2. Recorded Objects, Reproducibility, and
Identity
In
his now legendary paper "The Work
of Art In The Age Of Mechanical Reproduction", Walter Benjamin stated that
"to an ever greater degree the work of art reproduced becomes the
work of art designed for reproducibility". This was a provocative observation, implying that the machinery
of mass production was not an irrelevant ancillary to the creative process
but provided an underlying spirit,
structure, and raison d'etre for the art object itself.
Though
Benjamin's comments were inspired primarily by technological developments
related to film and the graphic arts,
it is not difficult to transpose his observations today to the making of
recorded objects.
The
primary property of all recorded objects is their reproducibility. Not just reproducible as objects in their
entirety, as in "I really liked that Elly Ameling recording of Brahms folk
songs and made a tape of it", but reproducible in part, as in "I know
you especially like the way she does 'O kuhler Wald', so I'll make a tape of
that song for you", or "Can you make a tape of the first 8 bars of
that song, please? The producers decided to use them in the film over the
credits." At the most general
level, we are often asked to identify the object under a single rubric. I might
refer to my recording of Schumann's Symphonies #1 and #4, played by the Vienna
Philharmonic conducted by Bernstein. (Or is it my recording of Bernstein
conducting the Vienna Philharmonic playing Schumann's Symphonies #1 and
#4? The rubric often wrests from us a
hierarchical description.) Or take
"New Music For Virtuosos", a recording of seven pieces (four by
Milton Babbitt, one each by Leslie Bassett, William O. Smith, and Charles
Wuorinen), performed by nine independent musicians (who may or may not have
known each other before this recording; the rubric sometimes suggests a
hierarchy -- this is new music for, not by, virtuosos). Or there are recordings like the Beatles'
"Revolver", a collection of 11 songs, unrelated except by the fact
that they exist "together" on the same disk.[i]
To
refer to recorded objects in this general way implies an ability to reproduce
separate entities, or parts of entities. On the simplest (and perhaps most
compositionally useful) level, a recorded fragment (no matter what medium) can
be copied and duplicated anywhere -- within the same, or any other,
composition. Within the same
composition the duplication might offer some recapitulative advantage; the
recurring, resounding church bell in Varese's Poeme Electronique, for
example. Examples of the second case,
recordings pieced together from other recordings, are rarer (perhaps for
copyright reasons). Take, as one
example, English composer John Cardale's "1956". The piece is constructed from fragments of
Karlheinz Stockhausen's "Gesang der Junglinge" and Elvis
Presley's "Heartbreak Hotel",
very cleverly mixed and juxtaposed. One
of the things that makes the piece interesting is the way short (less than a
second) excerpts from works so dissimilar in just about every way, can sound
like they originated from the same piece.[ii]
Reproduction of material at the lowest level involves the copy of
separate tracks, either to other tracks on the same device, or to different
devices. Because a track is as much a
conceptual category as a physical one, boundaries can blur, as the output from
one level becomes the input to another.
An interesting example is Harry Partch's "And on the Seventh Day
Petals Fell in Petaluma". The
piece -- inseparable from the recording --consists of twenty-three one minute
duets or trios ("verses"), followed by ten composites of those
verses, into quartets or quintets.
verse 1 duet ---
|
verse 24 quartet
verse 2 duet ---
verse 3 duet ---
|
verse 25 quartet
verse 4 duet ---
..
..
Theoretically each level is exactly reproducible, though
depending on the media and mode of transfer a certain amount of noise may be
introduced.[iii]
Reproduction of disk recordings depends on a metallic "mother"
which has a limited wear of some several thousand pressings, the last disk in
the run being substantially inferior to the first. With tapes it depends, of course, on whether all copies were made
from the "master" or from other copies. Digital copies are about as noise-free as we have at the moment,
though devices differ in quality, of course, from manufacturer to manufacturer,
model to model. One may well question, then, to what extent any two
reproductions are the same. Moreover,
it's worth asking whether, and how, an object is affected by multiple playings;
a single disc recording, for example, can often easily be identified by the
"sound" of its wear, i.e., the physical (and hence temporal/musical)
position of pops and scratches.[iv]
Reproducibility
and Originals
If a score's primary function is to generate multiple, yet accurate,
performances -- to make the music reproducible -- then the reproducibility of
the recorded object frees the score from that function (and obligation). The
recording can be turned into something potentially more than an electromagnetic
image of a live performance. Once the
recorded image is equated with the event itself, once playback equals
performance, then the recording is freed from its "stenographic"
function, from its need to be representational, to document reality. The
recorded image, creating its own reality ("the choral production,
performed in an auditorium or in the open air, resounds in the drawing
room"[v]) becomes a "new" music in its own
right, engendering new ways of composing and hearing.
Still, the score offers us a unique point of reference, a yardstick by
which any and all compliant performances can be measured. We have a certain amount of faith in the
fact that our copy of Beethoven's Seventh Symphony is related in very
significant ways to an original manuscript prepared by the master himself, and
it's that faith which gives the score its authority. The score is an original.
