James Boyk Home
Articles by James Boyk

Dynamic Inflection and the Beauty of Live Music
Hi-Fi News & Record Review (England), June, 1985.
Copyright © jwb 1985, 1997.

[This essay introduced the term dynamic inflection. Twelve years later, no one else seems to have adopted it; but I still find it useful.]

An important aspect of live musical sound, and one which is very helpful in judging reproduction systems, is not recognized by existing terminology. I have coined the phrase dynamic inflection for it. My Caltech students and I find it useful, so perhaps others will, too.
       Dynamic inflection refers to the natural rise and fall of volume in any communicative sound, whether it's saying, "Pass the salt," singing "Freude, schöner götterfunken," or playing music with no words at all. A musical phrase may be marked forte (loud) in the score. That does not mean that all notes of the phrase are equally loud. Rather, there's an expressive moment-to-moment flexibility to them; and this is what I'm calling attention to by the term dynamic inflection.
       Music uses much larger-scale dynamic changes for expressive purposes, too, whether on a short or long time scale; but this is acknowledged by the term dynamic range. And then there are the features of the signal that are dynamic insofar as the test bench is concerned but are not heard as part of musical dynamics: all the various transients, such as the attack of every note on the piano. We hear these as contributing to liveness or immediacy rather than as changing the level of musical dynamic.
       Raymond Chandler said that the best short film scene he had ever written had no dialog at all except a woman saying "Uh huh" three times with three different inflections. In music, too, subtle inflections carry subtle meanings; so dynamic inflection does refer to something important in live sound. In fact, if you try to speak or play music ‘deadpan,' without dynamic inflection, you find that it's not so easy to do; but even a good attempt will have listeners laughing. Evidently this aspect of communicative sound is so important that it's innately human; and indeed, one convention for the speech of robots in films and television is that these have no dynamic inflection.

       Since very little audio gear is good at reproducing this aspect of sound, it's a useful listening category for evaluation. For example—and these are only one man's examples—I find that tube equipment tends to be convincing in portraying dynamic inflection, while solid-state components tend to give you graduated levels of loudness instead of a flexibly- rendered continuum; it's something like getting the "stepped" or "plateau" dynamics of a harpsichord rather than the smooth gradations of a piano, though much more subtle, of course. To be sure, there are some fine solid-state power amps in the world. I've heard two or three that were as good as anything I've ever heard. But most fall down in this area, while even run-of-the-mill tube amps do quite well.
       Digital recording is weak in portraying dynamic inflection also, as it is in so many ways. [I should point out here that James Boyk is one of the few commentators on digital to have performed carefully-controlled listening tests on digital processors using a live mike feed as a program source --Ed.] But unlike the situation with solid-state amplifiers, where I have at least heard a couple of absolutely top-quality units, with digital I have heard none at all, even after several years of trying. These perceptions are the reason that we do All-Tube AnalogTM recording. They also underlie our hyperbole, "Digital finishes what the transistor began." T-shirts with this legend available from us if enough people want them! [1997: These sold out long ago. —JB]
       Digital vs. analog and tube vs. transistor are not my topics here; I'm simply using them to show how dynamic inflection is a useful term for discrimination in listening.
       Relating this new phrase to various other terms, I note that Gerald Sindell refers to dynamic openness, this being the feeling that while the music may be very loud at this moment, you can hear the possibility of its being soft; and that it seems to have no dynamic limit in either direction. (This sense of dynamic freedom also seems lacking in solid-state equipment and absent in digital recording. Even with the very best equipment, of course, it's never so convincing as in live sound.)
       Doug Sax talks about something he calls the jump factor or the startle factor, two good terms having to do with musical attacks and perhaps with gross dynamic range. Sax's Sheffield Drum Record has lots of jump factor, lots of dynamic openness, lots of dynamic inflection. Is it a coincidence that it's an analog recording made using only tube electronics?
       Trying to find the technical correlates of a component's quality of dynamic inflection promises to be difficult. Perhaps a careful examination of gain vs. input level would show some microscopic plateau effects in offending components, but perhaps not. There are, as we all know, many problems that show up with music signal while evading detection with test signals. An example from a different area of music perception: the pitch of reverberant sound goes flat with some equipment; put the same tape or disc on other equipment and the pitch is true. It's hard to imagine a mechanism for this, or a measurement which would predict it. And some pure-digital recordings show pitch instability, even though all digital machines have near-zero wow and flutter measurements. What measurement should be taken instead?
       Ultimately the problem is that we can never have enough kinds of measurement because the number of perceptual categories is unbounded, or at least very large. This is so because what is important in the musical sound changes with the meaning of the music. (For more on this, see "The Music of Sound," my editorial in The Audio Amateur, issue 5/82.
       Our search for a complete set of perceptual categories and technical measurements is thus probably in vain. But it does good if it leads us to deeper attention to our perceptions of live and reproduced music, and to the relations between measurement and perception. At this moment, so far as I know, there are no measurements which have been demonstrated generally to have a positive correlation with listening-test rankings. [1997: This is still true.—JB] And as for the usual "specs" so dear to manufacturers' advertising departments, none of them has any demonstrated correlation with perceptual quality; while the most common, total harmonic distortion (THD), has been known for decades not to correlate at all! Yet the audio establishment goes on using it as though it meant something.
       In this vexing situation, we must be sure that we do not confuse our language for music perception with our language for technical measurement. Let's talk about musical pitch and remember that it correlates with more than just frequency, and that loudness correlates with more just amplitude.
       Let's talk about pitch stability and not assume that wow and flutter tell the whole story. Let's speak of the jump factor, and not think that a measurement of slew rate preempts our perception. Let's listen to the perceived dynamic range and note that it doesn't relate very well to rated dynamic range (so it's unfortunate that the two languages use the same term).
       And let's not be surprised that technical measurements have a hard time dealing with such musically important perceptions as dynamic inflection, dynamic openness, or the most important one of all, the beauty of live music.

James Boyk Home
James Boyk's articles