AAWM Journal 13/No. 2 (2025)

Six Analyses of Kokū (虚空)

Amine Beyhom and Bruno Deschênes

shakuhachi, Japanese music, spectral analysis, video analysis of music, graphic analysis of music, spectrogram, honkyoku

This article is a sequel to four of Amine Beyhom’s previously published articles concerning the VIAMAP (Video Animated Music Analysis Project) as applied to Middle Eastern music. This time, it examines six recordings of Kokū, a well-known piece from the Japanese shakuhachi solo repertoire formerly composed by the monks of a Buddhist sect during the Edo era (1603-1868), as played by different musicians. This particular type of analysis is the first to examine the recording of a single piece performed by different shakuhachi players (four Japanese, two non-Japanese players). Three main points have emerged. Firstly, a piece is never exactly played in the same way and there can be, at times, wide variations between players. Secondly, no two musicians produce the same tuning when performing the solo pieces on any of these bamboo flutes, since no two shakuhachis can be tuned exactly the same (either of the same length or of different lengths). Thirdly, it appears that some of the techniques that are unique to the shakuhachi, thanks to the varieties of ways they can be produced, play a key role in allowing musicians to develop their own individual styles.

Amine Beyhom was the editor in chief of NEMO-Online and director of CERMAA (Centre de Recherche sur les Musiques Arabes et Apparentées / Centre for Research on Music from Arabian and Akin countries), a research center affiliated to the FOREDOFICO foundation in Lebanon.

Bruno Deschênes received his shihan (licence) in 2016 from Alcvin Ryūzen Ramos alongside a Japanese artist name, Chikushin. He studied the shakuhachi under Kurahashi Yōdō II from Kyoto and Alcvin Ryūzen Ramos from Vancouver. He is the author of the first book published in French on the shakuhachi (2017), a book on the philosophy of listening (2018), and his latest book on the phenomenology of listening (2024).

Click for DOI, citation, PDF version.

This analysis is probably the last major analytical work of Lebanese musicologist Amine Beyhom (1960–2022) before his untimely death in December 2022, at the age of 62. With a background in engineering and mathematics, he became a specialist of the musics of the Middle East, unveiling the complexities of its micro-tonalities. He was, unfortunately, unable to complete this last project. The bulk of his analysis is available here, though a few paragraphs had to be adjusted, corrected or removed here and there. I sincerely thank Lawrence Shuster and the reviewers for allowing the publication of this thorough analysis of six recordings of a well-known piece for the Japanese shakuhachi: Kokū. It is a last homage to Professor Beyhom’s outstanding analytical work.

– Bruno Deschênes, Montréal

Introduction^[1]

This article is a sequel to four previously published articles concerning the VIAMAP^[2] as applied to Middle Eastern music, and it examines six performances of the piece Kokū (虚空) played by different musicians on the Japanese shakuhachi. In June 2019, I received, as the editor of NEMO-Online, an exploration by Bruno Deschênes of how the melodic lines of the traditional and modern pieces for the shakuhachis^[3] appear to be composed. At that time, I made a few remarks to Bruno on how the pitch of a same note can vary with time and according to the intensity of blowing in the flute. Moreover, two shakuhachis with the same length, played by different performers, would produce small but effective differences in pitches, which confirmed the need for a more thorough exploration of this phenomenon.

[2] Two years later, as I was rediscovering shakuhachi performances by Japanese shakuhachi master Kitahara Kōzan,^[4] I proposed that we co-author an article with Bruno about the different versions of the same piece by different performers. Our choice fell on Kokū for many reasons, among which the – relative – simplicity of the melodic line as well as its unusual scale and flow, and the very different interpretations of this piece that we were able to find. We chose five among at least a dozen available versions of this piece, which is a panel wide enough to give a good idea about the differences in interpretation – be these in the pitch, the spectrum, or in the various techniques used by each performer – and small enough to be analyzed and handled within one single article. Later on, a sixth version, by shakuhachi master Yokoyama Katsuya, was added, as today’s most recorded versions of Kokū follow the score from his school.

[3] The article is divided into two parts. In order to understand some terms used in the analysis section, Part I presents a few technical points about the shakuhachi, including briefly introducing the piece as well as the six musicians. Part II provides detailed analyses of the six recordings and explains the general principles upon which they are based. Each recording is accompanied by a detailed video spectral analysis of each of the full audio performances, and from which the figures analyzed were extracted. The score of Kokū and its transnotation are presented with some explanations in the Appendix.

[4] This particular type of analysis is the first to examine the recording of a single piece performed by different shakuhachi players. Three main points have emerged. Firstly, a piece is never exactly played in the same way and there can be, at times, wide variations between players. Secondly, no two musicians produce the same tuning when performing the solo pieces on any of these bamboo flutes. Thirdly, it appears that some of the techniques that are unique to the shakuhachi, thanks to the varieties of ways they can be produced, play a key role in allowing musicians to develop their own individual styles.

**Part I. Introducing the Shakuhachi and the Musicians**

[5] No two shakuhachi’s bamboo pieces have the same size, thickness, density of fibres, tuning and timbre.^[5] Thus, no two shakuhachis can be perfectly in tune with each other, as well as to the Western tempered scale. Comprising of only five holes, the base tones of this bamboo flute are D, F, G, A and C for the standard 1 shaku 8 shakuhachi. As to the intermediate ones, they must be produced by modifying the angle of the head and the position of the upper lip, as well as partially opening one or sometimes two holes.

[6] The width of the holes’ opening differs depending on tone, octave, and flute, especially when these are of different lengths. One tone may require an opening of as little as a millimetre on one flute, or of three or more millimetres to produce another tone. A few of the intermediate tones are soft, they cannot be played loud.

[7] This characteristic has become part of the aesthetic of the shakuhachi, though some modern schools have developed ways to produce these intermediate tones more loudly. Furthermore, it is impossible to maintain a steady tone for a long period of time. The lips and head movements, including the different vibratos used on the shakuhachi, cause the pitch of any tone to fluctuate.

[8] Additionally, by decreasing the force of the breath, the tone is lowered while, by intensifying it, it rises. The player has to take all of these factors into consideration when playing each tone. She or he must learn to coordinate the head and lips positions, the intensity of the blowing, and the opening of a hole, which can be quite a challenge when playing a large number of flutes, since each one is different.

[9] One particularity of the shakuhachi is that each flute has its own distinctive character, which particularly applies to long flutes. It is impossible for a player to develop a lips position that is applicable to all flutes, regardless of their lengths. Each player must learn to adjust her or his breathing to the flute, as well as her/his lips position. For example, the small size and small bore of a 1.2 shakuhachi will require a different lips position than a 3.3 one with a large size bamboo and a large bore.^[6]

[10] There can even be differences between flutes of the same length because of the different shape of the bamboo and the size of the bore. If a player has been playing a single standard flute (the 1.8 shakuhachi) for a number of years, whenever he tries to play a much longer flute, it might take him a few weeks to learn to adjust her or his lips position to this new flute. If, in concert, a player uses shakuhachis of different lengths, he has to warm up with each one of them before the concert.^[7]

[11] An interesting aspect of the solo pieces for shakuhachi is that they can all be played on flutes of different lengths. A score can be used for any length of shakuhachi. The notation is a tablature, and the characters used refer to fingering, not to tones or notes. There is no need to transpose it as is the case with the Western notation. For a given piece, a player might prefer a standard flute (1.8), while another might prefer a longer one (for example, a 2.1, or 3 shakus). Lately, the popularity of longer flutes has grown, in particular among non-Japanese players, as some shakuhachi makers have confirmed to Bruno Deschênes.^[8] As pointed out above, one key feature of this bamboo flute is its timbre. When performing a piece in concert, most players choose a flute not for its key (for example, a 2.1 being in B, while a 3 shakus being around E or F), but for its timbre.

[12] The question of playing in “tune” according to a particular mode is not a requirement for playing solo pieces for both modern and traditional shakuhachi, though modern schools might prefer to be closer to the Western tempered scale. The well-known Japanese shakuhachi master Yamaguchi Gōro (1933-1998) used to say that the most important thing is not to play in tune, but to make beautiful sounds.^[9]

The techniques referred to in the analyses

[13] The techniques used on the shakuhachi are diverse. Some of the most distinctive techniques that we refer to in the analyses are the following:

Nayashi ():^[10] Downward movement of the head and the upper lip followed by a return to the starting tone. This downward movement can be more or less up to half a tone. When, in the score, the nayashi follows a tone, it is heard following the attack of a tone. When inserted before a tone, the player starts the tone one half-tone lower, to bring the sound up to the tone indicated in the score.
Muraiki (群息): Explosion of air. This sound can be intense, or it can be a small effect which makes a tone hoarse and less pure. This particular sound is usually not indicated in a score, being left to the player to use it according to his style of playing.
Koro-koro (コロコロ):^[11] A tremolo produced by swinging the index and ring fingers over the first and second holes (from the bottom of the instrument). The index finger of the right hand hits the second hole while the ring finger remains on the first hole; the ring finger is then lifted to cover the first hole while the index finger remains on the second hole, and so on, alternately, and very quickly. When a hole is hit, the other hole must be closed. This tremolo produces a sound with a tone-colour unique to the shakuhachi.
Yuri (ユリ, vibrato); there are three main vibratos:
- Yoko-yuri (横ユリ): Head movement from left to right, the pitch of the note remains (nearly) constant.
- Tate-yuri (縦ユリ): Up and down head movement, the note may vary by one full tone.
- Mawashi-yuri (回しユリ): Head movement in a circle, going left to right and up and down. The tone could also vary by one tone.^[12]

The piece and the musicians

[14] Kokū (虚空) is one of three original pieces thought to have been composed sometime in the 13^th century. The other two are Kyorei (Empty bell) and Mukaiji (Flute in a Misty Sea). Kokū is usually translated as “Empty Sky”. This title’s characters (虚空) both mean empty.

[15] The character Ko (虚) refers to an absolute emptiness which cannot be expressed in words. Kū (空) can mean sky, air, the heavens or paradise.^[13]

[16] Most solo pieces for shakuhachi have more than one version. These differ in content and length. A few among them even have more than 20 different versions composed by different musicians. Some of these versions might be similar, others completely different. Some are traditional, while others have been modernized. A same piece played by different schools might have different titles. We can see this in the choice of the recordings for this analysis. Four of these recordings follow mainly the same (Yokoyama’s) version, while the versions from Nishimura Kokū and Kitahara Kōzan are, as will be shown in the analyses, different.

[15] The score in the Appendix (FHT 1) is from the school of internationally known shakuhachi master Yokoyama Katsuya, the most performed version today.^[14]

[16] A few points particular to the version of Kokū used in these analyses are worth mentioning. Although being Yokoyama Katsuya’s version, it has somehow become today’s standard version, in particular for non-Japanese shakuhachi players.^[15] Some of the phrases in his version have a large number of tones, compared to more traditional pieces (see Appendix (FHT 2—bars 39 & 43). Yokoyama accentuated a few groups of tones in the second half of the piece (indicated by an accentuated “-” over the tones in the transnotated score—see Appendix (FHT 2—bars 42, 43 & 52)). Furthermore, Yokoyama developed a particular type of trill, unique to his style (see Appendix (FHT 2—bar 54)).

Presentation of the Six Musicians

Nishimura Kokū (1915–2002)

[17] Nishimura Kokū was a master shakuhachi player, teacher, and craftsman. Like his teacher, Tani Kyochiku, Nishimura became a komusō (a mendicant shakuhachi player), wandering around Japan for ten years. Nishimura played an antiquated shakuhachi with no plaster added to the bore, unlike the modern version in which the bare is filled with plaster and lacquer. He also favoured long instruments. He decided to call this flute style kyotaku in order to differentiate it from the shorter, modern shakuhachi.^[16]

Kitahara Kōzan (1925–2015)

[18] Kitahara Kōzan II is a musician from the Tozan school of shakuhachi created by Tozan Nakao (1876-1956) in 1896.^[17] His father, Kitahara Kōzan I, who died in 1947, was one of Tozan Nakao’s first students.

Sogawa Kinya (1962–)

[19] Sogawa Kinya studied shakuhachi with Yokoyama Katsuya (1934-2010). His core focus is the Watazumido-style honkyoku solo musical tradition of the komusō wandering monks of medieval Japan. He is active in many genres, playing with orchestras, in theatre productions, and has done a large number of recordings for computer game music, including ChronoCross and Ghost of Tsushima. Sogawa Kinya is also a devoted shakuhachi maker.^[18]

Andrew MacGregor (1949–)

[20] Andrew MacGregor began his studies of the shakuhachi with Riley Lee in 1985, and with Yokoyama Katsuya in 1989. He also lived in Japan as a pupil (uchi deshi) of Tajima Tadashi in 1993. Studying with Miyata Kōhachirō from 2001, Andrew has recorded solo albums, plus collaborations with Riley Lee, guitarist Marion Schaap and Japanese koto players Yamaji Miho and Watanabe Haruko.

Alcvin Ryūzen Ramos (1969–)

[21] Alcvin Ryūzen Ramos is a teacher, performer, and maker of the shakuhachi based in the outskirts of the Sunshine Coast of BC, Canada. He teaches and makes jinashi shakuhachi flutes, fue, and tenkan overtone flutes, and leads the Shakuhachi Roots Pilgrimage to Japan every few years.^[19]

Yokoyama Katsuya (1934–2010)^[20]

[22] Yokoyama Katsuya learned the shakuhachi from his grandfather, Yokoyama Kōson, and his father, Yokoyama Ranpo. At the age of 24, after hearing Watazumi’s performance, he was so impressed by the passionate player who emphasized the spiritual aspect of playing the shakuhachi, that he became his student. Being primarily a performer, in 1961, he created the Tokyo Shakuhachi Trio with Miyata Kohachiro and Muraoka Minoru and, in 1963, the three of them helped form the Nihon Ongaku Shudan (more known as the Ensemble Nipponia) with the intent of creating a new Japanese music using traditional instruments.

[23] In November 1967, he performed at the premiere of Toru Takemitsu’s “November Steps” in New York City, alongside biwa player Tsuruta Kinshi, with the New York Philharmonic conducted by Ozawa Seiji. He has also performed with the NHK Symphony Orchestra on various occasions, including at the opening of the Sydney Opera House and at the 1974 United Nations Peace Concert.

Preliminary Comments on the Six Recordings

[24] Interestingly, among the six musicians whose recordings we have selected, the four Japanese musicians have, sometimes, very different versions from the one we have transcribed here, while the two non-Japanese musicians play this transcription with very few changes. By version, we mean each musician’s performance version.

We asked, for example, Sogawa Kinya about his version of Kokū and he answered:

“[it] is from Yokoyama Katsuya Sensei. […] among [my] Western students there is a tendency to ‘believe’ in the score that is handed to them, and be under the impression that that is the only way to play the piece, and that it must be played only that way.

