Konpon Daito Pagoda, Koyasan Temple
(Electronic Music)
1966 (17'30")

     In 1966, Stockhausen was invited to Tokyo, Japan for several weeks, where he took advantage of the audio equipment at the Japanese radio broadcasting studio NHK.  It was here that he first started to really explore the use of "found sound", i.e. - material generated from outside sources on tape or later, radio (nowadays, this practice is of course widely known as "sampling", or for environmental sounds, "field recordings").  In the resulting 5-channel work TELEMUSIK, Stockhausen manipulated recordings of traditional world folk music with ring-modulation (and other effects) and combined them with synthetic electronic tones.  In fact, the synthetic elements were intermodulated with the world music through the use of new ring-modulation circuit combinations.  Stockhausen would soon further expand his usage of "historically pre-formed objects" to create HYMNEN, a roughly 2 hour work based on the national anthems of the world.

"(In Japan) there were more modulators, which made it possible to transform
and modulate existing music and then combine it with electronically produced sounds.
And I thought it might be possible to produce a piece in which the entire music plays in a very high pitch-range...
now and then I would reflect a few parts downwards
(sections, so to speak, of this music that seems to be so far away because the ear cannot analyze it),
so that it entered the normal audible range and suddenly became understandable.
This way, I would then be able to mix electronic and existing music in the very distant
(or seemingly distant) range, and depending how I modulate parts of it downwards,
I would be able to mix the two categories and modulate them with each other.
Later, I derived the title of the work, TELEMUSIK, from this process:

“Tele” means also to bring the “distant” close up."
- Stockhausen from 1966 Lecture CD 16 (edited Jayne Obst transl.)

     Stockhausen only had a few weeks available to him in order to complete the work, so he and his assistants recorded "live" to the 6-track master tape, adding section by section.  Oftentimes 1 or more tracks were copied to another track ("bounced") through a signal processor.  On average, Stockhausen and his team completed 1 section (out of a total 32) per day.

     Before Stockhausen left the Tokyo NHK studio in 1966, he mixed down the 5-channel tape (from an 8-channel tape machine) to 2-track (with the stereo distribution described down below in the Form Structure section), since at that time there were no tape players available to play this kind of tape anywhere else.  Many years later (1987, 1995), Stockhausen contacted one of the original engineers who assisted him in the 1966 recording in order to get a copy of the original tape.  The engineer was able to refurbush a 5-track machine to operability (just before it was to be junked) and made a copy to a modern storage tape, but unfortunately the frequencies were distorted.  Stockhausen writes in the CD booklet, "When I heard this tape (referring to a 1987 dub)...literally nothing of my work (which I have in my possession as stereo copies since 1966) was recognizable.  Only a completely distorted maze of vibrations had been sent to me."  The stereo version was however, remastered from the 1966 2-track dub to a new transfer by Stockhausen himself and is available on the TELEMUSIK Lecture CD, TEXT-CD 16 (see link at bottom).

Electronic Music
     The synthetic tones which Stockhausen created are mostly very high or very low-pitched.  This alone makes the sonic "band-width" of TELEMUSIK pretty unique (he had just previously explored a full range of pitched synthetic timbres in his previous electronic work, KONTAKTE).  This middle range "bandwidth canal" allowed room for the ring-modulated world music samples to bubble to the surface (in fact, most of the samples used are so warped by the ring-modulation, that they often sound synthetic themselves, and almost unrecognizable as "folk music").  The equipment Stockhausen used to create the electronic layers included the following sound generators:
  • sine wave generators
  • beat frequency oscillators (creates various pulses/sine waves/bands of sound)
  • function generator (used to create irregular non-sine curve waves, including irregular square waves)
  • delta generator (low frequency pulse generator used as 2nd input to control amplitude modulator)