With
recordings for which there is no score, or for which the score serves only as a
blueprint, we must ask ourselves where an original exists, if anywhere. For example, a piece which has been
synthesized digitally on a computer system and eventually released to the
public in both LP and cassette form might have several "masters" : a
two-track quarter-inch reel-to-reel "master tape" and a "master
cassette", from either one of which can be cut a hard, vinyl
"master", from which the metal mother can be made so that the records
themselves can be pressed. But one may
consider the data files themselves (or on older systems, the stack of 80-column
punched cards) as constituting the "master"--after all, it is from
those files (which may be many, containing representations of several stages of
the composition) that new master tapes will originate.[vi]
If,
as mentioned above, we might well question to what extent any two reproductions
are the same, then from the foregoing we will correctly conclude that unless
and until a relationship can be established between a recorded object and its
original, the identity of a recorded object can never be taken for
granted.
Identity (a digression)
"What
more easy than to conceive a tree ... existing by itself, independent of, and
unperceived by any mind whatsoever?"
(Hylas to Philonous in Dialogue I of George Berkeley's Three
Dialogues.)
People know quite a lot about the properties of sound nowadays. Enough,
anyway, that "Berkeley's paradox" (if a tree falls in a forest and no
one is around to hear it, then does it make a sound?) doesn't seem to be so
problematic. If a tree falls in a
forest and no one is around to hear it then no, it doesn't make a sound.
When a tree falls it disturbs the air around it; how much of the air, and the extent of the disturbance, depend, of course, on the size
of the tree. The air is disturbed
according to the laws of physics, whether anyone is around or not. But air disturbed according to the laws of
physics does not in and of itself constitute sound. That is, the word
"sound" is what we use to describe what happens not just when air is
disturbed, but when the disturbed air reaches our eardrum. It's only then that we say we
"hear" something. A tree
falls in the forest causing millions of air molecules to be disturbed, but if
there's no ear for those molecules to come into contact with, and similarly
disturb, then there is, very simply, no sound.
What if someone puts a tape recorder in the forest, starts it running,
and then leaves before the tree falls?
Won't the tape be evidence enough that sound did in fact occur, even
though no one was around?[vii] No,
this just postpones the moment at which we'll be forced to define sound from
the moment the tree is falling to sometime later when we're listening to the
tape.
Strictly speaking, sound won't occur until the tape is "played
back", and the loudspeakers (or headphones) disturb the air molecules in a
way that we hear something ‑‑ in this case something rather like a
tree falling. Something like a
tree falling, even something like that tree falling, but since we are
not in the forest within earshot of the tree itself, we can not be actually
hearing a tree falling. What is it
then, that we are hearing?
The tape is essentially a storage device, holding a sonic imprint on
it. It takes one kind of device to make
the imprint ‑‑ a microphone ‑‑ and another kind of
device to turn the imprint back into sound ‑‑ a loudspeaker. A tree falls, pushing molecules of air in
all directions. Somewhere close by is a microphone (connected to a tape
recorder). Air molecules disturb the
microphone, causing it to vibrate similarly. The pattern of disturbance is then transformed into an electromagnetic analog, imprinting itself on
the millions of particles of oxide
making up the tape. Sometime later,
during the playback process, these particles are turned back into an electrical
signal and passed to a loudspeaker, which must translate that signal even
farther back to a mechanical one, literally pushing the air in front of it ‑‑
disturbing it ‑‑ in a manner we hope will be nearly identical to
the disturbance the tree made when it fell in the first place.
Uniqueness
of the Sonic Event
It's not difficult to see that there's a lot of room for
"noise" to be introduced into the record‑playback system. If the microphone had been positioned
differently, or if a different microphone had been used, or different
loudspeakers, or if the tape was played back in a different room ‑‑
any one of these changes, and countless more ‑‑would have changed
the way that falling tree sounded when we played back the tape. How do these different ways of sounding
correspond to the way the tree actually sounded ‑‑that is, when
someone was around to hear it when it fell?
Much as manufacturers of recording tape would like us to believe
otherwise, we don't ever confuse the recorded event with the event itself.[viii] In
that sense the question is something like asking how a photograph of the tree
compares to the actual tree. The tape
recording is an imprint of a sonic event which has already occurred ‑‑
an event which took place at a particular place and time in history. And if, as Benjamin Boretz has demonstrated,
"to be in a given place in a chronology is to have a unique sound, and
because to have a unique sound is to be a unique thing, we may truly suppose
that no two musical entities can be
alike, that musical qualities, as elicited by
attribution through a common theory, are all ontologically distinct,
rather than repeatable in the sense of qualia." [ix] One
can only conclude that, if no two trees are identical, then no two trees can
sound the same falling. Our tape, then,
being a recording of a unique event, becomes the only tape of its kind.
But couldn't someone have just as easily hooked up two tape
recorders in the forest, producing two tapes of the same event? Now even if we were to imagine that both
tape recorders were identical (made by the same manufacturer and carrying the
same model number), it would be impossible for both to be in exactly the same
place at the same time. The molecules
of air set into motion and reaching the
microphone of the first tape recorder would not be the same as those reaching
the second. The sonic imprints would be
different, reflecting the different spatial perspective of each tape
recorder. (The tape is, in effect, saying
"what it sounds like from here", much as a photograph is "what
it looks like from here".) And
because each perspective is unique, then, once again, each of our two tapes is
unique.