[…] [A]round the bone structure of the basic melodic line, you are allowed to ornament using established conventions, like playing around tones of a raga or a maqām.

Basically, the accomplished player doesn’t play the piece the exact same way twice. The piece is alive and fluid. Yes, you imitate your teacher [as] exactly as possible at first, until you understand what the teacher is doing, and then you have to ‘break it,’ so to speak, and go on to make the piece your own.”^[21]

[25] Very generally speaking, Kitahara’s and Nishimura’s versions^[22] are clearly different from the Yokoyama school score, while Sogawa’s version is also from Yokoyama Katsuya^[23]–whose pupil he was–with no major changes from the Yokoyama school score reproduced in the Appendix. Yokoyama’s version of the performance of Kokū is, like Sogawa’s, slightly different from his own school’s score. Finally, Ramos’ and MacGregor’s versions are very close to the Yokoyama school score reproduced in the Appendix.^[24]

Part II. The Analyses

[26] We have followed a step-by-step methodology (Table 1 below) for choosing the versions of Kokū that we analyse in this article, and for the analyses as such and the redaction of the article, which took us more than eighteen months of work.

[27] In what follows, we explain the general principles of the VIAMAP studies as well as a few particularities of each performance. Then, we examine and comment on the six recordings, the basis of our analyses being Yokoyama’s score, and, obviously, the video analyses, starting with Alcvin Ryūzen Ramos’ and Andrew MacGregor’s performances, followed by Sogawa’s, Nishimura’s and Kitahara’s. We evaluate the increasing differences from Yokoyama’s version, which ends the series of analyses.

Table 1. Step-by-step methodology followed for the six analyses.

General Principles for the VIAMAP Analyses

[28] All six video analyses of Kokū follow the VIAMAP general guidelines explained in a series of articles from preceding years.^[25] The 1920 x 1080 pixels screen is divided in two horizontal stripes (Fig. 2).

[29] These cover roughly one third (for the upper stripe^[26]) and two thirds of the video (for the lower one). Both stripes show the graphic analysis but with more details in the lower, “zoomed” stripe, while the upper stripe shows an approximately threefold time-lapse of the analysis.

[30] The lower stripe of each video analysis comprises five main components: the pitch line (in dark green for instruments), the graphic scale (Fig. 1), the intensity line (in light maroon), the spectrogram (or “spectrum”) and the annotations.

[31] The annotations signal the special techniques used in shakuhachi performances (in black or purple for the muraikis, to differentiate them from other techniques with which they are frequently, and simultaneously, used), or for particulars of the analysis (in lighter green).

[32] Due to the subtle uses of intensities by the performers, the beginning (“attack”) and extinction of each “phrase” are reproduced as exactly as possible, knowing that the intensity line shows whether the beginning and end of each sound are clearly audible, or nearly inaudible. In the upper stripe, the cursor moves from the left to the right while the graphic analysis moves simultaneously from right to left.

[33] This allows for a wider view of the analysis and gives an indication about the current relative position of the piece timewise.^[27]

[34] Note that Table 2 shows a few landmarks (namely bars 4, 21-22, 46 and 52) in the performance of the four musicians who principally follow Yokoyama’s score, i.e. Yokoyama, Sogawa, Ramos and MacGregor.

Fig. 1. The graphic scale used for the six analyses.^[28]

[35] For the clarity of the graphic, and since the two stripes serve different purposes, no spectrogram is shown in the upper stripe and the scale is limited to the Tonic/Octave(s), Fourth and Fifth lines, respectively in red, green and blue.^[29]

Table 2. Landmarks for bars 4, 21-22, 46 and 59.
with MacGregor, Sogawa, Ramos and Yokoyama.

Fig. 2. One frame of the video analysis of Kokū
by Kokū Nishimura at analysis time = 0 seconds.

[36] In the lower stripe, a light blue^[30] central cursor remains still and indicates the current time of the analysis. This stripe includes the graphic mapping of the pitch (in dark green) and intensity (in light maroon) with literal or graphic annotations, as well as a graphic scale with the indication of the tonic (here chosen as b) in bold font (see Fig. 1).

[37] The scale may move vertically–without any changes in its proportions–to fit small discrepancies in the position of the tonic or of a stable, important pitch (generally the Fourth, the Fifth or the Octave), but it is kept stable in these six analyses as the tonic itself remains mostly stable.

[38] Note that the scale doesn’t follow absolute pitch but is relative in all VIAMAP analyses. This means that the most adequate division of the octave is chosen for a given part of the piece, regardless of the actual pitch of the tonic.^[31]

[39] Additionally, a reproduction of the spectrogram of the emitted sound in the background of the graphic analysis shows the distribution of its harmonic components, with the limits shown at the beginning of the analysis. (In the particular case of Nishimura’s analysis in Fig. 2, the spectrogram shows frequencies between 0 and 5,000 Hz.)

[40] The lettering of the notes of the scale changes with each octave from lower to upper case and vice versa, with the main octave notes being all in lower case.

[41] Finally, a horizontal VU-meter is added as a supplementary aid for the determination of the intensity at t-time.

Determining the Tonic and the Reference Pitch (the upper Fourth)

[42] In the particular case of Kokū, we initially hesitated between the relative tonic being e or b; a tonic on e would have provided the advantage of having both Just Fourth and Fifth in the scale. However, the score as well as the analyses of the six performances show that the fifth degree is generally one semi-tone below the Just Fifth, while the Just Fifth (corresponding to f# or F# for the scale on tonic b) as such is only used in modulations or transpositions such as at 389-396 and 399-404 s_a^[32] (Fig. 3) in Nishimura’s performance, then stabilizes at the (lower) F at 432-434 s_a. Thus the choice of b as a relative tonic is justified by both performances and notation.

[43] Another particularity in Kokū is the initial call of Fourth from b to e (from 2 to 10 s_a in Fig. 6). In the case of the performance of Nishimura, for example, this call of Fourth is repeated once (13-22 s_a) then followed by a note one semi-tone higher (24-33 s_a, f) and a one-whole-tone rise to the tonic (36-44 s_a, from A to b).

[44] As the first Fourth is an essential note in the piece, it seemed relevant to have an approximate measurement of its pitch in comparison with the tonic b. Fig. 4 shows an example for measuring the pitch of the tonic in the case of Andrew MacGregor while Table 4 shows the measurements and the times for the tonics of the six performers.

[45] We can notice the differences in the tonics of Nishimura Kokū (368 Hz)–who evidently uses a longer shakuhachi–and Kitahara Kōzan (589 Hz)–who clearly uses quite a smaller instrument – while the tonic of Yokoyama (468 Hz) is intermediate between Kitahara’s tonic (589 Hz) and the tonics of the three other performers, which are quite close to one another (around 410 Hz).

[46] Provided that the performers use the same fingerings and that the rule of lengthening of the shakuhachi 1 sun per one semi-tone lower tonic applies, and considering that Kitahara uses a standard D 1.8 instrument, MacGregor and Ramos are using a larger–nearly three tones lower, i.e. 2.4–instrument, with Sogawa using a still slightly longer (2.4-2.5) shakuhachi while Nishimura’s instrument should be 8 sun (813 cents difference in the measured tonics) longer (and 4 whole-tones lower), which corresponds to a 2.6 shaku instrument.

[47] According to Bruno Deschênes, Kitahara uses a standard 1.8 shakuhachi, while Yokoyama uses a 2.2 shakuhachi and Nishimura a long, 2.7-2.8 one. Ramos uses indeed a middle-length (2.4) shakuhachi,^[33] while MacGregor uses a 2.55 instrument^[34] and Sogawa a 2.5 one^[35].

[48] This shows that the “rule” of lengthening the instrument by 1 sun for 1 semi-tone is an approximation, and that the exact tuning of the tonic of the instrument depends on characteristics other than the mere length.

Fig. 3. Part of a frame showing the video analysis of Kokū by Nishimura Kokū between 364 and 454 s_a. The f# is clearly discernible at 389-394, 399-404 and around 427, as well as the return to f plain^[36] at 395 and 405, and to F plain at 432-434 s_a. Notice the symmetry around the central modulation (401 s_a), preceded by 4 nayashis then a muraiki, and followed by a muraiki then by 4 nayashis.

Table 3. Measurements of the pitch of the tonics for the six performers. The lengths of the shakuhachis are calculated with the “one sun longer = one semi-tone lower tonic” rule.

Fig. 4. Determining the pitch of the tonic with Praat for Andrew MacGregor’s performance. With slightly varying tonics such as this one, the analyst generally chooses a stable, medium placed segment and asks the program to compute the mean of the pitch on this segment.^[37]

Table 4. Measurements of pitch and corresponding times for both “Reference Note” (the initial e) and the following tonic b; all measurements are rounded to the 2^nd decimal place. Note that for Andrew MacGregor the tonic is very stable and was measured on a very short time lapse (0.29 seconds), while it was measured at 40.29 s_a (for “Seconds of the Analysis”, to differentiate it from the video analysis time) for Alcvin Ryūzen Ramos.

[49] Another difference between the six performers lies in the measurement of the Fourth between e and b (Table 4 above, last column to the right). With Nishimura (495 c), Kitahara Kōzan (504 c) and (to a lesser extent with) Yokoyama Katsuya (507 c), it is quite close to the Just Fourth (498 c), while with the three other performers the fourth is slightly exaggerated (between 518 and 522 c), approximately one comma^[38] greater than the Just Fourth.

[50] This is one first indication of varying performances of the same musical piece in the shakuhachi repertoire. The detailed analyses will show even more particularities for each performer playing the same piece with the same notation.^[39]

[51] Note that, in order to avoid too many repetitions in the six video analyses, a single introductory video was added explaining most of the generalities of the analyses listed in the sections above.

Six Analyses of Kokū (虚空)

Preliminary Remarks

[52] Each analysis goes from the general (the correspondence with the score) to the particulars of each performance, commenting on the performances the performances of some key points in the score, such as:

The micro-modulations at bars–also called “phrases” in the analyses, as mentioned above–8, 19, 30, 43, 52, 60 (upper f#) and bar 46 (lower F#).
The groups of nayashis–not performed by everybody–for example at bars 2, 7, 14, 18, 25, 36, 60 and 69.
The four koro-koros, with two incorporating breath pauses at bars 21-22 and 65-66, and two others (both with no breath technique) at bars 41 and 50.
The repetitions at bars 35, 59 and 61, knowing that it is not necessary to perform them exactly as written.^[40]
The accentuations (notes are underlined or highlighted) in bars 42, 43 and 52.
The sizes and particularities of the semi-tones e-f, E-F and B-c.
The portamento from E to G of bar 46, with successive transitions through E F F# G and back to E through F# and F.

[53] Other peculiarities, which characterize more specifically one of the performers, are also highlighted in the following analyses.

[54] Note also that the determination of the different yuri (vibrato) types was made both by ear and by determining the corresponding graphic line. However, each performer develops particular lips techniques. These allow him to perform different techniques which are, sometimes, very close to a certain type of yuri (or to another technique).

[55] In such cases, and for the four performers following mostly Yokoyama’s score, we opted for mawashi-yuri–when it was found in the score at the corresponding bar–rather than, for example, for tate-yuri.^[41]

**Analysis of Alcvin Ryūzen Ramos’ performance of Kokū^[42]**

[56] Preliminary remarks on Ramos’ performance: please note a possible nayashi at 38.3 s_a, which, however, corresponds to the suite of three notes f e f in measure 10 in the transcription of Furuya’s score (FHT 2).

[57] The first school of shakuhachi in which Alcvin Ryūzen Ramos studied was the Yokoyama Katsuya school.^[43] He studied under Kakizakai Kaoru, one of the most prominent students of Yokoyama.

[58] Compared to Ramos’ performance with Yokoyama’s version (FHT 2 and FHT 3), Ramos did not make major changes to his playing of Kokū. He slightly altered a few phrases. In phrase 24 (232-242 s_a–with “s_a” meaning “seconds of the analysis” to differentiate the analysis time from the video time), he adds a second mawashi-yuri (large vibrato–see Fig. 5) to the phrase.

[59] In phrase 30 (285-295 s_a), he repeats some tones toward the end of the phrase. As explained by Sogawa Kinya,^[44] such type of repetitions is common in all schools of shakuhachi. Students (as well as shakuhachi players)^[45] do not have to play exactly the phrases as they are written. They are allowed to alter them slightly.

[60] At phrase 53 (461-468 s_a), he adds again a second mawashi-yuri. He appears to link phrases 65 and 66 (582-589 s_a) into a single phrase. On a few occasions, Ramos alters a tone. For example, in phrase 3 (37-47 s_a), the score indicates that the performer should play an F followed by a sliding down to an E. Ramos does not do it and inserts a nayashi instead.

[61] Otherwise, his performance respects almost entirely Yokoyama’s version transcribed in FHT 2 and FHT 3.

[62] A particular of Ramos’ performance is the use of very high harmonics in some passages, which is why an extended analysis (to 8 000 Hz, whereas for the other musicians it was limited to 5 000 Hz) of the spectrum was undertaken. This translates, for example, into markedly separated border frequencies at the end of the two first phrases with a nearly empty spectrum in between (Fig. 2). This is a “hissing” technique, which seems to have been developed on purpose by the musician.

[63] We also noted a tendency to accentuate the muraikis,^[46] with some of them being very strong (and stretch for a whole second such as at 12–13 s_a–see Fig. 7). The spectrum is clearly “scrambled” at such passages, highlighting the inharmonicity of that sound.

[64] Similarly, Ramos’ groups of nayashis are very strong sometimes, sounding as if they were replaced by repetitions of d and e, and more or less resembling a mawashi-yuri (244–249 s_a, Fig. 10). At 76–80 s_a (bar 7–Fig. 8). However, he actually plays the nayashis.

Fig. 5. Extract from a frame of the video analysis of Alcvin Ryūzen Ramos’ performance of Kokū showing the added mawashi-yuri at 232–242 s_a with considerable changes of intensity, surrounded by a festival of muraikis and nayashis.

[65] The koro-koros are accented, with a clear technique while including sometimes a muraiki (see one example in Fig. 9). We found also many repetitions of notes such as at 279.1–380.2, 473.3–474, 510.4–511, 524.8–526.9 and 546.9–547.8 s_a, plus one with accentuations at 327.2–330 s_a (and preceded by a muraiki at 327 s_a).

Fig. 6. In phrase 3 (37–47 s_a), the score indicates that the performer should play an F followed by sliding down to an E, an indication omitted by Ramos. (Note the nayashi at 38.5 s_a.)

[66] Accentuations are always present, with or without tempo accentuations. The accentuations of all the descending degrees f# e c b A F E can be found at 385.6–387 (and until the ascending b around 389 s_a) and at 457.3–458.7 s_a (similarly, until the ascending b around 461 s_a), with a strongly accented mawashi-yuri at 470.2–473.2 s_a, while the portamentos are very rare (such as the three at 352-353.2, 486.2–487.2 and 507.2-–08.2 s_a).

