Tone Generation
and Tone Row
of the Jew's Harp

+ Tone Generation
+ Scale of the Jew's Harp: The Natural Harmonic Row
+ Resonances: Sounds and Pitches, Melodies and Accompaniment
+ Open and Closed Vocal Tract
+ Literature

zurück

Inhalt

vorwärts

back to playing techniques

to the contents

forward to playing
on several jew's harps



Audacity Logo  frequency analysis: Audacity  Libreoffice Calc - Logo LibreOffice Draw further calculations and plotting of exported frequency data: Libreoffice-Calc und -Draw


This chapter presents some background knowledge: It starts with tone generation and the specific scale of each jew's harp, dependant on the tuning of the reed - the natural harmonic row. The sound of playing is formed in the mouth and throat. These are the resonating cavities also forming the sounds of speech. 

Tone Generation

It is difficult to place the jew's harp in the system of musical instruments. On the one hand it is classified as plucked idiophone, together with the musical clocks: The plucked part of the instrument sounds itself. On the other hand, the jew's harp belongs to the aerophones, together with the wind instruments and the instruments of the accordion type: In this class of instruments the sound is generated by a vibrating air column (flutes etc.) or by a stream of air stimulated to sound by a reed (harmonica, accordion). The affinity to the accordion instruments becomes clear when jew's harps, hold against the lips and teeth respectively, begin to sound without being plucked, just by breathing in or out. Clemens Voigt of Dan Moi (see links) told me how this effect works: The reed has to be shifted a little from the level of the frame, and the air flow has to pull it back towards the frame level (figure 1, cross section of a jew's harp).

figure 1

On bow-shaped jew's harps the effect works if the reed is pushed gently at the basis by the thumb of the hand holding the frame as in figure 1 to shift it a little. In this way the jew's harp works just like a harmonica. In sound example 14 (64 KB) the tone c is heard, first played on a harmonica, than plucked on the jew's harp, and finally stimulated by breathing in on the jew's harp.

Scale of the Jew's Harp: The Natural Harmonic Row

Seitenanfang
to the top

In our common tonal system the octave consists of 12 equal half-tone steps (enharmonic temperament). This means, that the number of vibrations per second (frequency, unit: Hz) of one tone multiplied with 1,059463 equals the frequency of the following half tone. If any frequency is multiplied again and again with this factor, after 12 steps the frequency is just doubled, and the respective tone is transposed in an octave above. Thus, as the frequency of successive half tones always is altered by the same factor, the absolute difference of the frequencies is less at deep tones (low frequencies) and higher at high tones (high frequencies).

The tone row of the jew's harp (as well as that of any instrument with a natural harmonic row) is different: The reed of every harp has its specific fundamental frequency, e.g. 98 Hz (tone G) or 58.3 Hz (B1). The vibration that produces the fundamental note is always accompanied by its overtones. The overtones of a harmonic sound, like the one produced by the jew's harp reed, are called partials or harmonics. They have frequencies that are multiples of the fundamental frequency. Thus, the jew's harp tuned G plays the multiplication tables from 98 Hz, the one tuned B1 that from 58.3 Hz. This can be seen in figure 2: The sound of a jew's harp was recorded without contact to the mouth, that is, without resonances of the oral cavity (sound example 15a, 126 kB). The freeware programme Audacity was used to determine the frequency spectrum. Each peak of the green curve is one tone played by the jew's harp reed. They are situated at regular intervalls of 58 Hz. The first peak on the left under the first grey triangle is the fundamental note. The next, smaller peak to the right is the first overtone or partial, the following higher peak is the second partial and so on. The yellow curve shows a frequency spectrum of the same jew's harp as played normally with contact to the mouth (sound example 15b, 130 kB). 

Note: Without resonance (green curve) the peaks of odd-numbered overtones (no. 1, 3 and so on as counted from left to right) are lower than the even ones (no. 2, 4 and so on). This is the typical in terms of physics for springs that are fixed on one side and free on the other. For all vibrations forming the overtones the fixed end is a node. The free end is a strong antinode and a weak node, resulting in the stronger even-numbered overtones. Further explanations are given in the literature, see bottom of this page.

figure 2
frequency spectrum

Fundamental note and overtones form the natural harmonic row. The first overtone vibrates with two times the frequency of the fundamental note, being its octave (tone B minor in figure 2, under the second grey triangle). No tone inside this octave can be played on the jew's harp. The next octave (tone b) again vibrates two times as fast (third triangle, fourth partial, 58,3 Hz x 4 = 233,2 Hz). Here we have one tone in between: 3 x 58.3 Hz = 174.9 Hz, which about corresponds to the tone f with an enharmonic frequency of 174.6 Hz. In the next octave until b1 (fourth triangle) there are already three overtones. Thus, the higher we get in the scale, the more tones can be played with a natural harmonic row.

