Introduction
On December
2nd 1889, Theo Wangemann, a representative of Thomas
Edison recorded
Johannes Brahms performing two segments of music at the piano. The works recorded included part of a paraphrase
of Strauss'
Libelle, preceded by measures 13-72 of Brahms' 1872 arrangement of the
first
Hungarian Dance for solo piano.
While this unique historical
document should have been hailed as an important 'window' to the world of nineteenth century music
performance practices, the musical information was almost entirely masked by
noise. The recording was, in fact, so noisy that musicologist Gregor
Benko wrote that, "any musical value heard [in the cylinder recording] can be charitably
described
as the product of a pathological imagination."
Indeed, despite various
attempts at filtering and enhancing the recording, the poor quality of the cylinder recording resulted
in a general concensus that the recording was of no significant musicological
value.
Using an innovative
approach to audio signal analysis (see Berger, Coifman
and Goldberg, 1994) PhD candidate Charles Nichols
and I were able to transcribe the music by painstakingly removing layers of noise
to reveal the music embedded within.
Transcription was followed
by careful analysis of the numerous performance nuances, agogic inflections, improvised
segments and added elaborations.
|
Who was speaking?
The
spoken text at the start of the cylinder recording has been wrongly attributed
as belonging to
Brahms. Numerous writers, scholars and amateurs alike, have presumed that
Brahms introduces
himself with the words "I am Doctor Brahms, Johannes Brahms". However a humber of factors raise serious doubts
as to who is speaking. The only mention of the recording by someone who was present (in the
published memoires of Fellinger's son) states that Brahms was introduced. Considering the
time between the announcement and the start of the music it seems improbable that
the same person could segue from speech to playing so quickly particularly given the technological
limitations.
The denoised excerpts
reveals enough of the speech to suggest that the speaker (probably Wangemann) introduces Brahms as follows:
"...Dezember Achtzehnhundertachtundneunzig. Haus von Herrn Doktor Fellinger,
bei mir ist Doktor
Brahms, Johannes Brahms".
("...December Eighteen Hundred Eighty Nine. House of Mr. Fellinger,
with me is Doctor
Brahms, Johannes Brahms.")
(acknowledgements to Guido Behnke for his assistance in transcribing
the text)
|
In reconstructing the
performance we measured and compared
various temporal aspects of Brahms' playing. Since the dance
uses a limited number of rhythmic patterns we subdivided the
data according to rhythmic types. The most recurrent of these
is the dotted-quarter/eighth-note measure unit. There are sixteen
recurrences of this unit in four consecutive six-measure phrases.
Each of these phrases culminates with a half note accompanied
by arpeggiation.
The second section of the recorded segment of the dance is
charaterized by two four-measure phrases of sequential upper-
neighbor melodic patterns in three groups of sixteenth-note,
sixteenth-note, eighth-note separated by a measure of eigth-note, quarter-note,
eighth-note. This section is cadenced by a two measure sixteenth-note sequence
of descending conjunct tetrachords followed by a sixteenth-note pattern
terminated with [eighth-note, sixteenth-note, sixteenth-note,quarter-note.
The score repeats the last phrase with the neighbor note embellishment replaced
by octave skips and a sextuplet arpeggio replacing the sixteenth-note sequence.
For reasons not immediately obvious, Brahms commenced the recording of
the 'Ungarische Tanz' segment on the consequent of the first
phrase, thus starting on a V9 harmony. The phrase structure
and harmonic rhythmof measures 1-12 is:
[(((2
+ 2) -- 2) -- (2 (1+ 1) +
2))]
|
i vii7 V i i V7/III
V7/iv iv
|
Brahms performance commences at measure 13, continuing the six measure
phrases as follows:
[(((2
+ 2) -- 2) -- (2 (1+ 1) + 2))]
|
V9
ii4 ii6 V
ii6 V V7
i
|
Measures 25-48 continue the 6+6 phrase structure.
The second section of the piece is subphrased into four-measure groups
preserving the twelve meashre phrse structure as (4+4+4) rathern than (6+6).
Although the recording ends at this point, the score follows with a transition
of one (4+4+4) phrase followed by one (6+6) phrase that leads into a recapitulation
of the opening section.
While the overall durations of the three large phrase groups that incorporate
the dotted-quarter-eighth-note units (mm 13-24,25-36 and 37-48) do not differ
radically, the internal lengths and proportions of each measure unit
is remarkably variable. Measures 13-24 have an overall duration
of 10.82 seconds, measures 25-36 are 9.28 seconds, and measures 37-48 9.86
seconds. Measure lengths range from .69 seconds (m 21) to
.93 seconds (m 33).
The most outstanding preformance feature of this section is Brahms' tendency
towards 'underdotting'. The eighth note exceeds the duration of its preceding
dotted-quarter in measures 14 and 19, and is approximately equal in duration
in measures 20, 34 and 40. The dotted-quarter is given its
full durational value only once in measure 38.
The other distinctive performance practice is Brahms' elongation of the
middle quarter note of the amphibrach figures of the second
section (measures 49-68).
In addition to the agogic features of Brahms' performance there are a
number of instances where Brahms departs from the score, both
by melodic insertion or alteration, and by modification of the phrase structure
to facilitate closure at a nonterminal musical point.
Complete details
of the analysis are available in our 1996 publication
in Leonardo, and in our paper from ICMC94 .
|
|