Fig. 7. Extract from a frame of the video analysis of Alcvin Ryūzen Ramos’ performance of Kokū showing the gradual extinction of the intermediate frequencies between the fundamental and the highest harmonics towards the end of the second phrase (12–25 s_a), in complete contrast to the scrambled spectrum (notably for the higher harmonics) of the initial muraikis in the same phrase.

[67] The portamento in micro-modulation of bar 46 (at 407.2–410.7 s_a) omits the F in both ascent and descent (Fig. 11). (We also noted a tendency for yuri or portamento when performing the semi-tones E F and b c.)

Fig. 8. Extract from a frame of the video analysis of Alcvin Ryūzen Ramos’ performance of Kokū showing the actual performance of a series of nayashis (76–80 s_a), with a clear difference between the changes of intensity in the corresponding passage (244–248 s_a) in Fig. 5. (The nayashis are more regular, in both pitch and intensity, at 76–80 s_a.)

[68] We shall also note a non-systematic tendency to perform small half-tones between f and e, such as at 498–504 s_a or at 518–521 s_a (Fig. 12–with a clear quarter-tone transposition as at 522.2–529 s_a, but also at 531.5–533.5 and 286–291.7 s_a à Contributing to a high f# in modulation), with quarter-tonal “play” around the b such as at 203–207 s_a.

[69] As for the micro-modulations to f#, they are all played higher than the (equal-) tempered f# and occur at 90.5–91.5 s_a (bar 8), 202–203 s_a (bar 19), 291–292 s_a (bar 30), 385.8 s_a (bar 43), 457.5 s_a (bar 52) and finally at approximately 564.5 s_a, with a near-g (bar 60–Fig. 10).

Fig. 9. The last koro-koro in Ramos’ performance (at 578–587 s_a–corresponding to bars 65–66 in the transcription of FHT 3), with a muraiki at the end of bar 65 (corresponding to 579.5–581.5 s_a).

Fig. 10. The micro-modulation corresponding to bar 60 in the transcription is performed very high (a near g) by Ramos at approximately 564.5 s_a.

Fig. 11. Ramos’ micro-modulation in portamento from f# to g (and back) at 407.2–410.7 s_a, beginning with a low f# (407.2 s_a) and ending (410.7 s_a) with a high one.

Fig. 12 Small semi-tones in Ramos’ performance at 498–504 and 518–521 s_a, with a clear quarter-tone transposition at 522–529 s_a (and 531.5–533.5 s_a).

**Analysis of Andrew MacGregor’s performance of Kokū ^[47]^,^[48]**

[69] Preliminary remarks on MacGregor’s performance: please note that MacGregor is probably performing a nayashi at 471.5 s_a, while the nayashi box around 248 s_a could be extended to 249 s_a.

[70] Andrew MacGregor studied mainly under Tajima Tadashi,^[49] who was also a pupil of Yokoyama Katsuya. Playing Tadashi’s version as he was taught,^[50] MacGregor’s recording shows a few more changes than Ramos’.

[71] In phrase 3 (38 to 51 s_a), similarly to Ramos, MacGregor does not perform the sliding head movement from F to E. In fact, he does not perform it for all occurrences in the entire piece. In phrase 8 (119–136 s_a) he does not play the C appoggiatura and E. In phrase 9 (137–155 s_a), he adds a nayashi after the last tone. In phrase 10 (157–167 s_a), he does not play the 3^rd tone, B, and at the end, he repeats many times the appoggiatura before the last B. As mentioned above, this is common to all schools of shakuhachi, and includes repeating a tone several times, or adding a nayashi. In phrase 11 (167–176 s_a), he also does not play a few notes. In phrase 23 (321–327 s_a), he does not play the last tone. At the beginning of phrase 27 (381.5 s_a), MacGregor adds a tone (the E). And at the end of phrase 28 (389–393 s_a), he does not play as many repetitions of A as indicated. A first change occurs at phrases 29 and 30 (394–408 s_a), combining and shortening both into a single phrase. At phrase 32 (417–423 s_a), he also adds a tone, a G, after the 2 Es. He also alters and merges phrases 34 and 35 (431–442 s_a). Phrases 39 and 40 (463–476 s_a) are combined into one phrase. At the beginning of phrase 43, he adds two tones, while combining it with phrase 44 (488–498 s_a). Phrases 47 and 48 (515–525 s_a) are also combined into one phrase, as well as phrases 52 and 53 (544–555 s_a).

[72] Phrase 54 (555.1–556.5 s_a) is performed differently (it should be a mawashi-yuri with repetitions). In phrase 63 (651–666 s_a), he does not play the two middle tones, while phrase 67 (709.5–712 s_a) is different from the score. These are the most noticeable differences from Yokoyama’s score.

[73] We can note additionally that MacGregor has a very clear technique. He does not perform suites of nayashis (or pairs of them as with Nishimura), but only single nayashis. Notably, he is the only musician among the six who performs a rare, stretched upper nayashi around 248 s_a (Fig. 13).

Fig. 13. A rare nayashi in the upper range of the shakuhachi.

[74] Note at 157 s_a an accentuation which could be perceived as a muraiki. Another muraiki at 613.4 takes place within a series of accentuations at 613–614 s_a, with repetitions of notes at 165.4–166.4, 234–236.6 and 427.8–428.3 s_a, as well as a repetition with accentuation at 435.7–436.7 s_a (which is characteristic of MacGregor’s performance), and the aforementioned 613–614 s_a with muraiki at 613.4 s_a.

[75] As for the yuris, we shall note three absent (when comparing the performance with Yokoyama’s score) mawashi-yuris at 535.8–536.5 s_a (corresponding to bar 50, and replaced by a tate-yuri–Fig. 14), at 552-553 s_a (corresponding to bar 53), and at 709.4–711.9 s_a (bar 68).^[51]We can also note a mawashi-yuri at 176–179.2 s_a, with a soft muraiki which stops at 178.4 s_a (Fig. 15), as well as a short mawashi-yuri at 482.6–483.4 s-a.

Fig. 14. A tate-yuri (replacing the mawashi-yuri in Yokoyama’s score) around 536.2 s_a.

Fig. 15. A muraiki within a mawashi-yuri around 178 s_a, followed (and preceded) by a nayashi.

[76] We can also note one single discernible yoko-yuri at 610.2–611.2 s_a. (MacGregor seems to be the only performer with Kitahara to do it. As no yoko-yuri is indicated in the score, this is a decision of the performer.)

[77] MacGregor’s koro-koros (four as in Yokoyama’s score) are mostly regular and fitted on the graphic equal-tempered scale, with two long ones at 315–327 and 687–699 s_a (Fig. 16), one short one at 475–479 s_a, and an even shorter, last one at 532.7–534.3 s_a (Fig. 17).

Fig. 16. A long koro-koro at 687-699 s_a.

Fig. 17. A very short koro-koro at 532.7-534.3 s_a.

[78] With regard to semi-tones, and while they are mostly performed on the equal-tempered scale by the musician (especially the e f semi-tones–for example around 264, 279 and 284 s_a–Fig. 18), they are rather small for B C at 237.4-249 s_a, for c b around 292 or 498 s_a (in this case within a mawashi-yuri–Fig. 19), and still rather small around 662, 683 and 713.5 s_a (in the form of c b c), and reaching an approximate quarter-tone c b⁺ c around 433.5 s_a.

[79] Furthermore, and unlike Kitahara or Nishimura, MacGregor uses only a few portamentos (and no initial nayashis with portamento at all), with some exceptions such as (descending) around 460.2–460.6, 682.9-685 and 712–714 s_a, and a distinctive, very rare long and ascending portamento in the upper octave at 619.2–621 s_a (Fig. 20).

Fig. 18. “Regular” (nearly equal-tempered) e f semi-tones around 264, 279 and 284 s_a.

[80] Even the E F semi-tones can be rather small around 424.5 s_a, with a descending F^– (F minus one quarter-tone, or F^{“half-flat”}) E around 469–472, 502.5 (Fig. 19), 502.5 and 522 s_a, and ascending around 466 and 518.5 s_a, etc. He does not, however, perform the portamento from E to G (and back to E) of bar 46.

[81] MacGregor has generally a decreasing tempo in his accentuations for bars 42, 43 and 52, plus some tempo accentuations. These accentuations occur at 419.8–420.8 s_a (mostly on the A–and not in Yokoyama’s score), 492.3–493.7 s_a (mostly b A and F), and the whole descent c b A F at 548.3–549.5 s_a–a particularity of his performance among the six analyzed in this article.

Fig. 19. A small c b semi-tone within a mawashi-yuri around 498 s_a with an F-E around 503 s_a.

Fig. 20. Long, very rare ascending portamento in the upper octave at 619.2-621 s_a.

[82] One other accentuation is clearly discernible at 435.7–436.7 s_a for a repetition of notes (here the b), while other repetitions can be heard at 165.5–166.3, 392.1–392.8, 427.9–428.4, 574–575, 613–614.2 and 638–640.8 s_a.

[83] Finally, and in contrast to the other five performers (except, partly, for Sogawa), MacGregor’s micro-modulations to f# are generally either low-pitched such as at 128+s_a, 401.3–403.1 (with an addition around 404.8 s_a–Fig. 21), around 492 and 547.5 s_a, or (nearly) equal-tempered at 288–289 and 658.4–659.3 s_a.

[84] Note also the beautifully performed, and rather low-pitched and accented F# around 509.2 s_a, with a small portamento at 509.3–509.9 s_a and followed by a mawashi-yuri around E⁺ (at 510.1–511.7 s_a).

[85] We shall finally note his insistence on this rather equal-tempered F# as he repeats it around 527.5 and (rather low-pitched) at 531 s_a, with semi-tonal accentuations such as a (low-pitched) G at 526.2–527 s_a and a short plain F around 528.8 s_a. (Fig. 22 shows the two F# micro-modulations surrounded by two f# micro-modulations–Note also the octavial semi-tonal accentuations at 566 and 580 s_a.) All in all, a rather neatly performed piece with many particularities which could be attributed either to the performer or also partly to his teacher Tajima Tadashi.

Fig. 21. Low-pitched micro-modulations to f# at 401.3–403.1 and around 404.8 s_a.

Fig. 22. Two micro-modulations with F# surrounded by two micro-modulations to f#.

**Analysis of Sogawa Kinya’s performance of KOKŪ ^[52]**

[86] Preliminary remarks on Sogawa’s performance: please note that with Sogawa, the muraiki at 293 s_a could be interpreted as an accentuation, while there is a possibility for another muraiki at 295.4 s_a, preceding the nayashi at 295.7 s_a. Sogawa Kinya, similarly to Andrew MacGregor, adds a tone on occasions, for example at the beginning of phrase 8 (73–85 s_a).

[87] On other occasions, he removes a tone, for example in phrase 9 (86–94 s_a), the tone before the last. He also does not perform the head movement found at the beginning of phrase 6 from F to E (54–64 s_a). He merges and condenses phrases 11 and 12 (105–117 s_a). On phrases 19 (184–197 s_a), 20 (198–208 s_a) and 24 (224–231 s_a), he does not play the last two notes of each of these phrases. On a few other occasions, he does remove the tone at the end of a phrase. Afterwards, he combines and condenses phrases 29 and 30 (280–292 s_a), as well as 34 and 35 (312–324 s_a), turning them into a single phrase. The repeated tones in phrase 35 are switched between the high and the low octave. He also combines phrases 39 and 40 (343–352 s_a). At the beginning of phrase 52 (434–441 s_a), he plays a few extra tones. In phrase 58 (506–510 s_a), he adds a mawashi-yuri, ending with a muraiki. Overall, Sogawa’s performance essentially respects Yokoyama Katsuya’s score. The main changes are that, on a few occasions, he combines two phrases and alters both phrases along the way. He also includes small muraikis here and there in the piece.

[88] As a first additional remark on Sogawa’s performance, his pitches are generally well adjusted to the equal-temperament graphic scales,^[53] notwithstanding his nearly systematic use of small semi-tones e f^— in ascending or descending portamento (Fig. 23) from the very beginning (see 21–24 and 54–60 s_a, etc.)

Fig. 23. Semi-tones e f^— in ascending or (rapidly) descending portamento with Sogawa.

[89] Semi-tones c b (descending or ascending) are also small (for example at 81–85 or 119–124 s_a–Fig. 24), with the c slightly lower than its equal-temperament counterpart.

[90] He seems to favour muraikis over the nayashis in the series of nayashis on e or E of bars 2, 7, 14, 18, etc., as can be seen in the first supposed series at 11–16 s_a (Fig. 25)–the second series for bar 4 being absent.

Fig. 24. Small semi-tones b c b with Sogawa in Kokū.

Fig. 25. First configuration of muraikis (replacing the series of nayashis in Yokoyama’s score) with an initial one, then three close one to another, then a concluding nayashi for bar 2 at 11-20 s_a.

[91] But we find the same configuration (one initial muraiki, then three muraikis instead of nayashis, followed by a nayashi) for bar 7 at 65–71 s_a (Fig. 26) and for others, including in the penultimate bar 69 (Fig. 27–607.5-619 s_a).

Fig. 26. Second, typical, configuration of muraikis (replacing the series of nayashis) with an initial one, then three close one to another, then a concluding nayashi for bar 7 at 65-71 s_a.

Fig. 27. Last typical configuration of muraikis on E in the penultimate bar 69 at 607.5–619 s_a.

[92] Sogawa performs–as in Yokoyama’s score–four korokoros which take place at 353.7–357 and 419.9–425.9 s_a (Fig. 28–both rather short, regular, and based on the equal-tempered scale), and at 209.7–217.1 s_a (with a mawashi-yuri within at 215.7–216.7 s_a–Fig. 29) and 582.1–590.3 s_a (both regular, long, also based on the equal-tempered scale).

Fig. 28. Short koro-koro by Sogawa at 353.7–357 s_a, based on the equal-tempered A.

Fig. 29. Long, regular koro-koro (based on the equal-tempered A b) at 209.7–217.1 s_a, with a mawashi-yuri within at 215.7–216.7 s_a.

[93] He performs repetitions rather frequently–with different and sometimes accelerating tempos–such as the ones at 103.7–104.3, 278.3–279.9 ( Fig. 30), 476.5–476.9, 500.1–500.4 s_a, between the four muraikis at 511–515 s_a, and at 538.8–541 s_a. The last repetition (Fig. 31) is long with an acceleration, and finishes in portamento, which seems to be a technique particular to this musician–at least among the six performances we are examining here.

Fig. 30. Repetition of notes–with an accelerating tempo–at 278.3–279.9 s_a.

[94] Other notable portamentos occur, as aforementioned, systematically between e and f, but also at 365.8–367, 417.2–419.7 (slight, descending then ascending), and at 518.8–521.1 s_a.