Another example is shown in the animation of figure 3. The grey curve is the frequency spectrum of a jew's harp tuned G without resonance: The frame was fixed to a vice, the reed was plucked and the sound recorded. The red curves show frequency spectra of the same instrument played normally at the mouth, that is, with resonance. Each curve belongs to one tone of a tone row played one after the other (sound example 16, 260 KB) and displayed in the animation successively.  As the playing moves up the tone row, the resonances of the mouth amplify higher and higher frequencies of the jew's harp spectrum.

figure 3

(x-axis: frequency, Hz; y-axis: volume, dB)

offen-Animation
Robert Vandré, 2016




The overtones of the fundamental note deviate from the tones of the enharmonic scale. The 10th overtone is almost in the middle between two tones of the enharmonic scale (see below, figure 5, fourth peak from the left). These deviations are measured in hundredth of a half-tone, the so-called cent. As long as the fundamental note is a tone of the enharmonic scale, the harmonics allways show the same pattern: For any fundamental note the 10th harmonic will diverge by 49 cent from the enharmonic scale.

Table 1 shows how the tones of the natural harmonic scale (marked dark grey) fit into the normal (enharmonic) scale. This is a concise overview of the tones available for playing melodies. A more detailed overview of the specific scales of all tuned jew's harps is given in the overtone table (with explanations, pdf, 68 KB).

table 1
Explanations follow to the right or below.
tone scale of the jew's harp Explanations

The enharmonic scale is listed from top left to bottom right, starting from the fundamental note of the jew's harp. Each column represents the 12 tones of one octave and the first note of the following octave. Each row represents a tone and its octaves.
The fundamental note and its overtones/ harmonics are marked dark grey.
The bold figures show the number of the overtone. The small numbers list the deviation of the respective pitch from the enharminic note as cent. Negative numbers show that the overtone is below the enharmonic note.
To the left and right of the list of tones circles indicate the mayor and minor scales respectively, starting from the fundamental note. Horizontal bars below the table indicate the tone range playable on different jew's harps. The broad part of the bar shows the main resonance (second formant, see below), the narrow part the accompanying lower resonance (first formant, see below).
On jew's harps with a lower fundamental note more different pitches can be played than on those with a higher fundamental note. E.g., a harp tuned A1 plays the multiplication tables from 55 Hz, wheras a harp tuned G plays that from 98 Hz. With the multiplication tables from 55 you will get more steps in the resonance range between 250 and 2000 Hz than with those from 98 Hz.
In sound example 16 (260 kB) the 9 tones of the 4th octave (4th column of the table) are played successively on a jew's harp tuned G.

 Sounds and Pitches, Melodies and Accompaniment

Seitenanfang
to the top

When playing the jew's harp, it is resonances that produce the sound: The jew's harp reed vibrates before the mouth. It excites vibrations of the air column enclosed in the cavities of mouth, throat and chest. Depending on the shape of the cavities the vibrating air column enhances certain tones of the natural harmonic row of the jew's harp. This is shown in figure 2 and 3: Played normally with contact to the mouth (yellow curve, red curve) the jew's harp has the same tones as played without contact to the mouth (green curve, grey curve) - the peaks of both curves are at the same frequencies. But the yellow and red peaks are higher than the green and grey ones, that is, the sound is amplified by resonance. Certain frequency regions are especially amplified by resonance. They are called formants. Thus, the resonances of the mouth cavity are a filter that amplifies or reduces the volume of the jew's harp reed at certain frequencies - just like an equalizer that is a filter for audio signals.