What
we are admitting here is what we already know to be true in everyday life,
namely that two (or three, or a hundred) witnesses to an event witness that
event uniquely, if for no other reason then because they witness it from a
position unique in time and space. The
event itself, then, can be defined only from one or more specific points of
view. The question asked above:
"How do these different ways of sounding correspond to the way the tree
actually sounded?" must now gives rise to another: how do these different
ways of sounding correspond to each other?
The tapes might correspond rather closely to what a human ear at that
exact spot might have heard ‑‑ and to each other ‑‑ or
they might not. But how important is it
that they do? To the person who thinks
of the tape recording as a sonic moment preserved, as an historical document,
then it's probably desirable that recording and event correspond as much as possible. The recording is representational, like a
landscape, to be rendered faithfully.
"Noise" ‑‑that is, anything which alters or
corrupts the original pattern of sound waves ‑‑ is something to be
eliminated. To another person, however,
for whom the challenge of recording is not in finding ways of eliminating noise
but of controlling it, a tape can be a different thing entirely ‑‑a
means of not only preserving sound, but of shaping it. For this person, the joy
is not in hearing now, in his living room, for the first time, the sound of a
tree which fell in a forest fifty miles away four days ago, but in knowing that
the tree falling in his living room sounds, in fact, like no tree ever sounded
before.
[i] The
question becomes further complicated by the fact that the British release of
"Revolver" (the original) contains 3 additional songs not contained
on the American release.
[ii] And
of course the fact that they are, now, inhabiting the same piece is not without
irony. We are reminded that both
Stockhausen and Elvis, musical contemporaries (seven years apart in age), each
of whom epitomized, in the year 1956, a significant musical "genre",
had, in all likelihood, been totally ignorant of each other's work. Of course all this makes sense only from
the point of view of someone who can identify the strains of "Gesang der
Junglinge" and recognize Elvis' voice, and understand the connection
between the year and the title. Which raises the most interesting -- and
self-reflective -- aspect of the piece: the fact that Cardale's
"1956" could only have been made some time after 1956. (My thanks to David Hicks for this last
point.) In a similar vein, there is
Richard Trythall's "Omaggio a Jerry Lee Lewis", a recording made from
fragments of Jerry Lee Lewis' "Great Balls of Fire".
[iii]
Noise can be eliminated or introduced (often intentionally) with
filtering devices. Of course all
recorded objects are, in a sense, filtered sources of some kind. No sound can find its way onto any recordable
medium without having first passed through some kind of filter, be it a
microphone, tape recorder, or synthesis program. Then, the recorded object must be subjected again to filtering on
playback; it must pass through an amplifier, and then through some sort of
bandpass filter (both of which may be variably adjusted by the listener). And of course there is the loudspeaker
itself, the final voice of all recorded objects, which is itself a filter. (In
mixing popular records it is customary for there to be two pairs of playback
speakers in the studio: one of high quality, capable of accurately reproducing
sound over as much of the audio spectrum as possible, and another
"cheap" pair, more representative of those owned by the average
listener. Some studios have a pair of
car speakers as well. The idea, of
course, is that the mix should be tailored for what the record will sound like
in a "normal" environment, not an ideal one.)
With analog filters it's quite easy to trace a path from source to filter,
with the output of one filter perhaps becoming the source for another. A composition might be "built up"
in precisely this way, during preliminary or final mixing, as a sequence, or
string, of filtering devices. One
should include here all devices designed to modify the signal during recording
and/or mixing: reverberation units, harmonizers, and distortion boxes. Software
filters, on the other hand, are essentially part of the same program which
generates the source; depending on the implementation, it can be very difficult
to separate "source" from "filter", except to identify
lines of code in terms of their functionality.
With digital filtering, then, we must expand our concept of filter from
that of a "black box" to that of a process. (Digital-to-analog
and analog-to-digital converters may be considered high level filters which
combine, in effect, both process and black box.) Filtering then becomes
embedded in the act of composition itself, whether we are generating sound for
recording, or composing a string quartet for the concert hall.
[iv] For
example, I might identify my recording of Schumann's Symphony #1 played
by the Vienna Philharmonic conducted by Bernstein, by the sound of a small
scratch which I know precedes measure 50 of the first movement.
[vi] Or,
regarding my scratchy Bernstein/Vienna Philharmonic recording of Schumann's
Symphony #1: we can be sure that the recording is mine because we can be
absolutely sure that the scratch is not part of the piece. But what about some much more abstract
object; might not a pop, scratch, or distortion (the disk may be warped) be
heard as part of the piece? Only the
original knows for sure.
[vii] This
formulation in no way refers to the Newhart Show episode (related to me by Paul
Lansky), where the tape recorder was positioned in the path of the falling
tree, destroying the evidence.
Obviously the show's producers weren't themselves so sure of the outcome
-- i.e., what, if anything, would have been recorded -- that they felt it
necessary to invent a deus ex machina.