Fig. 31. The last repetition of notes by Sogawa (at 538.8–541 s_a) is long with an acceleration, and finishes in portamento, a technique which is particular to this musician among the six performances we are examining in this article.

[95] Accentuations of notes are rare–notwithstanding the frequent muraikis performed by the musician–including the descending phrase from e at 303.2–305.6 s_a (note the absence here of the f# preceding the e) and at 439–440.3 c b A F (with an f# preceding the e). There is no accentuation (there is, however, an accelerated tempo) for the descending phrase f# e c b A E at 369–371.5 s_a. Note also rare quarter-tone accentuations such as the one at 282.7–286 s_a (Fig. 32), within the semi-tone e f.

Fig. 32. A quarter-tone accentuation at 282.7-286 s_a, within the semi-tone e f.

[96] The micro modulations to f# occur around 81 (low f#) and 193.5 s_a (very low f#) and at 286.3–290.8 (variable, as can be seen in Fig. 33, with portamento for the second f#–which is a particularity of Sogawa’s performance).

Fig. 33. Micro modulation to f# at 286.3–290.8 s_a.

[97] They can also be found around 370 (low), at 389-392.5 with an F# with portamento (Fig. 37) to less than G followed by a micro-modulation to F# at 411.5 (with semi-tonal/tonal accentuations around F# to G and E), at 437 (very high–a near g), with a second one around 438.5 in small portamento (still in Fig. 37–another characteristic of Sogawa’s performance),^[54] followed by the aforementioned c b A F accentuation at 439-440.3, and finally a nearly equal-tempered f# around 563.5 s_a.

[98] The semi-tones e f e, B C B and b c b are nearly systematically small (with lowered f) with frequent, non-systematic portamentos. They are, in particular, frequently in mawashi-yuri for the b c semi-tone (at approximately 228–231, 314–316, 374–376, 445–448 and finally at 603.2–607.2, but also sometimes with f e, as at 506.5–508.5 s_a (Fig. 34).

Fig. 34 A mawashi-yuri within the semi-tone e f at 506.5–508.5 s_a.

[99] Additionally, note that at 328.1-330.2, there exist–not shown in the graph–quick alternations between e and E (octave changes). The whole phrase encompassing these octave changes (325.8–330 s_a–Fig. 35) is particularly rich in mixed techniques, with two muraikis at 326 and 328.2 s_a, four muraikis within a possible–but not indicated in the graph–mawashi-yuri^[55]⁹⁰ at 328.2–329.2 followed by another mawashi-yuri at 331–331.9 s_a, then by a nayashi at 333 s_a.

[100] Conversely, the mawashi-yuri at 313.9–316.2 s_a sounds like a mawashi-yuri but the head movement by Sogawa in the YouTube video suggests a tate-yuri, if not a series of nayashis. The same happens at 374.2–376.3 and 445–447.2 s_a. Moreover, the mawashi-yuri at 392.9–394.7 s_a (Fig. 37) is even more ambiguous as Sogawa’s movement of the head in the YouTube video^[56] is so minimal that it is not possible to discern a clear technique.

[101] In all these instances, we decided to continue identifying the technique as a mawashi-yuri as it is what the sounding of the instrument, as well as the graphic line, suggest.

[102] Finally, all yuris were identified through the graphic line and by hearing the pitch. When compared with the YouTube video of Kokū by Sogawa, the technique of the proposed tate-yuri at 227.5–230.5 s_a (122) is, in fact, a lateral movement of the head and a movement of the upper lip, which would indicate a mawashi-yuri. We maintained, however, the tate-yuri in this case.

Fig. 35. A phrase encompassing octave changes at 325.8–330 s_a, particularly rich in mixed techniques, with two muraikis at 326 and 328.2 s_a, four muraikis within a possible–but not indicated in the graph–mawashi-yuri at 328.2–329.2 s_a. followed by (another) mawashi-yuri at 331–331.9 s_a, then by a nayashi at 333 s_a.

Fig. 36. The tate-yuri at 227.5–230.5 s_a corresponds, in the YouTube video of Sogawa, to a lateral movement of the head and a movement of the upper lip, which would indicate a mawashi-yuri. We maintained, however, the tate-yuri in this case.

Fig. 37. Micro-modulation at 389–392.5 to an F# with portamento to less than G followed by a micro-modulation to F# at 411.5 (with semi-tonal/tonal accentuations around F# to G and E) and at 437 s_a (very high–a near g), with a second one around 438.5 in small portamento–another characteristic of Sogawa’s performance. The mawashi-yuri at 392.9–394.7 s_a is ambiguous as Sogawa’s movement of the head in the YouTube video is so minimal that it is not possible to discern a clear technique.

**Analysis of Nishimura Kokū’s performance of Kokū ^[57]**

[103] Nishimura Kokū’s version of Kokū was recorded long before Yokoyama Katsuya published his score, and is quite different from and much longer than Yokoyama’s. The first few phrases are quite similar, although the first three present slight differences (e.g., adding a nayashi, or removing a few tones), like most of the other phrases that follow. Phrase (bar) 4 is absent. Phrase 7 (62–81 s_a) becomes two phrases, similar to phrases 1 and 2. As with phrase 7, he also splits phrase 14 (159–178 s_a), 18 (215–235 s_a) and 25 (307–335 s_a) in two. Phrase 20 (247–270 s_a) is also split in two, while removing two tones. The koro-koros of phrases 21 and 22 (272–283 s_a) are totally different from what the three other musicians do–much softer and calmer, furthermore becoming a single phrase. Phrase 28 is absent, while phrase 29 (387–396 s_a) is also split in two. From phrase 32 (from 409 s_a) on, he does not follow Yokoyama’s version.

[104] We hear some similar phrases, but in a different order, as well as adding a few of his own not found in Yokoyama’s version (e.g., from 498 to 519 s_a, from 570 to 586 s_a, or from 523 to 629 s_a). As for the two last phrases, they are the same. As before, he splits phrase 69 (964–994 s_a) in two, then plays phrase 70 (997–1018 s_a). Note that, on a few occasions for this performer, we indicate a muraiki, though we do not know if they are intentional or not. Traditionally, a musician would not produce them on purpose, though they could occur.^[58]

[105] We can formulate a few additional remarks for Nishimura’s performance. He does not perform repetitions of tones, his pitches being sustained instead (see for example Fig. 44 representing the sixth phrase at 62–70 s_a, in which the intensity of the sustained pitch with hisses sinks rather regularly, with a stable spectrum), and that he also uses–as with Ramos–hisses at the end of some phrases, but with a rather regular spectrogram all over the piece (except for muraikis).^[59]

[106] Instead of repetitions of nayashis (as in phrases 2 and 7) he contents himself with two nayashis as shown in the same figure at 12–22 and 72–81 s_a, and uses a few quarter-tone accentuations of a phrase. Such example is also shown in Fig. 17, in which the fifth (50–60 s_a) and eighth (81–92 s_a) phrases are repetitions of the third phrase (24–33 s_a) with quarter-tone (upper) accentuation.^[60]

[107] Similar accentuations become effective transpositions, like the one at 488-497 s_a, which includes a micro-modulation to f# (Fig. 39), and at 622–639 s_a (Fig. 46).

[108] Concerning Nishimura’s nayashis, we notice that he is, with Kitahara (next analysis), the only performer(s) among the six who use(s) initial nayashis (before the note) in (slight) portamento, beginning an approximate semi-tone–or even one whole-tone–below the intended pitch. (See for example Fig. 40, around 434.5 s_a, but also at 434.3–435, 738.8–739.2, 772.1–772.4 and 863.5–864.2 s_a.)

[109] One particularly wide nayashi by Nishimura is situated at 612.3 s_a (Fig. 41): it is notable that such nayashis are performed with the second hole from the bottom swiftly covered by the finger.

[110] As for the koro-koros, and if we include the very peculiar near-koro-koro in micro-modulation to f# at 814–823 s_a (Fig. 38), the four from the transcription are present in this musician’s performance, and are particularly slow and soft (Fig. 42). Furthermore, Nishimura’s use of micro-modulation in his performance is noteworthy if compared to the five other musicians’ performances. Firstly, he rarely uses very short, nearly imperceptible modulations to generally high f# (such as the one at 568.7 s_a–Fig. 45, with examples in the same figure of longer-lasting modulations). His modulations are more accented and sometimes last a considerable time (such as the one at 398.5–406 s_a–see below). More particularly, his use of an additional micro-modulation to c# ( Fig. 45, at 584–586s_a ), clearly linked to the two surrounding micro-modulations to f# (at 568.7 s_a) and F# (at 595–597 s_a), distinguishes his performance from the the other five performances. A similar passage, without, however, the last F#, occurs at 691-710 s_a, with a nearly imperceptible f# at 691 s_a, and a c# at 707.5–710 s_a.

Fig. 38. The near koro-koro at 814-824 s_a in direct modulation, with a soft muraiki, to f#, the fourth modulation in a series of four beginning at 784 s_a, with a perceptible, lower, f# at the end, immediately followed by a call of Fifth A e.

[111] Note also the two micro-modulations to f# in the phrase at 387–396, both of them in portamento, a particularity that we noticed with Sogawa Kinya (see Fig. 23).

[112] Nishimura’s semi-tones are mostly equal-tempered except during the frequent quarter-tonal accentuations (with portamento) of the f, while the semi-tones E F are frequently in portamento or yuri.

[113] In another example of long-lasting f# (Fig. 43, 398.5–406 s_a), the ending f# (same figure, approx. 405–406 s_a) seems, at a first listening, to be higher in pitch than the preceding f# (399–404 s_a), in contrast to the graphic statement.

[114] A thorough and repeated listening to the extract shows, however, that the real pitch of this second f# is similar to the former one, and only very slightly higher, but that its lower intensity changes its perception.

[115] A few other modulations to f# are distinguishable at 490–494 s-a, with a muraiki in f# at 751 s_a lasting till 761 s_a. At 784–825 s_a, a suite of four phrases beginning with a soft muraiki in f# from the very beginning. In the fourth phrase–the aforementioned near-koro-koro^[61] at 814–824 s_a–we can notice a perceptibly lower–and most probably consciously performed by the musician–f# at 823–824 s_a, immediately followed by a call of Fifth A e at 826.2 s_a.

Fig. 39. A phrase in quarter-tone (upper) transposition at 488–497 s_a, with a central micro-modulation to f# and a nayashi.

Fig. 40. One example of initial nayashi with portamento, beginning a whole-tone below the intended pitch.

Fig. 41. A particularly wide (vertically) nayashi at 612.3 s_a, performed by Nishimura.

Fig. 42. One example of slow and soft koro-koro by Nishimura, at 640–650 s_a.

Fig. 43. Long micro-modulation to f# (398-403.5 s_a) with a reprise at 405–406 s_a which, because of the considerable difference in intensity, seems much higher. (Sound before 398 s_a is one octave higher.)

[116] One other micro-modulation to f# can be observed at 847–857 s_a, with the last one occurring at 900.5-901.5 s_a, which considerably raises the number of modulations by this musician when compared to the performances of the five other musicians (and to Yokoyama’s score).

[117] Comparatively, Nishimura uses rare mawashi-yuris, namely three at 430–432, 574–575.5, and 696.5–698.5 s_a. His accentuations are generally limited to the two descending b A notes in the descent f# c b A corresponding to bar 43 in the transcription of Yokoyama’s score, and as can be noticed at 425.8–428.3, 570.7–572.8 and 693.5–696 s_a..

[118] As for the portamento of bar 46, it may be taking place, in ascent only, at 630.4–632.2 s_a (Fig. 45 and Fig. 46) within a phrase in (upper) quarter-tone transposition.^[62] All in all, a considerably longer performance filled with subtle accentuations and an array of techniques, a real pleasure for the aficionados.

Fig. 44. Extract from a frame towards the beginning of Nishimura’s performance of Kokū showing many techniques within one minute and a half, such as sustained notes at 62-70 s_a, quarter-tone accentuations of phrases (here of Phrase 3) at 50–60 and 81–89 s_a, and the use of a pair of nayashis instead of repetitions of those (in the transcription of Yokoyama’s score) around 17 and 78 s_a, notwithstanding the three muraikis. The spectrogram remains mostly homogeneous, partly due to the date of the recording, with much background noise and hiss.

Fig. 45. Series of three linked micro-modulations to f# (very short, nearly imperceptible at 568.7 s_a), then longer micro-modulations to c^# (at 584–586 s_a) and to the lower F# (at 595–597 s_a), surrounded and interspersed with nayashis, a mawashi-yuri, a phrase in quarter-tone transposition as well as a closing koro-koro.

Fig. 46. A phrase in quarter-tone (upper) transposition, with a long portamento at 630.4–632.2 s_a.

**Analysis of Kitahara Kōzan’s performance of Kokū ^[63]**

[119] The version heard in Kitahara Kōzan’s recording is a different and shorter version compared to the four other recordings. We do not know which version it is, and who is the originator. What we do know is that his performance is prior to the publication of Yokoyama Katsuya’s score. Right from the start, Kitahara does something that the others do not do. We hear three tones instead of the two tones indicated on the score. This technique is never noted and is performed by some more traditional schools. For example, the school of Yokoyama Katsuya does not apply it (except for one piece, Shika no tōne, for two shakuhachis: it is composed in such a way that a musician can play with other musicians from other schools). The technique is simple: when playing tsu-re (D-E on a standard 1.8 shakuhachi), the player usually taps both the second and the fourth holes at the same time.^[64] In some schools, the player will tap only the second hole. But what Kitahara does is to tap the fourth, and then, very rapidly, the second alternatively, then playing the second tone of the phrase, giving three tones instead of two as a result.

[120] Though he starts with the same two first phrases, the third one is different. For example, he skips phrases 6 and 7 in Yokoyama’s version to play phrase 8 and 9 (60–79 s_a), though differently from the phrases on the transcribed score.

[121] He plays phrase 10 (88–97 s_a), while skipping phrases 11 to 13. However, he plays similarly, though not exactly as written, phrases 14 to 19 (98–159 s_a). He skips phrase 20 to go directly to the koro-koro on phrases 21 and 22 (160–170 s_a). From this place on, his performance is totally different from the other recordings. Although we find a few phrases that can occasionally be found in the original score of Yamamoto (in a different order), they are interspersed with a few phrases of his own that are not found there. However, the conclusive phrase is similar to the last phrase of all the other recordings.

Additional Remarks

[122] Kitahara begins (7.2 s_a) with a yoko-yuri (Fig. 47), which will remain nearly constant throughout the piece, interspersed with frequent tate-yuris. (See one example of three successive and distinct tate-yuris in Fig. 48, and of a tate-yuri transforming into a throat vibrato in Fig. 52.)

[123] While other musicians seem not to use yoko-yuri, Kitahara’s particularly clear technique allows for a distinct identification of this procedure. As a matter of fact, one particularity of this musician’s performance is his distinctive mastery of the different techniques he uses throughout the piece.

