared audio examples that words are inadequate to convey—is that our perception of frequency extremes (but especially the bass) changes with changing volume to a much greater degree than the average loudness changes. The "Equal Loudness Curves," reproduced in the accompanying book, show this graphically.

I recently made a demo recording of violinist Arturo Delmoni accompanied by a wonderfully full, rich-sounding pipe organ. The climax of that piece measured 84dBA at the microphone position—plenty loud. If your system doesn't sound good at 85dBA, there's probably something wrong with it (or your room) that turning up the volume knob won't fix (footnote 2).

I don't have the space to give as much detail as I would like of the other volumes, but they're meaty and thought-provoking. To give just one example: The lecture in Equalization in which different octave bands of music were boosted at times sounded to me as if I was changing loudspeakers. The Stereo lectures' use of different microphone patterns and methods was a succinct introduction to that fascinating kettle of fish (footnote 3). What can I say? Highest recommendation for all four volumes of Playback Platinum.

If you're brave enough to get out of your easy-chair mode and want to develop your skills seriously, Dave Moulton has another set of CDs that are not only ear-training but ear-testing: the Golden Ears series. How's this for serious: Golden Ears Vol.2 gives drills and tests on "31 possible signal processing changes, grouped into simple families: amplitude change, gross and subtle distortion, slow and fast release compression, equalization changes, stereophony anomalies and time-delay/reverberation settings."

Be forewarned: the Golden Ears series does not consist of lectures you can enjoy passively. These "graduate-level" discs require hours of dedicated, serious critical-listening work if you are to get anything significant out of them. To buy the Golden Ears discs and not take the blind-listening tests would be like buying a health-club membership and never working out. No pain, no gain.

Finally, if you or someone you know is thinking (again, seriously) about audio engineering or music production as a career, Moulton's encyclopedic 450-page college textbook, Total Recording, includes 300 illustrations, comes with its own CD of audio examples, and seems fairly priced at $89.95. Ordering information on all these products can be found at David Moulton's website.


Footnote 2: Measuring the loudness of music with an A-weighted meter can give misleading results, because the A-weighting filter discounts the influence of low and high frequencies. (See my article describing the design of a true peak-reading sound pressure level meter in the October 1981 issue of Hi-Fi News & Record Review. However, the violin's energy is concentrated in the region where the A-weighting filter has minimal effect, so JM's recommendation is basically valid in this instance.—John Atkinson

Footnote 3: I sat up bolt upright when one of the stereo-miking examples was a soprano with a truly luscious voice singing one of Dvorák's Gypsy Songs in Czech.—John Marks