“The judgments of jazz in this book are based on recordings, which have reached a state of technical perfection that makes such an approach valid. Besides, the recording is the most trustworthy witness we have in dealing with an art form of which nothing that is essential can be set down on paper. The reader should not be surprised, therefore, if the words work and record are used interchangeably throughout” (2). While acknowledging the limited efficacy of notation, Hodeir asserts that recordings are similar to written scores in that they offer analysts access to musical works.10 Like Hodeir, Gunther Schuller posits an equivalence between recordings and works when he writes that the jazz historian must evaluate “the only thing that is available to him: the recording” (1968, x).11 Schuller does question whether such “one-time affairs” can be viewed as definitive, but he feels that—in absence of other texts—they, as “primary source[s],” are all that historically minded analysts have at their disposal. And since the most prominent methods of musical analysis were developed for notated music, jazz researchers who want their work to be intelligible have to transcribe music from a recording—to transform it into a score/work, and in the process reduce complex sonic events to the parameters of pitch, rhythm, and volume—before analyzing it.12
The work perspective, though, founders partly because recordings are not “acoustic window[s] giving access to how the music really sounded” (Rasula 1995, 135). Or, as Anthony Seeger explains, “[no recording] preserves sounds. What it preserves are [selective] interpretations of sounds—interpretations made by the people who did the recordings and their equipment” (Seeger 1986, 270, emphasis in original; see Jairazbhoy and Balyoz 1977). Microphone selection and placement, recording media, room construction, frequency equalization, dynamic range compression, and countless other choices affect what we hear on a recording. A change in any one of them can appreciably alter the final product.13 Each of these choices constitutes a human decision, whether a producer’s, engineer’s, or performer’s, oriented toward getting a specific kind of sound, doing something in one way rather than in others.14
Indeed, based on evidence from a number of recordings in the 1990s, one might assert that the now-standard reliance on multitrack recording and on digital editing has led to a broader anxiety regarding the fidelity of recordings to a live performance ideal.15 In the notes for pianist Jacky Terrasson’s 1995 release Reach, for example, Mark Levinson offers the following account of the CD’s recording:
Years ago, musicians recorded music as they played—informally, in close physical proximity, without much editing. What they played was what people heard on the record. Today that approach has been all but lost. Studios separate the musicians, put them behind glass booths, give them headphones and cue tracks, and leave most of the production decisions to engineers in the post-production process—mixing, editing, and mastering.… [In my approach, only] two microphones are used, positioned carefully in the optimum location. The balance between instruments is therefore created by the musicians themselves. There is no opportunity later to change this balance…. Musicians and engineer are in the same room with no glass windows or partitions between them. No headphones or monitors are used by the musicians.16
Here, again, the assumption is that recordings, at least when done well, can provide direct access to what musicians do. Many other releases from the 1980s and ’90s contain similar statements, such as “recorded live to 2-track,” perhaps intended to make them seem more authentically representative of live jazz performance and more accurate as historical documents.
In that capacity, they might also give us privileged access to the authorial intentions of individual musicians. This second kind of fidelity is compromised when we take account of how musicians decide what and when to record (see DeVeaux 1988, 127, 135). In his autobiography Reminiscing in Tempo (1990), for example, producer Teddy Reig explains that he allowed Charlie Parker to choose all of the tunes recorded during his sessions for Savoy—provided they were “original” compositions, that is, ones that did not require the record label to pay royalties to other composers. Orrin Keepnews, however, took a more hands-on approach in recording Thelonious Monk for Riverside:
My partner [Bill Grauer] and I had decided that our initial goal was to try to reverse the widely held belief that our new pianist was an impossibly obscure artist; therefore, we would start by avoiding bebop horns and intricate original tunes. We proposed an all-Ellington trio date: certainly Duke was a universally respected figure and major composer…. [Monk] agreed without hesitation, despite claiming to be largely unfamiliar with Ellington’s music. I insisted that Thelonious pick out the specific repertoire, and eventually he requested several pieces of sheet music. (Keepnews 1988, 122–23)
Monk’s second session for Riverside produced another album of jazz standards, again at the request of Grauer and Keepnews. It was only with his third release that Monk was allowed to record his own compositions.17 Likewise, Joshua Redman (1995a) told me that his decision to record Eric Clapton’s “Tears in Heaven” for the album Wish (1993) came after he listened to a cassette compilation suggestively given to him by Matt Pierson, his producer.18
Finally, it is rarely clear whether the compositions on a given release were rehearsed by a band prior to recording or whether they were created in the studio. Charlie Parker’s most celebrated quintet—with Miles Davis, Duke Jordan, Tommy Potter, and Max Roach—was a working band that almost certainly had performed some of the tunes they recorded in live settings before entering the studio. At the very least, they had performed tunes with similar harmonic progressions. But according to both Teddy Reig (Reig and Berger 1990, 22) and Miles Davis (Davis and Troupe 1989, 88–89), there were also numerous tunes recorded by Parker and his groups whose melodies were composed and learned by musicians on the spot. Such instant composition and performance were certainly affected by the time constraints of recording. As Helen Oakley Dance and James Patrick point out, prior to the 1950s it was standard practice to record four tunes in a three-hour recording session.19 Up to forty-five minutes then, on average, could be allotted to the recording of each three-minute tune. Such generous amounts of time, however, could be diminished by in-studio rehearsal, by false starts and mistakes, or by decisions to change repertoire or modify arrangements. Alternatively, Robert Palmer (1985) suggests that the high quality of recordings on the Blue Note label in the 1960s was due to the label’s policy of financing two to three days of rehearsal prior to each recording session.20
These examples make clear that one cannot definitively say whether an individual recording truly represents a first-time, improvisationally brilliant performance.21 For those writers interested in locating such performances, the use of commercially released live recordings is not a viable corrective, for such releases are as subject to post-performance manipulation as studio recordings. When artists like Joshua Redman and Joe Lovano made their live recordings at the Village Vanguard in the 1990s, it was likely as apparent to other audience members as it was to me that these were not typical performances. Intricate networks of wires and cables ran from the stage to other areas of the club and up the stairs to large mobile recording units parked in front of the club on both occasions. If that weren’t evidence enough, the musicians took care to inform us in each case that the evening’s performance was being recorded for commercial release.22 Moreover, as is standard with studio recordings, some recorded material, such as the intervals between songs or “extraneous” audience noise, didn’t appear on the final releases. Finally, audience applause was recorded on separate microphones to be mixed in later, and the individual tunes chosen for inclusion on the final recordings were sequenced in a manner that didn’t replicate their order on the evening(s) of performance.23
Where musical analysis is concerned, the process of transcribing those same recordings strips dense sonic phenomena of all that cannot be translated into a particular notational system, discourages study of musics not easily transcribed, and privileges the aspects of sound that researchers dependent on Western notation have been trained to emphasize (Tagg 1982, 41–42).24 Consequently, through notational dependence the analysis of jazz has come to resemble the analysis of Western concert music (see Walser 1995, 170–71, for a strategic use of notation-centered analysis). As a result, the majority of jazz analytical work concentrates on the improvised solos of historically