Ritual & Dance World Music Samples
     For the folk music elements, Stockhausen used samples from records of the following traditional world musics (not in order of appearance):
Country Genre Instrumentation/Environment Title
Japan Gagaku court music ryuteki (flute), hichiriki (reeds), sho (harmonium), drums, etc.. "Etenraku" 
Kabuki dance Matsuri Bayashi ensemble: plucked string, flute, perc "Yasai Aikata"
Shinto dance Aoi Matsuri Festival "Daidai Kagura"
Noh theater music percussion (ōtsuzumi) w. M/F vocal, flute "Hashi Benkei"
Buddhist temple chants Koyasan Temple "Senbotsusha Irei Heiwa Kigan Daihoyoh ceremony"
"Shingon-Shu ceremony" 
Temple chant from gyōdō procession, circular walking chant w. wooden shoes, Yakushiji Temple "Jion-e" 
Conch music Omizutori Festival, Nara   field recordings
Bali Gamelan orchestra rhythmic metallic and wood percussion "Baris Bapan"
Vietman Music of the Montagnards female voice "Song of the Festival of Love" 
bamboo flutes "Concert of Bamboo Whistles" 
gamelan-like "Air de Gongs" 
China Peking opera Chinese flute and orchestra: flute, strings, fast percussion  "Keihosau"
Spain Flamenco guitar, castanets, rattles, male singing "Sevillanas"
Hungary Hungarian folk song children's song for violin, fast rhythmic drum tom "Pista Bácsi, János Bácsi"
Africa South Saharan song kalimba-like percussion w. group singing "Ibani-Sansa" 
South America
Shipibo Indians song drums, ringing, blowpipes, girls exclamations "Dance for Adolescent Girls"
Javahé Indians song girls, panflutes, rattles, perc. "Lullaby"
Suyá Indians song low men's choir w rattles

     From the selection above, it seems that Stockhausen chose mostly ritual or folk dance music in order to bring the distant past into union with his very modern synthetic tones.  The Stockhausen official website has a Text-CD (lecture recording) of Stockhausen talking about TELEMUSIK available for purchase, and the page also includes brief audio samples of each of the world music selections he used.

     Stockhausen also used field recordings of Japanese temple percussion sounds as "signals" at the beginning of each of the 32 sections.  The resonance (decay time) of each percussive strike is proportionately related to the length of the ensuing section.  The sequencing of the 32 sections is based on durations in a Fibonacci series, which is a device which Stockhausen has used in other works as well.  Listed below from highest to lowest pitch, Stockhausen used the following bells/wood clappers:
    • taku (柝, wood clapper): 3rd A above middle C (least resonant, shortest sections)
    • rin (metal singing bowl): 3rd Ab above middle C & Bb above middle C
    • bokushō (墨床, large wood clapper): 2nd E above middle C
    • keisu (磬子, large bronze bowl): G# above middle C w rising gliss & D below C (longest decay, for longest section)
    • mokugyo (木魚, "wooden fish"): Eb below middle C
    • 5 kane (large temple bells): bass clef and below
    (High Wood)
    sp   ace
    (High Metal)
    (Med Wood)
    (Med Metal)
    Mokugyo (wooden fish)
    (Low Wood)
    (Low Metal)

    "TELEMUSIK is not a collage.
    Rather - through the process of intermodulation between old 'found' objects
    and new sound events which I made using modern electronic means - a higher unity is reached:
    a universality of past, present and future, of distant places and spaces:
    - Stockhausen, 1969