[124] He performs only three koro-koros–out of the four present in Yokoyama’s score–with generally an A⁺ (approximately one quarter tone higher than the equal-tempered A)–namely at 160–168.4 (Fig. 49), 196.5–203 and 334–342 s_a.

[125] Like Nishimura, Kitahara uses initial nayashis in portamento, for example at 189+ and 266.5+, 292+, 296–296.5, 324.5–326 and 346+ s_a (Fig. 50). He seems to avoid groups of nayashis, in contrast to the Yokoyama transcription/score and most of the other musicians’ performances, including Yokoyama’s.^[65]

Fig. 47. Long, initial yoko-yuri by Kitahara, which remains nearly constant throughout his performance.

Fig. 48. Three successive and distinct tate-yuris performed by Kitahara at 168.5–176 s_a.

[126] Moreover, Kitahara frequently uses portamentos, being shorter or longer, such as at 60–61, 124.5–125, 127.9–128.3, 154.8–155.2, 189.1–189.5, 191.6–191.9, 248.1–248.3, 292–292.4, 296–296.4, 297.5–298.5, 324.4–325.9, 328.8–329.7, 346–347, 348.8–349.3 (descending–Fig. 50), 351–351.8, 361.6–362.1 and 378.2–378.6 s_a.^[66] The e g e portamento of Yokoyama is, however, missing in his performance.

[127] More unique to this musician are the repetitions of alternations between two neighboring notes, such as at around 121.7, 194.2, 254.5–255 and 271.5 s_a (see one example in Fig. 51).

[128] Kitahara is also characterized by his throat vibrato, particularly discernible at 309–314 s_a, where a tate-yuri transforms into throat vibrato. (Fig. 52–See also Fig. 53 and Fig. 54).

Fig. 49. A koro-koro with heightened A (practically A⁺ = A + one quarter tone) performed by Kitahara at 160–168.4 s_a.

Fig. 50. Nayashi with portamento followed by a descending portamento at 348.8–349.3 s_a then by an equal-tempered F#.

Fig. 51. Two alternations between neighboring notes around 194 s_a.

Fig. 52. A tate-yuri transforming into a throat vibrato at 309 s_a.

Fig. 53. An accentuation perceived as a muraiki (see the scrambled spectrogram) at 46.5 s_a with a long tate-yuri at 46.6–52.7 s_a, interspersed with (rather mild) muraikis at 47.8–52.8 s_a, while ending with an unidentified vibrato (maybe a throat vibrato).

[129] His muraikis are generally mild and can be perceived as either accentuations (see one example at 46.5 s_a–Fig. 53) or effective, and sometimes successive muraikis, transforming at times into a throat vibrato. (Such as at 284–287.7 s_a–Fig. 54.) He uses also at least one tate-yuri interspersed with (rather mild) muraikis at 46.6–52.7 s_a, ending with an unidentified vibrato (maybe also with the throat) at 53–57 s_a (Fig. 53).

[130] While he accentuates the descending phrase f# e c b A (F) E only once at 223.3–224.5 s_a (the whole descending sequence except for e), he frequently uses repetitions, such as at 106-108 (three times only), 121.3–121.6, 183–184.2, 251.5–254 and, (very subtly) at 277–280.3 s_a.

Fig. 54. A series of muraikis at 284–287.7 s_a transforming into a throat vibrato beginning with 286.6 s_a. (Observe the differences in spectrum and intensities.)

[131] The semi-tones e-f and E-F are mostly performed as equal-tempered in ascent, generally in portamento (at 150–152.5–Fig. 55). (Exceptions with a higher e–such as at 112–116 s_a, Fig. 56–exist.) In descent, he may use a quarter-tonal f e⁺ (see at 20.7 and 25 s_a in Fig. 47, but equal-tempered, for example at 223 s_a in Fig. 58, and also at 296 s_a), and an equal-tempered c-b (Fig. 55, Fig. 57 and Fig. 58.)

Fig. 55. A semi-tone e-f in portamento at 150–152.5 s_a, followed by a high-pitched f# also in portamento, and by a (mostly) equal-tempered descending c-b semi-tone around 158.5 s_a.

Fig. 56. Small semi-tone in (portamento and) ascent at 112–114 s_a, and 114–116.6 s_a.

[132] He uses frequent, mostly accented–compared to Nishimura, for example–and very high-pitched f# , such as at 329.7–330.7 (nearly a g–See Fig. 57), but also equal-tempered ones such as at 222.8 s_a (Fig. 58).

[133] Other uses of f# can be found at 74 and 155, with one equal-tempered F# near the end of the piece (at approx. 352 s_a–Fig. 50).

[134] Finally, he does not use extra-modulations such as with Nishimura and, as we will see, Yokoyama.

Fig. 57. Micro-modulation with very high f# around 330 s_a with a rather equal-tempered descending c-b semitone around 323.5 s_a.

Fig. 58. Nearly equal-tempered f#, connected to the other notes of the descent f# f e c b A E ending with a beautiful, long, tate-yuri. Note also the equal-tempered semitone c-b around 223.2 s_a.

**Analysis of Yokoyama Katsuya’s performance of Kokū ^[67]**

[135] Preliminary remarks on Yokoyama’s performance: please note that the tate-yuri (box) around 219 s_a could begin earlier, at 217.8 s_a., probably performing a nayashi at 471.5 s_a, while the nayashi box around 248 s_a could be extended to 249 s_a.

[136] The last shakuhachi player whose recording we analyzed is Yokoyama Katsuya. We compared one of his recordings with his own score. As Sogawa Kinya mentions, he effectively does not follow it exactly.

[137] He adds nayashis and a tone on a few occasions. He merges groups of two phrases. Towards the end, he replaces a few phrases with something else. He follows his own score, we would suggest, roughly around 90–95 % of the time.

[138] Here are a few examples. He merges phrases 1 and 2 (0.5–12.5 s_a), thus becoming a single phrase and does the same with phrases 14 (131–136 s_a), 18 (172–178 s_a), 25 (225–232 s_a)–although this phrase is in the lower octave–and 36 (310–314 s_a). However, as the score shows, in the second half of that phrase, there should be numerous nayashis. In his case, compared to the other players, his nayashis sound more like a repetition of d-e that starts the phrase, ending with a nayashi.

[139] Phrase 3 (14-22 s_a) is also interesting. As mentioned about the other musicians, they do not perform the head movement that produces these two tones. He is the only one that plays them as written. As mentioned just above, he merges groups of two phrases, modifying them along the way; for example, phrases 11-12 (98.5–113 s_a), 29-30 (264–277 s_a), 39–40 (322–336.5 s_a), 43–44 (347.5–355.5 s_a).

[140] At phrase 33, after the third tone, we see in the Japanese score the line that asks the player to hold the tone for one or two seconds, followed by a head movement down for another second or two, and then a quick up-down-up movement. Yokoyama plays that motive in such a lively manner that it sounds more like a mawashi-yuri, which is what we indicated in the video (290–293.5 s_a).

[141] As for phrase 35 (302.5–308 s_a), he is the only musician who does it exactly as written in the score, while phrases 59–52 (452–491 s_a) are different from the score.

[142] As a preliminary remark for the video analysis, two notes had to be manually corrected around 87.7 and 289 s_a (Fig. 63), the audio having been downloaded from YouTube, in low resolution sound.^[68]

[143] A few remarks can be further made about the yuris. The nayashi at bar 45, 359 s_a, should have been a mawashi-yuri. The nayashi at 417.2 s_a does not figure in the score, while the mawashi-yuri at 415.4–416.9 s_a (Fig. 59) corresponds to the one in bar 54. The tate-yuri at 353.6–354.2 s_a should have been a mawashi-yuri as indicated in bar 44 of the score. Finally, the tate-yuri around 443 s_a does not figure in the score.

[144] The semi-tones of Yokoyama are rather small (see 56–52, 68–70, 116–122, 136–142, 148–150, 156–156.6, 164–170, 218–224, 249–253, 299–302, 442–448, 513–519, 552–558 s_a), while the other notes are regularly disposed on the graphic scale. (See one example of differences between semitones at 249–253 and 299–302 s_a, both beginning with a tate-yuri, and the remaining, regularly disposed notes in Fig. 63–which shows a wide view from 246 to 304 s_a.)

[145] Yokoyama’s koro-koros (four as in the score) are either very long and slowed down (199.5–211.5–Fig. 61–and 534–543 s_a) or very short (337–339.5–Fig. 62–and 395–398 s_a).

[146] The repetitions of notes are frequent (262.2–263.1, 295.7–296.3 s_a, bar 35 at 303.2–304.7 s_a, 417.9–418.4, 465.8–466.4, 476.5–478.7 and 500.1–500.7 s_a).

[147] As for the accentuations, they are rather soft and occur at 349.7–351 (mostly descending A and–less–E, with an accelerated tempo) and 409-410.1 s_a (same accentuations, also with an accelerated tempo–Fig. 59).

[148] Other accentuations can be found at 418–418.4, 435.4–435.9, 465.9–466.5 and (mainly for the repetitions) at 475.7–478.5 s_a.

[149] The E G E portamento in bar 46 in the score can be found, transposed to the upper octave and limited to (a rather equal-tempered) f# in ascent, at 271–273 s_a (Fig. 63).

Fig. 59. Accented descending phrase f# e c b A E with the last two notes (and especially the A) being more accented than the other notes, preceded by 3 nayashis, and followed by two mawashi-yuris and two nayashis.

[150] Other occurrences of f# take place at 349.5–350.1 s_a (rather high-pitched), in portamento in the lower octave at 366–367.8 s_a, and a long, slightly low-pitched F# at 382.2–387.7 s_a, repeated at 389.8–393 s_a. (Fig. 60)

Fig. 60. A long, slightly low-pitched F# at 382.2–387.7 s_a, repeated at 389.8–393 s_a.

[151] The last two occurrences of f# can be found at 408.7–409.1 s_a (rather equal-tempered–Fig. 59) and at 519.1–519.7 s_a (very faint, rather equal-tempered).

[152] Last but not least, Yokoyama’s performance, when compared for example to Nishimura’s, is rather lively and even incisive, with a tendency–except for the semitones–to maintain the Western equal temperament’s semitonal notes’ pitches.

[153] Other noticeable portamentos occur for example at 68–70 (ascending small semitone–rather frequent), 147–149, 153.5–155 and 164–171 (descending) s_a, etc.

[154] With regard to the modulations to f#, they are (very) frequent and occur at 73–73.9 s_a (accented and high-pitched), 184.3–185 s_a (very soft and nearly equal-tempered), with an additional micro-modulation to the upper (high) B^b around 268 s_a (bar 29) preparing the aforementioned 271–273 s_a transposed portamento, with a remake of f# around 274.1 s_a (Fig. 63).

Fig. 61. Long and slowed down koro-koro by Yokoyama at 199.5–211.5 s_a.

Fig. 62. Short koro-koro at 337–339.5 followed by a mawashi-yuri.

Fig. 63. General view from 246 to 304.5 s_a in the video analysis of Yokoyama Katsuya’s performance of Kokū, showing small b c semitones beginning with tate-yuris, two linked micro-modulations to B^b and f#, as well as the manual correction around 289 s_a followed by a mawashi-yuri, a nayashi and another tate-yuri.

Comparing the Six Analyses

[155] As explained in the “Preliminary remarks” at the beginning of the analyses, in this section, we will compare the six performances’ correspondences with (1) Yokoyama’s score, (2) the execution of the series of nayashis at bars 2, 7, 14, 18, 25, 36, 60 and 69, (3) the koro-koros in number and particularities, then compare (4) the accentuations–be they in tempo or intensity or as quarter tone accentuations –, (5) the E G E portamento, (6) the use of the half-tones and (7) of repetitions, (8) the micro-modulations to f# and F# and (9) additional modulations and, finally, (10) additional particularities in each musician’s performance.

[156] These 10 comparisons are detailed below and in Tables 4 to 13.

[157] Additionally, in the course of our analyses and explorations of our six musicians’ different ways of performing Kokū, we noticed one important factor that we may have underestimated at the beginning of our research. This is the, sometimes, considerable differences in the six performances’ durations. We consequently added an eleventh comparison for these durations (Tables 14 to 16), and correlated them with the lengths of the instruments used by the performers: the results are interesting and open further perspectives for the analysis of the shakuhachi repertoire.

1. Correspondences of the performances with the score of Yokoyama Katsuya

[158] Surprisingly enough, Ramos’ performance is the most faithful to Yokoyama’s score. Yokoyama himself takes some–small–liberties with his own score, very much like Sogawa and MacGregor (Table 5).

Table 5. Correspondences with Yokoyama Katsuya’s score.

[159] While MacGregor’s differences may come from Tajima’s teaching, Nishimura’s and Kitahara’s versions are very far apart from one another’s performances, and they also consistently differ from Yokoyama’s written version.

[160] In the case of the last two performers, these consistent differences seem to be normal, as both had a different teaching, and both recorded the analyzed version before Yokoyama’s score was published.

2. Execution of the groups of nayashis

[161] The groups of nayashis seem to be in disfavour among the six performers (Table 7).

Table 6. Correspondences with Yokoyama Katsuya’s series of nayashis in his score.

[162] Some of them prefer replacing them, sometimes, with muraikis (Ramos and Sogawa), while even Yokoyama does not perform the whole series in his score. Kitahara and MacGregor definitely do not use them, while Nishimura frequently uses pairs of nayashis before and after a note.

3. Performing the koro-koros

[163] As a first observation on the koro-koros, Kitahara and Nishimura perform them differently from the other musicians. While Kitahara–who has the shortest version–inserts only 3 koro-koros in his performance of Kokū and uses an A⁺ instead of a tempered A, all five other musicians insert 4 koro-koros, generally two short and two long ones, mainly based on the equal temperament scale (Table 7). Nishimura is the only musician among the six who performs his koro-koros in a very different, distinctive manner, particularly slow and soft.

Table 7. Correspondences with Yokoyama Katsuya’s koro-koros in his score.

4. The Accentuations

Intensity (and Tempo)

[164] Accentuations in intensity and tempo seem to be the favourite technique allowing the performers to distinguish themselves from other musicians. (Table 8, 2^nd column.)

[165] Some of them accentuate all the corresponding notes in the score, and others only a few of these notes. MacGregor, for example, accentuates all the accented notes in the score, while using frequently a decreasing tempo in his intensity accentuations, and also frequent tempo accentuations. At the other end of the beam, Kitahara accentuates the f# c b A descent only once, excluding from it the e.

Table 8. Comparison of the use of accentuations for the six versions.

[166] Yokoyama, surprisingly, rarely uses–rather soft–intensity accentuations, mostly limited to the descending A. Sogawa uses preferably muraikis in place of accentuations, but he does accent at least once the descending f# e c b A phrase.

[167] As for Nishimura, he uses frequent accentuations, but these are generally limited to the two descending b A notes.

