Must We Test? Yes, We Must! Page 2

More importantly, it is long enough that clipping a transient of this length is plainly audible to the ear. The IHF transient test, in which the toneburst lasts just 20ms, takes no account of the dynamics of the hearing process itself. According to NAD, pyschoacoustic studies have shown that the ear/brain system determines the subjective loudness of a sound by integrating the energy arrival over a 200ms window. Such a rating would therefore directly correlate with how loud an amplifier would play.

To be honest, I think that NAD's proposal has about as much chance of changing the primary FTC definition of an amplifier's power rating as I have of becoming President (footnote 5). I remember becoming enthused back in late 1979 about the idea of rating an amplifier's power in dBW, which would give a number that would indicate directly the subjective difference between amplifier loudnesses. A 100W amplifier, which MMCs seem to think is much more powerful than a 50W one, would be rated as 20dBW into 8 ohms. But as a 100W amplifier is actually only slightly louder than a 50W amplifier, the 50W model's dBW rating of 17dBW would more correctly indicate the subjective difference. Accordingly, my magazine at that time, Hi-Fi News & Record Review, published all its amplifier specifications in dBW. Adding a -3dB correction factor every time the load impedance halved would also indicate the quality of an amplifier's power supply. An amp that, on this basis, delivered 20dBW into 8, 4, and 2 ohms would be clearly superior to one that managed 20, 17, and 14dBW, respectively. I thought I was on to a winner.

No one cared. No one even noticed!

Of course, I should have known better. I had only to look at the fact that linear, and therefore distorted, ASA film-speed ratings had never been supplanted by the altogether more sensible logarithmic ISO or DIN figures to realize that being right doesn't make a case irresistible. But I still feel a little ashamed that I haven't insisted on Stereophile including amplifier powers in dBW in its review headings. Perhaps I will, from the next issue. Listen up, all you reviewers.

The point to note, however, is that, putting aside their commercial motivation, NAD has proposed a primary power rating that ties in with the subjective loudness of different amplifiers. By basing the system on the psychoacoustic parameters of the ear/brain, it becomes a measurement that actually means something, as opposed to the cocktail of traditional tests implied by High Fidelity's "laboratory testing (properly done)," not one of which—unless found to be grossly in error—means anything at all.

NAD has decided to defy the FTC and emblazon this "Musically Effective Burst Power" rating on all its amplifiers and receivers. Buy an NAD product and you get "Extra Loudness," or some such commercial tomfoolery. This may, of course, be a bold but empty gesture, as I understand that the FTC has not escaped the diminution of resources suffered by all Federal regulatory bodies these days. But I still applaud NAD—Bravo!—for trying to bring some sense to a fundamental but nevertheless currently vapid specification.

All of which is a roundabout method of bringing me to the main topic of this month's "As We See It," which is to enthuse over a paper presented to a recent AES Conference by Dolby Labs' Louis Fielder (footnote 6). Sitting among a bunch of serious-minded delegates in a cavernous Canadian conference hall, I felt like jumping to my feet during Louis's presentation and shouting "Right on!," "Tell it like it is!," and other typically English expressions of support. For Mr. Fielder, some of whose previous AES papers (on music's true dynamic range and the true specification for a speaker to be termed a "subwoofer" (footnote 7) reveal that he possesses a truly righteous mindset, appears intent on almost single-handedly imposing rationality on the measurements routinely applied to CD players.

Mr. Fielder starts with two basic sets of data: the dynamic range of the real thing, and human hearing ability. From them he postulates how good a D/A converter system needs to be in order not to be heard introducing distortion or noise modulation. This is the right way to arrange things: rather than pick an arbitrary number, 0.1% THD, say, and state that everything below is "good," everything above "bad," work out from first principles the absolute limits of performance. That way, you might be able to get a handle on why an amplifier with 1% of second-harmonic-only distortion, a typical classic tube design, for example, is objectively worse but subjectively better than a solid-state model with 0.1% of components ranging from the 7th to the 17th harmonics.

Footnote 5: As a resident alien, I can neither vote nor become President. But the Federal Government and the State of New Mexico graciously allow me to pay many, many taxes. Even my sin tax. No taxation without representation, I say.

Footnote 6: Louis D. Fielder: "Human Auditory Capabilities and Their Consequences in Digital-Audio Converter Design," AES 7th International Conference, Toronto, May 1989, Paper 4A. Check out the references, some 65 strong, for an essential reading list of relevant up-to-date psychoacoustic research.

Footnote 7: "Dynamic-Range Requirement for Subjectively Noise-Free Reproduction of Music" and "Subwoofer Performance for Accurate Reproduction of Music" (with E. M. Benjamin), Journal of the Audio Engineering Society, July/August 1982 and June 1988, respectively.

MattJ's picture

I like the idea of using dB for power ratings. Benchmark has similar ideas to use dB for distortion tests.