    At the NHK Tokyo studio, Stockhausen used the following signal processors:
    • 2 ring-modulators (creates sum and difference frequencies of 2 inputs, see MIXTUR)
    • amplitude modulator (one signal is modulated by a second input, creates fades/pulses)
    • high and low-pass filters, octave filters (to select and accentiuate/filter out frequency bands before and after ring-modulation)
    • transposing variable speed tape recorder (used mostly to create "echoes" on the Structure signal hits)
         These signal processors were used to modulate the synthetic tones, the world music fragments, and the synthetic tones with the world music.  For TELEMUSIK, Stockhausen designed a few unique circuits, one of which he called the "Gagaku Circuit", since it was first used to modulate the Japanese gagaku orchestra music in Structure 3.  This circuit basically used 2 ring-modulators, one to modulate the original signal, and another to modulate the ring-modulated signal (sometimes with glissandi in the 2nd ring-mod sine-wave) to create a "double ring-modulation" mix of the original sample.
    In this example, 12 kHz was used in both the 1st and 2nd ring-modulation, with a glissando in the 2nd ring modulation.
    .The music was also frequency-filtered in different stages at 6 kHz and 5.5 kHz.

         For example, in one scenario (above), the 1st ring modulation A used a very high 12 kHz sine-wave base frequency, resulting in a very high-pitched buzzing texture (for example, a piano note of A, or 0.440 kHz, would become a high 12.440 kHz and 11.560 kHz).

         The 2nd ring-mod B base frequency (in this case with a slight gliss. variation on the same 12 kHz base frequency) has the effect of "demodulating" the signal (bringing it back down to near A).  This demodulated signal is also frequency filtered to accentuate low frequencies (dark sound).

         These 2 elements (high buzzing from the 1st signal and low distorted sounds from the 2nd) are intermittently mixed together with faders.  By varying the 2 ring-mod base frequencies and the 3 frequency filters, different effects could be achieved.  This process of modulation and demodulation is what Stockhausen means when he says he was able to "reflect a few parts downwards". 

         This is only one example of intermodulation between a single 12 kHz tone and a second sound source using the Gagaku Circuit.  In 1967, Stockhausen described (in an introduction following the world premiere of HYMNEN) some of the many other ways in which he was able to intermodulate electronic tones and folk music in TELEMUSIK:
    1. Rhythmic Modulation: The rhythm of one (folk) music is modulated onto another fragment.  In other words, if a slow chant is the 1st input and a drum beat is used as the 2nd input, each strike of the drum will create a ring modulation "disturbance" onto the the chant input, in the rhythm of the drum beat.
    2. Harmonic Modulation: Synthetic chords are modulated to follow a folk melody (or the reverse, where a drone-like Koyasan temple priest chant is pitch-transposed by the melodic changes of a synthetic tone melody). 
    3. Dynamic Modulation: The dynamic envelope of a song (such as an Indian lullaby, softly sung but with intermittent loud accents and pauses) is modulated onto another (for example, an Hungarian folk song, with a constant rhythm accompaniment).  In a typical scenario, dynamic accents and rests from the 1st input source control the output volume of the 2nd input (for example, the Indian lullaby's accents and phrasing impose unpredictable volume spikes and gaps of silence on the otherwise-constant Hungarian folk song).
         In all honesty, it's very hard to point to a section in TELEMUSIK and identify specifically which of the above types of intermodulation was applied, and the realization score itself does not seem to highlight the intermodulation type either (though it's in German, so I may well have missed it).  In any case, with the Gagaku Circuit and the use of various color (bandwidth) filters, many shades of modulated world music sounds were obtained.  Several other types of circuit schematics were also used and are included in the score (and the CD booklet).  See MIXTUR for more on ring modulation.

         Because of limited equipment, often times multiple sine waves were recorded on different tracks and then bounced to a single track, which was then processed with ring-modulation (or the Gagaku Circuit) and bounced to yet another track.  Some explanations follow, though the CD booklet gives more information about the details and circuits (score excerpts copyright Universal Edition).