Quarter-tone accentuations

[168] Melodic accentuations using approximate quarter tones or quarter-tone transpositions of phrases or of parts of them are, in fact, very frequent in the six versions, especially for the semitones, which become thus quarter tones (Table 8, 3^rd column). The quarter-tonal accentuations within the semitones are frequently performed in portamento, while Kitahara distinguishes himself among the six by using an approximate A⁺ in his koro-koros.

[169] Ramos and Nishimura perform sometimes full quarter-tone transpositions of parts of phrases, if not of complete phrases (mostly with Nishimura). MacGregor uses semitone/whole-tone accentuations once around the f#, while Nishimura, who holds the prize in the gallery, uses very subtle quarter-tone variations in portamentos around one note.

5. The portamento from E to G and back to E

[170] The performance of the E G E ascending then descending portamento, found in Yokoyama’s score at bar 46, with successive passages through E F F# G and back to E through F# and F, is very differently performed, or the contrary, by the six musicians (Table 9).

[171] While Kitahara just does not perform it–and let us remember that he is evidently not from Yokoyama’s school,–we could not identify a clear E G E portamento with Nishimura. MacGregor–although from Yokoyama’s school through his sensei Tajima–simply does not perform it, while Ramos omits the F in both ascent and descent. Yokoyama himself performs it differently, as it is transposed to the upper octave and limited in ascent to the f#. As for Sogawa, he performs it effectively in the lower octave, but also limited to F# in ascent (thus a portamento E F# E).

Table 9. Comparison of the use of the portamento from E to G and back to E.

6. Use of the half tones

[172] The half tones e-f E-F and b-c are also a privileged ground for differentiation between the six performers. Yokoyama uses, rather unsurprisingly,^[69] mostly equal-tempered or slightly smaller semitones all along his performance (Table 10).

[173] Nishimura’s semitones e-f are mostly equal-tempered except during the frequent quarter-tonal accentuations (with portamento) to the f. His semi-tones E-F are frequently in portamento or yuri.

[174] Kitahara’s e-f and E-F semi-tones are mostly equal-tempered in ascent, generally in portamento or yuri–while exceptions such as the use of a higher e, such as at 112-116 s_a, exist. In descent, he frequently performs a quarter-tonal f-e⁺, but an equal-tempered c–b.

[175] Sogawa’s semitones are near systematically small with non-systematic portamentos, and frequently in mawashi-yuri for the b-c semi-tones, while with MacGregor the e-f semitones are mostly equal-tempered, except during the quarter-tonal accentuations, and with possible portamentos or yuris.

[176] Finally, Ramos’ semitones are generally smaller than the equal-tempered half tone between f and e, with frequent quarter-tonal “play” around the f. His E-F and b-c semi-tones are frequently used in yuri or portamento.

Table 10. Comparison for the use of the semitones e f, E F and b c.

7. Use of repetitions

[177] Repetitions are frequent with all the performers, except for Kitahara, who performs none (Table 11). While MacGregor uses frequent accentuations in his repetitions, Sogawa distinguishes himself by the time acceleration of his repetitions, with an end in portamento.

Table 11. Comparison between the six versions for the use of repetitions.

8. Micro-Modulations to f# and F#

[178] The modulations to f# and F# seem to be a must in Kokū, a clear characteristic of the piece (Table 12).

Table 12. Different uses of micro-modulations to f# and F#.

[179] The pitch of the two notes, when compared to their equal-temperament counterparts, can be high with some performers (mostly with Nishimura and Ramos, slightly less with Kitahara), low with others (mostly MacGregor) or variable, as with Yokoyama and Sogawa.

[180] Nishimura’s modulations are distinctively long-lasting (as–albeit less–with MacGregor for F#), and sometimes repeated and accented with initial muraikis, but also sometimes faint and very short. Kitahara’s modulations are mostly accented, while Ramos uses variable intensities for each modulation and Sogawa distinguishes himself through the use of portamentos while performing them.

9. Additional modulations

[181] Only two performers (Yokoyama and Nishimura) (dared to) use additional modulations (Table 13). Yokoyama modulates to B^b and links this modulation with the f# modulation, a fourth apart.

[182] Nishimura’s additional modulation to c#, central between two neighboring modulations to a preceding f# and to a following F#, is more elaborated and plays on both relations with the fourth and the fifth.

[183] Whenever Yokoyama’s additional modulations departs from his own score, we cannot, at this stage, determine if Nishimura invented his additional modulation or if he learned it from another musician.

Table 13. Use of additional modulations.

10. Additional particularities with each performer

[184] There is much to say about each version and each performer, and, clearly, all we can state about each musician’s abilities and particularities will always fall short at describing them completely.

[185] However, through our repeated listening and comparisons of the six versions researched in this article, we can try and establish some of the particularities for each of their styles, at least for the particular version by the particular performer of this particular piece, Kokū (Table 14).

Table 14. Additional particularities of each version and performer.

[186] We have noticed with Yokoyama a clear tendency to stick to the normative equal-tempered pitches of the scale with, however, very small semitones: e-f. We have also appreciated his incisive, lively style, and his mild use of muraikis.

[187] While the followers of his school–whether directly or indirectly–are mainly faithful to his score, the two versions by Kitahara and Nishimura are refreshing in their profound differences in interpretation of what is supposed to be the same piece.

[188] This does not mean that the versions of Sogawa, Ramos and MacGregor are devoid of interest–quite the contrary. Each of them, while showing maybe an overemphasized respect for Yokoyama’s written score, has his own style and way of performing Kokū, which distinguishes him clearly from the other performers of this piece, and from the other versions examined in this article.

[189] Sogawa uses, for example, frequent portamentos, notably for the F# micro-modulation and, for the end of phrases, a technique which is much less in use–if not absent–in the five other musicians’ performances. While his pitches are–as with Yokoyama–mostly adjusted to the equal-temperament grid, his semitones–also as with Yokoyama–are small, but he embellishes them further with frequent mawashi- or tate-yuris. He also uses frequent muraikis in place of the nayashis in Yokoyama’s score, a clear preference–or a stylistic particularity–of his.

[190] MacGregor has also his preferences and is the only performer who uses a stretched nayashi in the upper octave. While he seems to avoid mawashi-yuris–at least in this version of Kokū–he is one of two performers (with Kitahara) who clearly use(s) the yoko-yuri vibrato, and the only one to perform solely single nayashis instead of pairs or series of them.

[191] Ramos has a clearly different style, with long and numerous, mostly very powerful, muraikis. He is the only one among the six who uses a particular technique–of the lips?–for the end on one pitch of his phrases, resulting in a characteristic sound and timbre (with an empty central part of the spectrogram).

[192] As for Kitahara, his very clear technique, which can be traced quite effortlessly on the graphic analysis, is unique among the six performers, as well as his near-constant yoko-yuri all along the piece. He completely avoids the use of mawashi-yuris, but he does use frequent portamentos, including initial nayashis with portamentos, a particularity that he shares with Nishimura. While he seems to be the only musician among the six to favor rapid and double alternations of neighboring notes he, as a closing remark, clearly distinguishes himself with his throat vibrato–with a very particular spectrum and varying intensity–and with his three notes debuts of the opening phrases (instead of the two that the five other musicians stick to).

[193] Last but not least, Nishimura is clearly the most inventive and subtle musician among the six. His slow and meditative style clearly deserves the denomination “Zen shakuhachi”, with a mastery of different techniques and nuances of accentuation far from being equaled by the five other performers. His performance is inventive and imaginative–although, as aforementioned, we cannot determine if this inventiveness is purely his, or learned from another musician.

11. Duration of each version

[194] As mentioned above, shakuhachi performers may use instruments of different lengths, varying from 1 shaku up to 4 shaku. They choose the instrument which, according to personal musical taste or to the mood they wish to infuse, best expresses the spirit of a particular piece, but also in order to vary the timbre between the pieces, instead of using a single flute for the entire concert or CD.

[195] We noticed, however, that another factor could influence, or could be influenced, by the length of the shakuhachi, which is the duration of the same piece played on instruments of different lengths. In fact, the difference between the shortest version analyzed in this article, which is Kitahara’s 392 seconds^[70] version, and the longest one, which is Nishimura’s 1021 seconds^[71] performance, is considerable, with a ratio (Table 15) of 2.6 between Nishimura and Kitahara.

[196] Intermediate durations for the other versions are, between Kitahara and Nishimura in the ascending direction, Yokoyama’s version with 586 seconds, Ramos’ with 630 seconds, Sogawa’s with 642 seconds, and lastly MacGregor’s with 752 seconds.

[197] If we go back to the determination of the tonics and the effective lengths of the corresponding shakuhachi for each version and performer–reminded to the reader here in Table 17–it seems that there is only a loose correlation between tonics and the shakuhachi lengths, which we explained by the consistent differences in the bamboos used, and in the manufacturing of each instrument. For example, MacGregor’s instrument is a 2.55, but the frequency of his tonic is slightly higher than Sogawa’s with a 2.5 shakuhachi, whereas it should be slightly lower.

Table 15. Comparing the durations of the six versions.

[198] The tendency to follow an increasing tonic pitch with the length of the instrument remains, however, undeniable, and justified by acoustic considerations.

Table 16. Reminder about the tonics and effective lengths of the shakuhachi for each performer. (Nishimura’s and Yokoyama’s instrument lengths estimated by Bruno Deschênes.)

[199] Similarly, when comparing the ratios of durations over the shortest one (Kitahara’s as shown in Table 17, first column) with the ratios of the tonic frequencies over the lowest one (still Kitahara’s, Table 17, second column), we still have the same anomaly in the progression for MacGregor who, if the correlation between the two series were complete, should have an 𝑅₁⁄𝑅₂ ratio in between Nishimura’s and Sogawa’s.

Table 17. Comparing the ratios of duration with the ratios of tonic frequency.

[200] While this anomaly is corrected when comparing the ratios of durations with the ratios of the shakuhachi lengths (Table 18), with MacGregor’s ratio 𝑅₁⁄𝑅₃ in between Nishamura’s and Sogawa’s, the other three intermediate instruments have close values of the ratio 𝑅₁⁄𝑅₃, which may be explained by their close lengths (and perhaps by an erroneous evaluation of Yokoyama’s instrument’s length).

Table 18. Comparing the ratios of durations with the ratios of instrument lengths (with Nishimura’s shakuhachi length estimated to 2.8, and Yokoyama’s to 2.2).

[201] Nevertheless, there exists a clear correspondence, although not a straightforward one, between the lengths of the instruments used by the performers, and the duration of the version of Kokū that they perform in these recordings.

[202] While this is too small a group of shakuhachi player (statistically) to be able to draw firm conclusions, it is clear that further research on the whole repertoire of the solo shakuhachi could shed more light on this phenomenon, which raises ethnological, aesthetic and organological questions of the utmost interest in ethnomusicology.

Conclusions

Musician or musicologist?

[203] This series of analyses of Kokū was a real risk-taking enterprise for both authors of this article, the result of the collaboration between two musicians and musicologists, the first of whom (Bruno Deschênes) is a specialist of the shakuhachi and a self-taught ethnomusicologist, and the second one (Amine Beyhom) a musician and musicologist specializing in the analysis of traditional music, but confronting his skills for the first time with Japanese music.

[204] What would be the feeling of a musician when confronted to these analyses of Kokū, Beyhom asked Deschênes?

[205] As a shakuhachi player, Bruno Deschênes mentions that he usually listens to the character and the playing style of a recording. It is probably what most, if not all shakuhachi players do. This is confirmed by his discussions with numerous other players. Musicians do not usually pay direct attention to the technical aspects of a piece. As musicians, we might notice that Nishimura Kokū does a slow koro-koro compared to the other musicians, or that Alcvin Ryūzen Ramos has a more outgoing style of playing, but rarely will we pay attention to see if he follows a score or not, for example. The quality, the character of the playing or, should we say, the musicality of the musician, is what attracts our attention.

[206] However, in working on these thorough analyses–something quite unusual for him–Bruno Deschênes particularly grasped two aspects of the playing of these musicians that did not attract his attention before. Firstly, the subtleness in using the different techniques, being nayashi, muraiki, koro-koro and others. These techniques are probably what allows a musician to create her or his own style, even though most music lovers who are no musicians might not even notice them, just perceive them. Secondly, as mentioned earlier, many Japanese and non-Japanese shakuhachi players believe that these melodies are not melodies of notes or tones, but melodies of tone-colours. This can sometimes be inferred from the spectral analysis. On some occasions, for the same tone, it can be noticed that the spectrogram of a particular tone differs from one musician to another and even within the same piece, which is due in large part, Bruno Deschênes believes, to the way each one of them trained his lips and orofacial muscles.

[207] As mentioned earlier, no two shakuhachis of the same length have the same tone-colour, and the same shakuhachi played by different musicians will also not have the exact tone-colour. The tone-colour of the shakuhachi appears to be very dependent from the movements of the lips and these muscles, including the movements of the head and the position of the flute on the chin.

[208] On the other hand, and as a musician and a musicologist, Amine Beyhom was bewildered by the multiplicity of techniques and the subtleties in the performances on this very simple bamboo flute, and by the sounds it can produce, notwithstanding the infinite variety it demonstrates for pitch progression and handling, together with a very wide range of expressivity.

APPENDIX: NOTATION, TRANSCRIPTION, SCORES AND TRANSNOTATIONS OF KOKŪ

The Shakuhachi Notation

[1] The current notation for the shakuhachi was created at the end of the nineteenth century. It uses characters from the Japanese syllabary katakana (see THT 1). Each character or set of characters designates not a tone, but a fingering from which the tone is produced. The main characters correspond to the fingerings of the base tones produced by fully covered holes. The intermediate tones, that is, the tones produced by partial hole openings and head movements, are indicated by adding a smaller character representing that head movement.

[2] The characters meri (メ) indicate tones requiring a lowering of the chin and kari (カ) indicates an upward movement of the head. One particularity of this notation is that pieces can be played with any flute regardless of its length. The tones to be produced in the first octave are indicated by the character otsu (乙 or 呂), in the second by kan (甲) and in the third by daikan (大甲).

[3] Although this notation is standard, some schools use different characters for some tones. In some schools, the notation can be detailed, while in others it is minimal. Typically, shakuhachi scores are guides and not definitive representations of a piece since some aspects of the style of a school cannot be directly indicated on a score.

[4] THT 1 below shows the basic standard characters for the 1 shaku 8 shakuhachi. These characters refer to fingering, not tones or notes. The staff below gives the tones for the 1.8. They vary depending on different lengths of shakuhachi. The readers will notice in the scores of Kokū proposed below that some characters differ from the ones proposed in THT 1. The first score for that piece is from the Yokoyama Katsuya school (FHT 1, calligraphed by Furuya Teruo), and some of the characters are unique to that school. The score is read from top to bottom and right to left.

THT 1. The notation for the shakuhachi refers to the fingerings, not to the tones or their pitch. (© Bruno Deschênes.)