The frequency spectrum of the resonance-"filter" of the mouth can be illustrated, using one of the tones of the scale shown in figure 3 (cp. sound example 16 - overtone 11): The frequency spectrum of the jew's harp without resonance ("vice") was substracted from the curve of the tone played at the mouth (figure 4, top). The resulting differential curve shows the amplifying and muffling effect of the mouth cavern. In the region of 1200 Hz the formant is displayed as a mountain. Accordingly, this is the frequency of the tone pitch played. The filter can also be made audible: With the same position of lips, mouth and tongue, but without jew's harp, the corner of the mouth was knocked gently by the hand. The frequency spectrum of this sound resembles that of the differential curve (figure 4, bottom). Again there is the formant-mountain around 1200 Hz.

figure 5

filter sprectrum of the resonance in the mouth caverndraum


Conclusion: The tones of a certain jew's harp are unalterable, yet the sounds are many and diverse, reacting on the slightest changes of resonance in the mouth, e.g. by movements of the tongue.  

Several formants can be active at the same time. This is illustrated by another small sound experiment using a jew's harp tuned C (sound example 17, 70 KB): The reed is plucked regularly 12 times. The first four strokes I play the tone e3 and the accompanying lower tone g1 (figure 4, green curve). As this jew's harp is tuned C, g1 is the fifth harmonic (see table 1). At the following four strokes the tones e and g are switched: I play g3, accompanied by e1 (figure 5, red curve). e1 is the fourth harmonic of this jew's harp. The last four strokes I play 'openly', like forming the vowel a in "father" (not included in figure 5).  Figure 3 shows not the whole frequency spectrum, but only the peak tops. The amplified frequency regions, that is the formants, are again forming broad peaks with their summits near the pitches I intended to play. Again the primary peak tops of the two frequency specta, that is, the plotted dots of both curves, are at the same frequencies.

figure 5

resonances, formants


The most pronounced resonance region is the second formant of the vocal tract, being somewhere in the range from 500 to 2000 Hz. The vocal tract is the sum of the cavities between the voice and the mouth that form the different linguistic sounds. The highest peaks of the curves in figure 4 show the second formant (blue background colour). On the jew's harp, tunes and melodies can be played using the second formant (see playing techniques: Pitches, how to play melodies). In addition, the deeper and less pronounced first formantof the vocal tract can be used to produce accompanying notes. The first formant can be anywhere in the range between 250 and 1000 Hz. In figure 5 the lower peak on the left side of both curves (red background colour) show the first formant. In the piece of music "Kein schöner Land" (see Music) the first verse is played without, the second one with accompanying notes.

Different resonance cavities and organs are shown in figure 6, left. Especially the tongue influences the vocal tract and the jew's harp sound. The second formant is influenced by movements in the throat (figure 6, right, blue region), the first formant by movements in the oral cavity (figure 6, right, red region).  

figure 6
Resonanzräume   Formanten

The form of the vocal tract determines the resonances and the formants. Thus, it is movements of the voice box, the tongue, the jaw and the lips, that form the sound of speech - and of jew's harp playing. According to the literature, the relations between the sound and the form of the vocal tract are complex. But there are two basic principles that can be illustrated with a simple experiment with bottles. Here, the bottles are a modell for the vocal tract. The mouth of the bottle represents the mouth of the player.  When you blow over the mouth like with a panpipe, the sound and pitch of the bottle cavity can be heard. Bottles with the same volume but different forms have different pitches. Different 1l-plastic bottles were tried here (figure 7): The pitch is lower when the cavity is narrowed near the mouth, and equally when the cavity is widened at the far end. Accordingly, the bulgy bottle with narrow bottleneck (left) has a deeper sound than the uniformly shaped one (middle). If the cavity is shortened and the volume is decreased, the pitch rises. Thus, when the bottle is partly filled with water, the pitch is higher (right). 

figure 7

Flaschentöne

 



The resonance chambers of mouth and chest of the player can amplifie a certain range of frequencies, irrespective of the fundamental note of the jew's harp. This range of tones that can be played is certainly different for each individual player. In her book on the jew's harp, Regina Plate states that the pitches that can be played range from 500 to 2000 Hz, corresponding to two octaves. In my experience tones with a frequency of up to about. 2300 Hz can be played. The notes of the first formant can be as deep as about 250 Hz. On the basis of the fundamental note, the tones that can be played on a certain jew's harp can be looked up in the overtone table (with explanations, pdf, 68 KB).