    Structure 1
    In Structure 1, the sequence is described in these steps:
    1. Beat frequency generators were used to record 5 pairs of tones on tracks II-VI.
    2. Tracks II-VI were bounced to track 1 to create a 10-layer sine wave band.
    3. 4 pairs of tones (1 even, the other with slow vibrato) were recorded to tracks III-VI.
    4. Tracks III-VI were bounced to track II.
    5. 3 pairs of tones were recorded to tracks IV-VI in 3 fragments (and changing frequencies for each fragment).
    6. Tracks VI-VI were bounced to track III.
    7. 2 pairs of tones were recorded to tracks V &VI in 5 "tight" bursts (and changing frequencies for each burst).
    8. Tracks V &VI were bounced to track IV.
    9. Tracks I-IV were bounced to track V.
    10. Track V was processed with RM (ring-modulation) and a fragment was bounced to track IV.
    11. BOKUSHO field recording dubbed onto track III (erasing the previous tone fragments from Step 6).
    12. Track VI erased.
    (After this process, track V still has tracks I-IV combined, which is not shown in the score graphic)

    Structure 3
    1. 3 different tones (let's call them A, B and C) are recorded through a RM circuit onto track I.  A, B and C go straight through, but B & C also surface as 3 RM'ed fragments with different attacks/decays.  C also has a fast descending scale near the end.
    2. Repeat the same process with 3 different frequencies and attack envelopes for tracks II-IV.
    3. Record Japanese gagaku music through the Gagaku Circuit onto track VI.  The first RM signal is even (but pushed into a high frequency range), the second RM signal is in 2 de-modulated "audible range" fragments (the first even, the second with a slow glissandi, causing beat patterns). 
    4. Track VI and a MOKUGYO field recording (with accelerando) are simultaneously dubbed onto track V.
    5. Track VI erased.

    Structure 5
    1. The beat frequency oscillator was used to create 11 different pulse tones (in irregular rhythm) on track II.
    2. Repeat with a different set of tones and rhythms on tracks III-VI.
    3. Bounce tracks II-VI onto track I, resulting in 55 irregular pulses between 8000 and 11970 Hz (the score graphic in track I is an example of just 1 layer of 11 tones).
    4. Repeat steps 1-3 using using longer pulse tones on tracks III-VI and then bounce the 4 layers onto track II, simultaneously with a BOKUSHO recording.
    5. Combine tracks I and II (from steps 1-4) with 3 sine tones into a ring modulator, then process through an amplitude modulator (controlled by a low frequency wave from a delta generator).  The output is passed onto track III in 10 fragments of different attack shapes.
    6. Balinese gamelan music is processed with the Gagaku Circuit onto track IV (as in Structure 3, this creates one HF band and a few fragments of twice-modulated sound).
    7. African dance music with voice and kalimbas (?) is processed with the Gagaku Circuit onto track V (1 HF band, 1 layer of audible material beginning about a third in).
    8. Track VI erased.

    Form Structure
         The chart below shows the bell/clapper "signals" and the Structures where world music samples were used.  The "SYNTHETIC TONES" column summarizes my impressions of the most audible electronic elements, including some of the modulations of the folk samples, but in general every world music fragment was processed through the "Gagaku Circuit".  The channels are referred to in this column as I through V, but they are not actually the 1st through 5th sequential columns in the table, since these table columns are arranged from far left to far right (for stereo listening - they're not in order, since Stockhausen switched channel II with IV).  When the channel continues down into the next Structure without a border, that means that the sound structure continues in some way.  RM means ring-modulation, HF or LF means high-frequency/low frequency, X means that that channel was unused.  If a box is blank with no "X", then some electronic element was present (and probably described in the "SYNTHETIC TONES" column).  The CD track numbers are from Stockhausen Complete Edition CD 9 (tracks 1-66 are MIKROPHONIE I and MIKROPHONIE II, also included on CD 9).
         For convenience's sake, the bell/clapper signals can be summarized like this:

    TAKU:  High Wood spaceeeeeee RIN:  High Metal
    BOKUSHO:  Medium Wood
    KEISU:  Medium Metal
    MOKUGYO:  Low Wood
    KANE:  Low Metal  
    (The "ŌTSUZUMI" in the last section is an hourglass drum)
    Structure SYNTHETIC
    (far left)
    (mid left)
    (middle right)
    (far right)
    67 1 High Frequencies with intermittent low patches and a brief metallic RM cloud (IV) >[BOKUSHO]<