Transcription of Kokū

[5] As is usual with the transcription of any non-Western music, it is challenging (if not impossible) to transnotate entirely and perfectly the Japanese score (FHT 1) into the European notation (FHT 2, FHT 3). Furthermore, not all shakuhachi players will play a piece exactly as written. The notation is basically a guide for performance. There is some leeway that allows shakuhachi players to alter a phrase (or a “bar” in the score)^[72], add or remove a note or a phrase, add a nayashi or a yuri, or not playing them when indicated, play a short tone long, or the other way around. It is not expected that a player would play a piece in the same way each time; he may alter it with every performance.

[6] Because of the musical and timbre complexity, and the subtlety of playing this particular bamboo flute and its music, it is not possible to transcribe in details all that the musicians are doing. When a yuri is indicated, at times they can be so soft or narrow that it is not really possible to clearly distinguish which of the three yuris it is.^[73] Furthermore, a nayashi–approximately “” in the Japanese score and reproduced smaller (“”) in the European score–can sometimes sound like two tones, when normally it is a slight portamento with a down-up movement of the head. Usually, except for a nayashi, when there are head movements, the portamento must be avoided by making them quick, though in the end this is up to the player. Moreover, since the tones are never perfectly stable, when we think we might be hearing a yuri, it might end up being a fluctuation of the tone, among others.

[7] One particular aspect of shakuhachi solo pieces is that phrases are separated by obligatory pauses, which allows the player to take a breath before playing the next phrase. These create silences between phrases, which have become inherent to the aesthetics of these pieces. Their durations are not fixed. They are left to the players. Some might make them short, others longer. Some might “group” two or more phrases with short silences in-between and a longer one before playing the following group of phrases.

[8] The flow of the music is left to the player. The breathing marks are shown by short horizontal lines (– vertical semi-bars in FHT 1). The first lines on the two pages of the original score are the Japanese numbers (一, 二, 三, 四…) that number each line of the piece. These numbers have no particular purpose in the performance of the piece.

[9] In Kokū, single nayashis are after a tone. In some pieces they can appear before a tone. Nayashis can also be found in groups, the first example being the second phrase of the first column (bar 2), in which we see four nayashis, followed by a tone, then followed by a single nayashi. This group is also found in bars 7, 14, 18, 25, 36, 60 and 69. For reasons of space, we found it appropriate to present the nayashis in the transcription stepwise, instead of one after the other as in the original score. In some traditional repertoires, nayashis are not indicated on a score, but the player can add some here and there while playing.

[10] In the first phrase of the second column (bar 3), following the first tone, the broken line indicates a down movement of the head. This means that the tone starts as indicated by the character, then the player must move her or his head down to bring the tone about one half-tone down, while keeping the current fingering. This line appears several times in the score, it can be short or long.

[11] Another broken line is found in column 12 (bar 28), second phrase of the line. That phrase is produced by a movement of the head down followed by a rapid movement up-down-up.

[12] One important point about these broken lines is that the head movement must be fast, so as not to produce a portamento between the tones. When a portamento^[74] is requested, it is indicated as in the last two tones of the first phrase of column 12 (bar 27). In this case though, it cannot be avoided here since the first of these two tones is produced by positioning the head down to bring it up to the second tone. But the portamento has to be avoided as much as possible by a quick movement of the head.

[13] In column 4 of the transcription, bar 12 (column 6 in the original score), there is a wavy line () that refers to the third type of vibrato, yuri (here a mawashi-yuri), mentioned above. It is done with a circular movement of the head.

[14] These vibratos are found in bars 12, 24, 37, 41, 44, 45, 46, 48, 50, 53, 54, and 68. Similarly to the nayashis, we inserted it in the transcription since this particular type of vibrato does not exist in Western notation,

[15] On a few occasions, we added an “under (or ‘over’) score” (“_”) above or below some tones (, , for example, on bars 35, 43 or 52). This indicates an accentuation of the tone. However, in phrase 54, a few (“—…”) were added under the mawashi-yuri sign. This means that the lower tones of that mawashi-yuri have to be accentuated by tapping the fifth hole in the back of the shakuhachi while doing this yuri. The number of times that the hole is tapped is left for the musician to choose, thus the use of “…”. Usually, the musician taps it a few times at the beginning of the mawashi-yuri.

[16] Column 10 of the first page of the original score by Furuya, (bars 21 and 22 in the transcription) shows the koro-koro (–here rotated 90° clockwise, to be read from right to left–or in the transcription) presented below.

[17] In column 26 of FHT 1, second phrase (bar 59 found in line 20 of FHT 3), a group of repeated marks indicates a repetition of the tone that precedes them. There are a few others marks here and there, but the ones just described are the most important.

Scores and Transnotations

[18] Yokoyama Katsuya learned around thirty pieces from Watazumi, which included Kokū. This prominent shakuhachi player used scores, but his notation was, as Riley Lee indicates, extremely skeletal, differing from the commonly used notation by shakuhachi players since the end of the nineteenth century. While Watazumi was making his own shakuhachi that he called hocchiku (it was based on the one used during the Edo period by the komusō monks of the Fuke sect), Yokoyama, as a concert musician, preferred the jiari, i.e., the modern shakuhachi, made by professional makers.^[75] Yokoyama adapted Watazumi’s repertoire to be performed on the jiari, developed his own notation, alongside his own concertante style for modern audiences. Lee indicates that Yokoyama Katsuya did not publish scores at the beginning of his career because of the impossibility of finishing a score to his satisfaction. Scores are simple “memos”. Up until 1990, when his scores were published, Lee indicates that Yokoyama revised the scores of his honkyoku repertoire, never being fully satisfied.

[19] It is usually his students who try their hand at notating pieces.^[76] This is because Yokoyama believes that by placing too much emphasis on notation will inevitably limit honkyoku.^[77] The version of Kokū on which we base our analysis is Yokoyama’s.^[78]

FHT 1. The original score of Kokū for the Yokoyoma Katsuya school. Calligraphy: Furuya Teruo. (Used by permission, Furuya Teruo, first published 1990 by Kokusai shakuhachi kensyu kan, Tokyo – Japan.)

REFERENCES

Beyhom, Amine. 2007. “Des Critères d’authenticité dans les musiques métissées et de leur validation : exemple de la musique arabe,” Filigrane. Musique, esthétique, sciences, société 5, 63–91. https://revues.mshparisnord.fr/filigrane/index.php?id=168.

_________. 2014. “Dossier : Influence des théories européanisées du XIX^e siècle sur la notation et la pratique des modes de la musique arabe et d’autres musiques, à travers la mise en exergue du mythe du genre ḥijāz semi-tonal.” Near Eastern Musicology Online 2 3, 87–177. https://hal.science/hal-01447295.

_________. 2018. “MAT for the VIAMAP. Maqām Analysis Tools for the Video-Animated Music Analysis Project,” Near Eastern Musicology Online 4, 7, 145–258, https://www.researchgate.net/publication/340504126_Dossier_MAT_for_the_VIAMAP_Maqam_Analysis_Tools_for_the_Video-Animated_Music_Analysis_Project.

_________. 2019. “The Lost Art of Maqām–With four video analyses of performances by Evelyne Daoud, Neyzen Tewfik, Hamdi Makhlouf, and by ʿAlī Maḥmūd and Sāmī a-sh-Shawwā,” Near Eastern Musicology On-line 5 8, 5–64, https://www.researchgate.net/publication/349206036_THE_LOST_ART_OF_MAQAM.

_________. 2021. “Further Analyses from the VIAMAP,” Near Eastern Musicology Online 6/10, 5–36, https://cnrs.hal.science/hal-03680899v1.

Beyhom, Amine and Hamdi MAKHLOUF. 2021. “A VIAMAP exploration of the Tunisian ṭubūʿ,” CTUPM.http://ctupm.com/en/a-viamap-exploration-of-the-tunisian-tubu/.

Blasdel, Christopher Yohmei and Yūkō KAMISANGŌ. 1988. The Shakuhachi: A Manual for Learning, Ongaku no Tomo Sha.

Deeg, Max. 2007. “Komuso and ‘Shakuhachi-Zen’: From Historical Legitimation to the Spiritualisation of a Buddhist Denomination in the Edo Period,” Japanese religions 32, NCC Center for the Study of Japanese Religions, 7–38.

Deschênes, Bruno. 2017. Le shakuhachi japonais : Une tradition réinventée, Éditions L’Harmattan, Paris.

_________. 2018. Une philosophie de l’écoute musicale, L’Harmattan, Paris.

_________. 2021. L’écoute de la musique à l’esprit. Conformité, socialité et historicité, L’Harmattan, Paris.

Ensemble Yonin No Kaï. 1995. Sankyoku, CDC 560070, Ocora France, Paris. https://www.discogs.com/Ensemble-Yonin-No-Kaï-Sankyoku/release/13584573.

Gillan, Matt. 2021. “Sankyoku Magazine and the Invention of the Shakuhachi as Religious Instrument in Early 20th-Century Japan,” Yale Journal of Music & Religion 7/1. doi: 10.17132/2377-231X.1203. https://elischolar.library.yale.edu/yjmr/vol7/iss1/2.

Gützwiller, Andreas B. 1974. Shakuhachi: Aspects of History, Practice and Teaching, Wesleyan University Ph.D., University Microfilm, Ann Arbour, Michigan.

Johnson, Henry Mabley. 2014. The Shakuhachi: Roots and Routes, Brill, Leiden and Boston.

Kamisangō, Yūkō. 1988. “The Shakuhachi, history and Development.” In The Shakuhachi: AManualfor Learning, edited by Christopher Yohmei Blasdel, Ongaku no Tomo Sha, 69–132.

Keister, Jay. 2004. “The Shakuhachi as Spiritual Tool: A Japanese Buddhist Instrument in the West”, Asian Music 35/2, 99–131. http://www.jstor.org/stable/4098447.

Lee, Riley. 1992. Yearning for the Bell: A Study of Transmission in the Shakuhachi Honkyoku Tradition, Ph.D., University of Sydney.

Linder, Gunnar Jinmei. 2012. Deconstructing Tradition in Japanese Music: A Study of Shakuhachi, Historical Authenticity and Transmission of Tradition, Ph.D., Stockholm University.

Macgregor, Andrew. 2021. “Zen Shakuhachi – Koku 虚空 ‘Empty Sky’ by Andrew MacGregor,” YouTube. https://www.youtube.com/watch?v=tKH41oaVobA.

Nishimura, Kokū. 1965. Nishimura Kokū 西村虚空–虚鐸普化宗本曲 (1965, Red, Vinyl), LP TH-7010, Toshiba Records, Japan. https://www.discogs.com/release/6488923-西村虚空-虚鐸普化宗本曲.

Ramos, Alcvin Ryūzen. 2008. Japanese Traditional Flute for Meditation; Zen Shakuhachi Vol. 2 [K10-25CD], K10-25CD, Oliver Sudden Productions. http://www.oliversudden.com/catalog/index.php/main_page/product_music_info/cPath/4/products_id/4.

Sanford, James H. 1977. “The Fukeshu and Komuso,” Monumenta Nipponica 32/4, 411–440.

Sogawa, Kinya. 2014. “Kinya Sogawa plays Koku,” YouTube. https://www.youtube.com/watch?v=rT1aOR1SK44.

Soothill, William Edward and Lewis Hodous. 1934. A dictionary of Chinese Buddhistterms: with Sanskrit and English equivalents, a Chinese index and a Sanskrit-Pali index, Taipei, Taiwan: Buddhist Culture Service.

Thompson, Christopher S. and John W. Traphagan. 2006. “The Practice of Tradition and Modernityin Contemporary Japan,” Wearing Cultural Styles in Japan: Concepts of Tradition and Modernity in Practice, 1st ed., Albany, NY: SUNY Press.

Vlastos, Stephen. 1998. Mirror of Modernity: Invented Traditions of Modern Japan, Berkeley: University of California Press.

Watazumido. 1968. 海童道宗祖法竹(Hotchiku ), LP Polydor MN-4012 (LPJM-526?). https://www.discogs.com/release/14349964-海童道宗祖-法竹.

_________. 2000. 海童道宗祖海童道(法竹), CD Polydor–UDC-499. https://www.discogs.com/release/1098961–海童道宗祖–海童道–法竹.

_________. 2020. “海童道宗祖 – 虚空篪,” YouTube. https://www.youtube.com/watch?v=eTf6S99SHrY.

Yokoyama, Katsuya. 1997. Japon. L’Art Du Shakuhachi, CD C 560114, OCORA / Harmonia Mundi. https://www.discogs.com/release/3904177-Katsuya-Yokoyama-Japon-LArt-Du-Shakuhachi.

_________. 2011. Kokusai Shakuhachi Kenshukan, 2 DVDs BDVR-001 & 002, Tokyo: Bamboo.

_________. 2013. “Katsuya Yokoyama ~ Kokuu,” YouTube. https://www.youtube.com/watch?v=83of323uRo4.

[1]. The Introduction, explanations about the VIAMAP–or the “Video Animated Music Analysis Project”–and the video analyses are by Amine Beyhom. The first section and the Western notations of Kokū are by Bruno Deschênes. All analysis has been co-written by the authors. We wish to thank Japanese shakuhachi master Kurahashi Yōdō II for his help about the history of this bamboo flute and the life of Kitahara Kōzan. We also sincerely thank Sogawa Kinya for his kind answers to our questions and Laurie Sogawa for translating his answers and providing her insight into Japanese shakuhachi music, as well as Andrew McGregor and Alcvin Ryūzen Ramos for allowing the use of their recordings of Kokū. This article is accompanied by 6 video analyses.

[2]. Beyhom 2018, 2019, 2021; Beyhom and Makhlouf 2021.

[3]. We use a regular “s” with Japanese words in transcription for the plural.

[4]. From the CD Ensemble Yonin No Kaï, 1995. Note that we are mentioning Japanese names as with the Japanese usage, meaning the family name first.

[5]. For a history of the shakuhachi and its making see Lee (1992) and, in particular, Johnson (2014).

[6]. The name shakuhachi means 1 shaku 8 sun, shaku (30.3 cm/11.93 in) being an old Chinese unit of measure, while sun refers to dividing the shaku by ten. The length of the regular shakuhachi is 1.8 (See Johnson 2014 for a more detailed presentation of its making).

[7]. On a few occasions, Bruno Deschênes heard players in concert having trouble playing with a particular shakuhachi from the start of a piece. The reason is that these players did not warm up beforehand with that particular shakuhachi.