Open and Closed Vocal Tract

Seitenanfang
to the top

When playing the jew's harp with some breathing, the vocal tract is open at both sides, to the mouth and to the lungs. But when the voice chink (or glottis) is closed (see Playing Techniques, Sound Effects), and also when the tongue closes the oral cavity at the soft palate like when speaking "ng" in the suffix "-ing", the resonance chamber is open only at the mouth and closed at the other end. With a "closed" resonance chamber the jew's harp sound is different and shows the following peculiarity: The even-numbered overtones (that is, the tones with even numbers in table 1, see above) are amplified and can be played especially clearly, but the odd-numbered overtones are not. With "open" resonance chambers, also tones with odd numberes can be played clearly. For example, in traditional Norwegian jew's harp playing, musicians most often open and close the vocal tract for the respective tones of the scale. This is shown on the following picture from a workshop held by the Norwegian Jew's Harp Forum (Norsk Munnharpeforum).


Tonleiter mit offenen und geschlossenen Tönen, Norsk Munnharpeforum

Source: www.munnharpe.no/nmf_festivalar.html , 2015
Åpen = open, Lukket = closed.
Unfortunately, the counting of tones is different from the one used in this text: The 8th tone is marked "Grunntone" (an octave of the fundamental note). This is the 7th overtone. Accordingly, all numbers in the picture are the overtone number plus one. The reason is that the first partial is the fundamental note itself. The first overtone is the second partial.


To examine this effect, odd and even numbered overtones were played on a jew's harp tuned G, and recorded. The tones were played with open vocal tract and some gentle breathing (figure 8, blue curves), and with closed vocal tract, either by expressing "ng" (closed at the palate, red curves) or with closed voice chink/glottis (yellow curves). Freqency analysis was done with Audacity, seperately for the different tones and modes of playing. As an example of the results, in figure 8 the spectra for the playing of the 10th and 11th overtones are plotted.  As a reference, the sound of the same jew's harp without any resonance of the oral cavities was analysed. For this, one side of the jew's harp frame was fixed in a vice and the reed plucked and recorded (lower diagramm in figure 8, "vice"; this curve is also shown in figure 3).



figure 8

resonance with open and closed vocal tract

Without resonance (vice), the even numbered overtones are louder than the odd ones, which is normal for the vibrations of a spring fixed at one end. When holding the jew's harp to the mouth and intentionally playing the 10th overtone (about tone cis) the respective peak of the frequency spectrum is clearly pronounced (figure 4, upper diagramm), with open as well as closed ("palate", "glottis") modes of playing. With closed playing, the peak even seems to be a little higher than with open playing, pronouncing the tone more clearly.
Without resonance, the peak of the11th overtone (tone d) is smaller than the adjacent 10th and the 12th (lower diagramm, "vice"). When playing this tone, the respective peak only with open playing is sufficiently elevated to be at least a little louder than the adjacent ones. Closed playing pronounces the same frequency region, but the 11th peak is not elevated over the adjacent ones. Thus, tone d will only be heared with open playing.

This effect is more clearly shown when all even and odd overtones are plotted. The volume of the respective played notes (peaks) was compared to the adjacent peaks. For this purpose the mean dB-value of the respective two adjacent peaks was substracted from the value of the central peak. When the resulting difference is positive, the played peak protrudes from its surroundings - the played tone can be clearly heard. Figure 9 shows the mean of differences for odd overtones and even overtones (error bars show standard deviations).

figure 9

resonances with open and closed mouth cavern: mean of the difference of played tones and adjacent peaks

Odd-numbered overtones can only be heared clearly with open playing. The closed play modes "palate" and "glottis" amplify the even overtones very clearly, but not the odd ones: They are surmonted by the adjacent peaks.
The bars on the right side of the figure ("vice") reflect the typical alternate pattern of high and low peaks (due to the fact that the reed is fixed at one end, see above).

The respective sound recordings:

scale played with open vocal tract = sound example 16
scale played with closure at the palate
scale played with closed voice chink (glottis)

So, the frequency analysis confirms the idea that odd-numbered overtones are played most clearly with open vocal tract, and even-numbered with closed. The reason for this is not clear to me so far.



That is all. At the end some Music (339 KB).

Write down jew's harp music? Sound example "Music" in music notation (27 KB)

Literature



zurück

Inhalt

vorwärts

back to playing techniques

to the contents

forward to playing on
several jew's harps