    68 2 HF gliss and "choppy noise gliss", 2 perc. Echoes (III)

    69 3 CH. I-IV: HF w isolated notes, swells

    (accelerating echoes)

    70 4 CH I, II: Bandwidth fallling/rising while expanding/shrinking

    71 5 HF waves (I), intermittent low delta generator
    pulse bursts (III)



    (South Sahara):
    Song w "kalimba"
    72 6 CH I-IV: slow falling gliss with some gaps

    Falling gliss eventually becomes slow pulses with more gaps
    (function generator)

     w "punch"
    7.1 >[TAKU]<

    Alternating fragments of JAPAN (Gagaku),  BALI, and AFRICA, then adding SPAIN:
    73 7.2 CH. II, V: HF slowing to high pulses w 1 low punch-in >[KANE]<


    Buddhist chant (1)

    74 8 High and low drones, with various swells and bursts in between
    (function generator)


    75 9 LF RM drone, intermittent w vibrato, HF whistle, "tape scrubbing-like"
    drum & violin children's song
    Sevillanas Guitar song

    X 0:35
    76 10 CH I, II, IV: swells of wideband noise with asynchronous gaps >[BOKUSHO]<

    (echo of I)


    77 11.1 Octave filtered noise (III), low pulses changing speed (II), Kabuki RM in gaps (I)

    Shinto dance
    (3 transposed rising)

    78 12 multiple frequency fast pulses, warbling and scrubbing ("scratching") >[TAKU]<

    drum & violin children's song

    79 13

    Brazilian Indians,
    Shipibo Tribe
    "Dance for Adolescent Girls"
    80 14 Background HF


    drum and violin children's song
    (9x transposed)


    81 15 CH I-IV: 11 indiv. ascending RM "inserts" from 1, 4-9, 11, 13, 14 (Bali, Kabuki, Indian, Taku sound)
    (echoes, varying tempo)

    82 16 CH II, III: HF drone w fragments of falling/rising scales >[RIN]<

    X 0:57
    83 17 Descending RM "inserts" of 1, 4, 8, 9 (II-V), scale fragments (I) >[BOKUSHO]<

    Conch Music (lower)

    84 18 CH II-IV: brief RM/function generator fragments faded in
    Conch Music 

    (12x with falling/rising dynamic)
    85 19 staccato middle freq fragments w reverb
    (feedback echoes, ritard/accel.)

    86 20 Slow, rising TAKU pulses (V), 
    RM'ed world music fragments
    (I - II - III - IV) 


    Song w Kalimba sound

    87 21 TAKU pulses (V) begin rising and getting faster

    Peking Opera w flute

    Peking Opera w flute
    88 22 Quiet HF drone fades in and out

    Buddhist Chants
    (1 & 2)


    AMAZON (Brazil): Javahé Indians "Lullaby"

    Suyá Indians (at 0:48)
    (men's chant w rattles)

    Conch Music 


    89 23 >[TAKU]<

    Conch Music Low

    Conch Music High

    90 24 LF buzzing and distortion
    (Octave filters)

    Low creaking/scrubbing fragments


     Vietnam in brief spikes,
    Octave filter (II)

    Javahé Song reappears as an "insert" in I (from 24.II)

    AMAZON (Brazil):
    Javahé Song


    "Song of the Festival of Love" (girls) & "Concert of Bamboo Whistles" (flutes) (at 0:33)

    Shinto dance


    91 25 >[BOKUSHO]<

    (cont'd from III, less RM)
    X 0:21
    92 26 >[TAKU]<

    "Air de Gongs" (gamelan-like)

    93 27 1 brief quote from 24, low buzz VIETNAM:
    "Air of Gongs"
    (cont'd from IV) and "Concert of Flutes

