An electronic instrument used in musicological research for the continuous graphic representation of melody or any monophonic vocal expression with a defined pitch. The melograph displays acoustical information in the form of a melogram which generally shows pitch and loudness as functions of time. Although the computer has replaced the melograph, the manner of presenting the musical material remains an essential stage in research and has changed very little even though the information is now obtained directly from a computer. The computer is used primarily for the measurement based on the graphic representation (which without the computer is done manually) and with the final summation; these two stages may be based on melographic or computer representations.
The melograph was created in the 1950s for the analysis of those melodic elements which cannot be expressed exactly in traditional notation, such as intonations based on systems other than those of Western music, microtonal intervals, contours of glissandos, the attack and decay of notes individually and in relation to adjacent notes, vibrato and relationships between the contours of pitch and loudness. Such elements are important mostly in the study of non-Western music, but also in that of certain Western folk musics and in the performance style of Western art melodies, and in nonverbal vocal communication, such as sounds made by infants, the prosodic layer in speech, and animal sounds. Most of these ‘melodic’ elements, even when they play an important role in shaping the musical style, have not been formulated in musical theory; those that have been addressed in theory exhibit a discrepancy between theory and practice.
The melogram can thus be regarded as a complementary notation (to that based on learned schemata in the theory of the culture) revealing the latent schemata that appear in practice. Moreover, most of the elements discussed, by their very nature, vary during performance, either due to changes during a specific musical event or in its repetitions. Sometimes a considerable inconsistency in repetition is an essential condition for a ‘correct’ performance. Measurement of a single event is therefore meaningless, and thus in formulating the latent regularity one should take the statistical scatter into account. However, doing so requires precise measurements (within certain limits of precision). From an acoustic standpoint, the melogram provides continuous information about the fundamental frequency (the main factor in the perception of pitch) and the overall intensity of the note as a whole (the main factor in the perception of loudness). These two are represented in the melogram by two continuous, simultaneous graphs, which provide information about the intervals (when the pitch axis uses a logarithmic scale), durations of events, and the relative change in loudness (the last two can also be obtained for sounds without a defined pitch, as in drumming). The scales of the axes of frequency, intensity and time may be contracted or expanded so as to suit the degree of detail required in the analysis. Sounds of indefinite pitch (‘noise’; for example, the sound of breathing) cannot be recorded on this graph and their existence may be indicated in various ways, as a break in the graph or as some schematic symbol.
The actual registration of the graphs in the early days was carried out in various ways: by a mechanical needle, which notates in ink on paper; by light-beams striking light-sensitive paper; by photographing from a cathode ray tube, etc. For example, in fig.1 the graph above the notated melody shows the pitch (fundamental frequency) of the melody and its change with time; the graph below the notation is of the relative intensity. Here there is great concurrence between the change in the pitch curve and in the intensity curve. This, however, is not always the case. Moreover, the fact of concurrence or nonconcurrence is of great significance. For example, fig.2 shows two ‘syllables’ in bird calls: the first (a), which is sounded in a relaxed state, is concurrent; whereas the second (b), which is sounded in an excited state, is extremely nonconcurrent. The graph showing the intensity is relatively simple to obtain, and many machines besides the melograph can supply one. Obtaining the pitch graph is more complicated. The melograph is not concerned with the shape of sound vibrations (wave-form, which may be seen on an oscilloscope), but has to extract the magnitude of the fundamental frequencies from these wave-forms. One way of doing this is as follows: each time a vibration passes the zero point an electric pulse is produced, the number of pulses per time unit being proportional to the frequency of the vibration. The combined pulses form an electric current with an intensity proportional to the frequency of the original vibration. A stronger current indicates a higher frequency, and vice versa. Fig.3a demonstrates schematically a sequence of four different pure vibrations (A, B, C and D) as they would appear on an oscilloscope and, in the graph below (fig.3b), how they would be registered by the melograph. Since only fundamental frequencies are to be measured, the melograph incorporates a filter unit to suppress the upper partials, which are often of high intensity.
Before the invention of the melograph in the 1950s, exact intonation was determined with the aid of the simple monochord. The pitches of notes were compared by ear to the pitches of the monochord, the frequencies of which were known. From the 1920s researchers used electronic devices such as the oscilloscope, to obtain the fundamental frequency from the wave-form. The monochord had many shortcomings: in measuring pitch, it relied on the keenness of the ear of the researcher; only one note could be measured at a time; it was impossible to examine variations within a note, for the pitch was related to a single point in time; rapid notes could not be isolated; and the work was very slow. The deficiency of early electronic instruments designed to replace the monochord lay in the detailed calculations involved and in their inconvenience when examining a large body of material.