[8]. In particular, Nick Bellando from Hirosaki, Japan, and Alcvin Ryūzen Ramos, from Vancouver. Though most beginner players start with the standard 1 shaku 8 shakuhachi, Bruno Deschênes met a few who started with a long shakuhachi, for example Danish shakuhachi player Kiku Day. Moreover, Bellando explains (ibid.) that “there is a preference for longer flutes particularly among followers of Watazumi Dōso (1910-1992) – also known as Watazumi, Watazumi-do, or Watazumido, Watazumi being the name most commonly used by shakuhachi aficionados (including those who learned from Yokoyama) – as it was he who made it popular. (See also, for Watazumi, footnote no. 121.) It just happens that there is also a greater concentration of non-Japanese among these, as the Kinko and Tozan schools have near-total dominance in Japan. Many non-Japanese also tend to be physically bigger, and are thus able to play longer flutes compared to Japanese people (who tend to be smaller on average, and have shorter arms as well).” Note, however, that, in the particular case of Kokū and for the six musicians whose performances are analyzed further, two non-Japanese performers (McGregor and Ramos), as well as one Japanese performer (Sogawa), use a middle-length shakuhachi, while Japanese Nishimura uses quite a larger instrument, with only Kitahara using a standard 1.8 shakuhachi and with Yokoyama using an intermediate, 2.2 one. (The dimensions of the six instruments are detailed in the sub-section of Part II “Determining the Tonic and the Reference Pitch”.)

[9]. This anecdote was related to Bruno Deschênes by his Japanese master Kurahashi Yōdō II during a class in the Fall of 2021.

[10]. This is the usual notation symbol used in the shakuhachi notation for nayashi.

[11]. The way this technique and the ones that follow are presented differs from one school to another, the reason being that these is no full standard notation for all shakuhachi schools.

[12]. This technique is sometimes difficult to distinguish from tate-yuri when hearing it. There is also a fourth type of vibrato which consist of moving back and forth the shakuhachi itself. It is rarely used.

[13]. In Buddhist terminology, Kokū means “space” or “sky”: “śūnya; empty, void, space; ākāśa, in the sense of space, or the ether; gagana, the sky, atmosphere, heaven; kha, space, sky, ether, 虚 is defined as that which is without shape or substantiality, 空 as that which has no resistance. The immaterial universe behind all phenomena”–in Soothill and Hodous, following a system, 1934.

[14]. See the Appendix for the notation of Kokūji performed by Watazumi in 1968, and the transnotations of both versions by Bruno Deschênes.

[15]. For example, there is another version titled Fudaiji Kokū, Fudaiji being a temple somewhat half way between Kyoto and Tokyo. That version is totally different from Yokoyama’s version. Kokū is considered as one of the three “original” pieces for shakuhachi, the other being Kyorei and Mukaiji. See Johnson (2014).

[16]. https://en.wikipedia.org/wiki/Kokū_Nishimura, or https://www.komuso.com/people/people.pl?person=364. Accessed January 3^rd, 2022.

[17]. The Tozan school was influenced by Western music. For example, its notation uses boxes to represent measures, while keeping the katakana notation, and laying the score from right to left, from up to down. See Johnson (2014).

[18]. This short biography was provided by Sogawa Kinya, in an email dated January 3^rd, 2022.

[19]. Ramos, 2008. This short biography was provided by Alcvin Ryūzen Ramos in an email dated January 7, 2022.

[20]. Extracted in part from https://www.komuso.com/people/people.pl?person=700, accessed March 3^rd, 2022.

[21]. Published on YouTube (Sogawa, 2014) as a recording from an iPhone, this is the eleventh track of the CD Sogawa (2012), apparently published by Animul Records, entitled Take wo Fuku – Playing Bamboo (the information is taken from the web page http://www.animul.info/playing_bamboo.html–accessed May 1^st, 2022).

[22]. Nishimura’s version dates back to 1965 and was taken (on September 24, 2021) from https://www.youtube.com/watch?v=uGzng5BZRVU, which is reproduced from the third track on Face A of the LP (Nishimura, 1965).

[23]. Yokoyama Katsuya’s version was taken (on September 24^th, 2021) from a YouTube video: https://www.youtube.com/watch?v=83of323uRo4.

[24]. MacGregor’s version of Kokū used in the analysis was recorded “for a special CD prepared for an exhibition at the National Gallery Victoria, in Melbourne. The CD title was ‘Zen Golden Memories’ and was produced in 2007. It was not made available for public distribution. So now it can only be found on [his] YouTube channel [at] https://youtu.be/tKH41oaVobA (MacGregor, 2021). Ramos’ version is, as aforementioned, featured on track 2 of his CD “Zen Shakuhachi 2” produced by Canadian record label, Oliver Sudden Records.

[25]. Altogether, 66 analyses (including the 6 Kokū performances) have been published, and are listed and commented at https://analyses.foredofico.org/.

[26]. The lower part of the upper stripe contains a horizontal VU-meter above the CERMAA signature line.

[27]. The more the cursor is positioned to the left, the closer we are to the beginning of the piece; the more it is positioned to the right, the closer we are to the end of the piece.

[28]. (Shortened horizontal lines for a better view.) Left and right sides notations follow the North American usage, the central notation follows the French solmization. The tonic–here chosen as b–is in bold and in a different font from the other literal notes. The lettering of the notes of the scale changes with each octave from lower to upper case and vice versa, with the main octave notes being all in lower case.

[29]. This is meant to merely give guidelines to the listener/viewer, and not to give a precise position of the pitch.

[30]. The last VIAMAP analyses habitually used light green cursors, but the addition of the spectrogram, together with the choice of dark green for the graphic pitch line, made it difficult to distinguish the central cursor from the rest. From there came the decision to change the colour to another one, contrasting more with the surrounding graphic elements, in this instance a light blue colour (RGB = 148CEE). Similarly, while other techniques are indicated in black on the graphics, the muraikis, which are often mixed with other techniques, are depicted in purple (RGB = 6D108C).

[31]. This is a common procedure in the VIAMAP analyses and particularly fit to Aural music, which should always be transcribed (and analyzed) in relative pitch. In the case of the shakuhachi–which has a recently founded notation shown in THT 1 –, this is particularly adapted to the repertoire of the instrument as its notation refers to the fingerings, not to tones or absolute pitch.

[32]. We use “s_a” for “Seconds of the Analysis” to differentiate the analysis time from the video time (“s_v”).

[33]. In an email dated May 3^rd, 2022, Ramos informed Bruno Deschênes that he used a 2.4 shakuhachi for his recording of Kokū.

[34]. In an email correspondence dated April 29, 2022, MacGregor informed Bruno Deschênes that he used a 2.55 shakuhachi in one piece for his recording of Kokū.

[35]. This information is taken from the web page http://www.animul.info/playing_bamboo.html (accessed May 1, 2022), which says that Sogawa used a “2.5” for Kokū (which is the eleventh track on the CD).

[36]. We use “plain” instead of “natural” as commonly used in Western musicology.

[37]. See for example Beyhom, 2007, for more explanations, in particular p. 199, figure n° 3b.

[38]. This may be a Holderian or a Pythagorean comma: as is explained in the introduction, exact measurements in such comparisons are not relevant.

[39]. With major differences, as we will see, for Kitahara Kōzan and Nishimura Kokū.

[40]. This applies to all styles, modern or traditional.

[41]. For example, the mawashi-yuris are generally more rapid than the tate-yuris, and the latter generally result in a back-and-forth semitone in the rising direction.

[42]. To view the video analysis of this performance click here.

^[43] He studied, later on, under Okuda Atsuya and Tokuyama Takeshi. One of non-Japanese students’ advantages is that they can learn in more than one school, which Japanese would not usually do. In Japanese traditional arts, the relationship between teacher and pupils is one between master and disciples who cannot cross over to another teacher or school.

^[44] See footnote no. 58.

^[45] This remark is added by the authors of the article.

^[46] Moreover, we could say that Ramos has a systematic tendency to use muraikis.

[47]. To view the video analysis of this performance click here.

[48]. Two manual corrections had to be made for the graph of this analysis, at 131.7–132.9 for the first one (octave correction), and a second one at 638.2–639.2 s_a (residual sound pitch correction).

[49]. Tajima Tadashi was born in Sakai, Osaka in 1942. He began to study the shakuhachi with Sakai Chikuho II in the Kinko School. He later (at the age of 29) began studying with Yokoyama who, although originally also from the Kinko School, had studied a number of years with Watazumi who emphasized using very long flutes and added a lot of breathiness to the resulting sound. Yokoyama made this style popular among shakuhachi players and devotees, and Tajima is considered to be one of his best and most well-known students. (Taken with slight changes from https://www.komuso.com/people/people.pl?person=1198, accessed on the 30th of April 2022.)

[50]. This was confirmed by the performer to Bruno Deschênes in an email dated February 14, 2022.

[51]. Two other mawashi-yuris seem to be missing from Yokoyama’s score with MacGregor, such as at 337.6–341.5 s_a (but with maybe here a near-imperceptible vibrato), and at 431.8–433.2 s_a.

[52]. To view the video analysis of this performance click here.

[53]. This could be the result of his work with film music and Western musicians, as Sogawa lives in the United States and has students there, or the result of Yokoyama’s teaching (see Yokoyama’s analysis and the discussion in the comparisons between the six performances and in the conclusions) or, possibly, of Sogawa’s personal taste and aesthetics–maybe, only in this piece.

[54]. See also the analysis of Nishimura’s performance for the phrase at 387-396 s_a.

[55]. This could be the result of lips techniques used by the musician.

^[56] As a reminder: https://youtu.be/rT1aOR1SK44.

^[57] To view the video analysis of this performance click here.

^[58] While, today, the production of muraikis is an integral part of some styles of shakuhachi, in particular from the Yokoyama Katsuya’s school, Bruno Deschênes’ Japanese shakuhachi master, Kurahashi Yōdō II, told him during a lesson that, as a traditional player, he intentionally plays muraikis only in the known shakuhachi duet Shika no tōne. Its modern version implies the use of that particular sound. As for the more traditional repertoire he performs, he does not use muraiki, though it might occur at times without aiming for it.

^[59] Note that the recording of Nishimura’s Kokū dates back to 1965, with a lot of background noise which contributes to the production of parasite harmonics, artificially filling the spectrum of the produced sound.

^[60] Note also quarter-tone plays around the lower Fourth such as at 132–145 s_a, with an approximate third-of-the-tone span.

[61]. This koro-koro results from the particular technique of closing the first and second holes (from the bottom) in alternation.

[62]. Note also a portamento at 650.8–652.2, and another one around 656.5 s_a.

[63]. To view the video analysis of this performance click here.

[64]. The five holes are counted from down up.

[65]. Note, however, that while Nishimura avoids also the grouped nayashis, he frequently uses pairs of these in his performance–see Nishimura’s performance detailed analysis above.

[66]. 60–61 s_a is in fact a very short portamento. As for the possible one at 124.6 s_a. At 217–218 s_a, we can also notice a possible portamento, but this may rather correspond to a movement of the musician’s head or lips.

[67]. To view the video analysis of this performance click here.

[68]. As already explained in the articles relating to the VIAMAP, it is always better for such graphic analyses to work with high resolution recordings, at least the CD standards at 16 bits and 44 000 Hz.

[69]. See footnote 109.

[70]. Or 6 minutes and 32 seconds.

[71]. Or 17 minutes and 1 second.

[72]. As the mini-bars in the transcription show an undetermined rest time, the use of “phrase” instead of (or concomitantly with) “bar” is here fully justified.

[73]. For example, the mawashi-yuris are generally more rapid than the tate-yuris, and the latter generally result in a back-and-forth semitone in the rising direction.

[74]. Shakuhachi players refer to this technique as “glissando”, which is a common error among musicians.

[75]. Lee 1992, p. 246–247, adds that Watazumi did not value the product of his performance, i.e., the audible sound, but rather the process of the performance.

[76]. As for example with Sogawa Kinya. In an email (dated July 27th, 2022), he indicates that Yokoyama was not using scores during lessons. Sogawa was transcribing what he learned after each lesson, gradually creating a full score of the piece. For his performance of Kokū, as well as for his teaching, he relies on his own score. He never used Yokoyama’s score calligraphed by Furuya Teruo.

[77]. Lee, 1992, p. 246–247, 300–302. For the making of the hochiku (also written hocchiku or hotchiku), a short description is available (accessed July 31^st, 2022) at https://en.wikipedia.org/wiki/Hotchiku.

[78]. Bruno Deschênes acquired all Yokoyama’s scores in 2004 while training with Alcvin Ryūzen Ramos.

Analytical Approaches to World Musics

Analytical Approaches to World Musics

Beyhom_Deschênes 2025

AAWM Journal 13/No. 2 (2025)

Six Analyses of Kokū (虚空)

Amine Beyhom and Bruno Deschênes

Introduction[1]

Part I. Introducing the Shakuhachi and the Musicians

The techniques referred to in the analyses

The piece and the musicians

Presentation of the Six Musicians

Nishimura Kokū (1915–2002)

Kitahara Kōzan (1925–2015)

Sogawa Kinya (1962–)

Andrew MacGregor (1949–)

Alcvin Ryūzen Ramos (1969–)

Yokoyama Katsuya (1934–2010)[20]

Preliminary Comments on the Six Recordings

Part II. The Analyses

General Principles for the VIAMAP Analyses

Determining the Tonic and the Reference Pitch (the upper Fourth)

Six Analyses of Kokū (虚空)

Preliminary Remarks

Analysis of Alcvin Ryūzen Ramos’ performance of Kokū[42]

Analysis of Andrew MacGregor’s performance of Kokū [47],[48]

Analysis of Sogawa Kinya’s performance of KOKŪ [52]

Analysis of Nishimura Kokū’s performance of Kokū [57]

Analysis of Kitahara Kōzan’s performance of Kokū [63]

Additional Remarks

Analysis of Yokoyama Katsuya’s performance of Kokū [67]

Comparing the Six Analyses

1. Correspondences of the performances with the score of Yokoyama Katsuya

2. Execution of the groups of nayashis

3. Performing the koro-koros

4. The Accentuations

Intensity (and Tempo)

Quarter-tone accentuations

5. The portamento from E to G and back to E

6. Use of the half tones

7. Use of repetitions

8. Micro-Modulations to f# and F#

9. Additional modulations

10. Additional particularities with each performer

11. Duration of each version

Conclusions

Musician or musicologist?

APPENDIX: NOTATION, TRANSCRIPTION, SCORES AND TRANSNOTATIONS OF KOKŪ

The Shakuhachi Notation

Transcription of Kokū

REFERENCES

Analytical Approaches
to World Musics

Introduction^[1]

**Part I. Introducing the Shakuhachi and the Musicians**

Yokoyama Katsuya (1934–2010)^[20]

**Analysis of Alcvin Ryūzen Ramos’ performance of Kokū^[42]**

**Analysis of Andrew MacGregor’s performance of Kokū ^[47]^,^[48]**

**Analysis of Sogawa Kinya’s performance of KOKŪ ^[52]**

**Analysis of Nishimura Kokū’s performance of Kokū ^[57]**

**Analysis of Kitahara Kōzan’s performance of Kokū ^[63]**

**Analysis of Yokoyama Katsuya’s performance of Kokū ^[67]**