    (echoes rit.)
    X 0:34
    94 28 Vietnam (I) rhythmically modulated,
    falling octave filter gliss (IV),
     slow gliss with gaps (V)

    95 29 >[RIN]<
    transposed hits w RM echoes
    transposed hits w RM echoes
    96 30
    Temple chant
    with 4 "punch" echoes

    97 31 Various overlapping HF fragments and RM gagaku (I)

    4 Large temple bells w RM (unison, then increasingly apart)

    Conch music
    (at 1:29)


    Temple chant w wooden shoes (cont'd from III), RM'ed by KANE hits
    98 32 Rising glisses (I)
    Rising and falling RM Noh fragments (II, IV, V)


    Noh Music

    (Gamelan modulated by Noh Music!)

    Noh Music

    Noh Music
    99 "FINAL SOUND" >[TAKU]< >[TAIKO]< >[TAKU]< >[TAIKO]< >[TAIKO]< 1:34

    Sound Impressions
         TELEMUSIK is without doubt one of my favorite Stockhausen works.  The melding of synthetic tones with "ancient music" is an amazingly successful one, even without the ring-modulation as "lubricant".  It's no surprise that many in the avant-garde crowd are also fans of "early music" - both are somewhat beyond our general listening experiences that they might as well come from other planets.  The ritual music of Asia is especially appropriate since it's form mostly stays away from western classical "tonic-dominant" cadences and homophonic textures.  In general, they're somewhat more modal in tonality and use heterophony as the main form structure.  This actually has a very similar flavor to the manipulation of sine-wave bandwidth tones.  Having said that, the ritual music is very modulated, so that on initial listens I didn't even know most of the time what was synthetic and what was "found sound".  After spending some time exploring world music, these fragments are much more recognizable, and at that point Stockhausen's quote about finding an "apple on the moon" becomes much more appropriate.  Another thing that strikes me as fascinating is that the "high band" layer of the RM'ed world music still has every element of the original but in a parallel "higher sphere".  Like YLEM, this brings to mind the modern cosmological theory of "stacked" parallel universes.

         It's interesting to compare this with Stockhausen's previous ring-modulation work, MIXTUR.  There, Stockhausen processed melodic and rhythmic fragments and textures, so that the RM effect is audible when an instrument is playing.  In TELEMUSIK, the samples are all rhythmic (steady and periodic) and have a generally even density.  In this way, instead of creating a melodic figure with a trombone solo, faders and "solo" buttons are used to "play" isolated bits of the prerecorded music (for example, in Structure 1 channel III and IV).

         The score is absolutely fascinating to study.  Stockhausen  does everything he can to provide instructions on how to create TELEMUSIK on one's own.  The step-by-step instructions and photographs of the equipment used give a rare look at exactly how electronic music was created in the mid-1960s (at least by Stockhausen).  The copy at my local library was only in German, but the thick Stockhausen Edition CD booklet gives enough information that it's basically decipherable to an English speaker.  The Text CD with Stockhausen's lecture on TELEMUSIK and extended samples of the world music he used also includes a newly remastered transfer of TELEMUSIK from the original stereo mixdown.

    TELEMUSIK Sound samples, tracks listings and CD ordering 
    Stockhausen's 1966 Lecture on TELEMUSIK - CD 16 (with Sound Samples and new transfer!)
    Stockhausen's 1966 Lecture on TELEMUSIK - (English translation)
    Purchase the Score 
    Stockhausen discusses TELEMUSIK in the British Lecture 1972 (Youtube clip)
    "Serial Composition, Serial Form, and Process in Karlheinz Stockhausen's Telemusik." In Electroacoustic Music: Analytical Perspectives (Jerome Kohl), ed. Thomas Licata
    Telemusik – a system of planetary order (Asbjørn Blokkum Flø)
    Analysis of TELEMUSIK (Presentation Notes, Arshia Cont)
    Creating the Gagaku Circuit on MAX/MSP
    Creating the Gagaku Circuit on MAX/MSP demo 2