The melograph has opened up the field for accurate, reliable and convenient analysis of many details important in determining melodic style, and, because it can handle large quantities of material, it is useful for statistical analysis. For instance, it is now possible to determine the rules of intonation in different types of music and styles of performance, whereas previously these subjects had been a matter for speculation without a sound empirical basis.
Melographs have been developed and used at the following centres: at the University of California at Los Angeles (under the direction of Charles Seeger); in Norway (where Olav Gurvin pioneered development of ‘melody writers’ during the early 1950s); in Jerusalem (Cohen and Katz) from 1957; and in Uppsala, Sweden, from 1966. In each of these places the melograph underwent several different stages of development to improve convenience, efficiency and precision; it should be noted that the Seeger melograph model C incorporates the function of the sonagraph so that the melogram includes a sonagram which supplies information about the spectrum of the examined material. Researches with Seeger’s successive models have encompassed many areas of vocal and instrumental music (particularly oriental music) and heightened speech (some of this work is discussed by Crossley-Holland). In Norway a team at the Institute of Folk Music in Oslo examined the differences noted in various folksingers’ performances of a single ‘control’ melody, as well as the characteristics of modern performances of melodies collected 100 years earlier by Ludwig Lindeman and published in his Aeldre og nyere norske fjeldmelodier (‘Old and more recent Norwegian mountain melodies’, Christiania, 1853–67/R), and the study of lokk (cattle calls) and lullabies (see Dahlback). During the 1970s Anna Johnson was carrying out similar research at the University of Uppsala, Sweden. Efforts have been made in Jerusalem to formulate and understand musical phenomena that have previously not been formulated theoretically but are important for characterization of styles; examples include types of ‘intonation skeleton’, concurrence or nonconcurrence between parameters, and the degree of stability or instability of phenomena. Characteristics of the parameter of timbre, which is selected from various points on the melogram, also play a significant role. These characteristics were studied separately from and together with the study of components obtained from standard notation (e.g. motives and melisma) in various styles, including the music of Samaritans and of Syrian Jews, the sacred and secular music of Arabs (see fig.4), the recitation of the Rig Veda, bird calls in various situations, the prosodic layer in Hebrew speech and the playing of the shakuhachi.
The switch from the melograph to the computer (other than for the summaries of the information from the manual measurements) has been carried out gradually. First the information obtained from the melograph was converted from analog to digital, then the melogram was obtained directly from the computer, then the measurements were carried out and summed up by the computer. The final stage (measuring and summing up the quantitative components derived from the graph’s basic parameters of interval, duration, intensity and the interactions between them) is under development, though the sounds produced by keyboard instruments, which are of predetermined pitch, have been successfully measured and summed up by computers.
The addition of timbre to the list of parameters studied by means of the melograph has also been facilitated by the use of computers. Timbre, the subject of many studies, is generally examined without looking at the other parameters and the changes in them. As a complete understanding of the principles of musical organization (with attention to the aesthetic preferences of cultures, periods and composers) requires the examination of all musical components, the melograph in its various incarnations has been and remains useful for the accomplishment of this aim.
M. Metfessel: Phonophotography in Folk Music (Chapel Hill, NC, 1928)
C.E. Seashore: Psychology of Music (New York, 1938/R)
O. Gurvin: ‘Photography as an Aid in Folk-Music Research’, Norveg, iii (1953), 181
C. Seeger: ‘Towards a Universal Music Sound-Writing for Musicology’, JIFMC, ix (1957), 63–6
K. Dahlback: New Methods in Vocal Folk Music Research (Oslo, 1958)
C. Seeger: ‘Prescriptive and Descriptive Music-Writing’, MQ, xliv (1958), 184–95
D. Cohen and R.T. Katz: ‘Explorations in the Music of the Samaritans: an Illustration of the Utility of Graphic Notation’, EthM, iv (1960), 67–74
P.A. Tove and others: ‘Direct-Recording Frequency and Amplitude Meter for Analysis of Musical and other Sonic Waveforms’, JASA, xxxix (1966), 362–71
D. Cohen and R. Katz: ‘Remarks concerning the Use of the Melograph in Ethnomusicological Studies’, Yuval, i (1968), 155
D. Cohen: ‘Patterns and Frameworks of Intonation’, JMT, xiii (1969), 66–92
P. Crossley-Holland, ed.: Selected Reports, ii/1 (1974–5) [on Seeger melograph model C and its uses]
Z. Katsir: Messages in Vocal Communication: Investigation of the Variation in the Babbler ‘Shout’ (diss., Hebrew U. of Jerusalem, 1991)
DALIA COHEN, RUTH KATZ
