5D Library › Singer Formant Clustering, Glottal Closure Quotients and Acoustic Authority Psychophysics · 中文

Chapter 7. “Moved by Things, It Takes Form in Sound”: The Record of Music, the Guanzi’s Five Tones and the Guiguzi’s “Quan”

Why read these sources here

The earlier chapters measured the voice; old Chinese texts described it. This chapter checks three passages and asks which correspondences can be computed. One point comes out of the check: the “Quan” chapter of the Guiguzi, often cited for adjusting the speed and weight of speech, does not contain the words for “slow and fast” or “light and heavy” in the received text; what it does say is that speech should be adapted to the listener.

The Record of Music: emotion shapes the voice

Source text

Liji, “Yue ji” (Record of Music)

凡音之起,由人心生也。人心之動,物使之然也。感於物而動,故形於聲。

其哀心感者,其聲噍以殺。其樂心感者,其聲嘽以緩。其喜心感者,其聲發以散。其怒心感者,其聲粗以厲。其敬心感者,其聲直以廉。其愛心感者,其聲和以柔。

(孙希旦 1989)

“All tones arise from the human heart; the heart moves because things move it. Moved by things, it takes form in sound.” The second passage lists six states of feeling and the voice each produces: grief makes it “sharp and fading”, happiness “relaxed and slow”, joy “open and scattered”, anger “coarse and harsh”, reverence “straight and restrained”, love “harmonious and soft”. Modern studies of vocal emotion describe the same kind of mapping with rate, pitch level, pitch range and spectral balance (Scherer 2003). The text is a description of listeners’ categories, not of acoustic measures, and the correspondence can only be tested by having listeners match the six descriptions to recordings.

The Guanzi: five tones by adding and removing a third

Source text

Guanzi, “Di yuan”

凡將起五音,凡首,先主一而三之。四開以合九九,以是生黃鐘小素之首以成宮,三分而益之以一,為百有八,為徵,不無有三分而去其乘,適足,以是生商,有三分而復於其所,以是成羽,有三分去其乘,適足,以是成角。

(黎翔凤 2004)

Starting from a string of 81 units for gong, adding a third gives 108 for zhi, removing a third gives 72 for shang, adding again 96 for yu, removing again 64 for jue. Part (a) of p13 turns these lengths into intervals.

Calculation

# Chapter 7. p13: Two classical ideas as arithmetic.
# (a) The five tones by "adding and removing a third" (san fen sun yi):
# string lengths 81 (gong), 108 (zhi), 72 (shang), 96 (yu), 64 (jue).
# Pitch is inverse to length; tones are folded into one octave above
# gong and given in cents (1,200 cents = one octave), next to the
# nearest equal-tempered semitone count.
# (b) Guiguzi, "Quan": "with the wise, rely on breadth ... with the
# poor, rely on profit". Suppose three audiences prefer pause ratios of
# 0.15, 0.21 and 0.28 (trust falls by 30 times the squared distance, as
# in p09). One speaker uses the single ratio best on average for an
# equal mix; another adapts to each audience.
from math import log2

TONES = (("gong", 81), ("shang", 72), ("jue", 64), ("zhi", 108),
         ("yu", 96))

if __name__ == "__main__":
    print("   (a) tone   ratio to gong   cents   nearest semitones")
    for name, length in TONES:
        r = 81 / length
        while r < 1:
            r *= 2
        c = 1200 * log2(r)
        print("   {:>9}   {:13.4f}  {:6.1f}   {:3d} ({:+.1f} cents)".format(
            name, r, c, round(c / 100), c - 100 * round(c / 100)))
    opt = (0.15, 0.21, 0.28)
    best = sum(opt) / 3
    loss = [30 * (best - o) ** 2 for o in opt]
    print("   (b) single best ratio {:.3f}; loss in trust by audience:"
          .format(best), " ".join("{:.3f}".format(x) for x in loss))
    print("       mean loss {:.4f}; adapting to each audience: 0".format(
        sum(loss) / 3))
#    (a) tone   ratio to gong   cents   nearest semitones
#         gong          1.0000     0.0     0 (+0.0 cents)
#        shang          1.1250   203.9     2 (+3.9 cents)
#          jue          1.2656   407.8     4 (+7.8 cents)
#          zhi          1.5000   702.0     7 (+2.0 cents)
#           yu          1.6875   905.9     9 (+5.9 cents)
#    (b) single best ratio 0.213; loss in trust by audience: 0.120 0.000 0.133
#        mean loss 0.0847; adapting to each audience: 0

Folded into one octave, the five tones lie at 203.9, 407.8, 702.0 and 905.9 cents above gong, within 8 cents of 2, 4, 7 and 9 equal-tempered semitones. For the speaking voice this gives a familiar scale for the size of prosodic movements: the declarative fall of 6 semitones in Chapter 4 is a little less than the step from zhi down to gong.

The Guiguzi: adapt the speech to the listener

Source text

Guiguzi, “Quan”

故口者,機關也;所以開閉情意也。

與智者言,依於博;與博者言,依於辯;與辯者言,依於要。與貴者言,依於勢;與富者言,依於高;與貧者言,依於利;與賤者言,依於謙;與勇者言,依於敢;與愚者言,依於銳:此其術也,而人常反之。

(许富宏 2010)

“The mouth is a mechanism by which feelings and intentions are opened and closed.” “With the wise, rely on breadth; with the learned, on argument; with the argumentative, on essentials; with the noble, on standing; with the rich, on loftiness; with the poor, on profit; with the humble, on modesty; with the brave, on daring; with the foolish, on sharpness: this is the art, and people often do the opposite.” The chapter’s advice is about content and stance, adapted to each listener. Part (b) of p13 puts it into the language of Chapter 5: if three audiences prefer pause ratios of 0.15, 0.21 and 0.28, the single ratio best on average is 0.213, which costs 0.120 and 0.133 points of trust with the outer audiences and 0.0847 on average; adapting costs nothing. The cost is small next to the effect of pitch in p09, which suggests that in voice, unlike in content, a single good habit loses little.

Hypothesis · not established

The author’s interpretation (not yet tested): listeners can match recordings of acted emotion to the six voice descriptions of the Record of Music above chance. What would show it wrong: in a pre-registered matching task with recordings of six acted emotions and the six classical descriptions translated without the emotion labels, listeners’ accuracy is not above one in six.

Summary

The Record of Music describes how feeling shapes the voice in listeners’ categories; the Guanzi’s five tones give intervals within 8 cents of equal temperament; the Guiguzi’s “Quan” asks the speaker to adapt to the listener, not to speak slowly or heavily. Computed in the terms of this book, adapting the pause ratio to the audience is worth little compared with pitch.

References

Scherer, Klaus R. 2003. “Vocal Communication of Emotion: A Review of Research Paradigms.” Speech Communication 40: 227 to 256.
孙希旦. 1989. 礼记集解. 中华书局.
许富宏. 2010. 鬼谷子集校集注. 新编诸子集成续编. 中华书局.
黎翔凤. 2004. 管子校注. Edited by 梁运华. 中华书局.

Fan, Y. (2026). Singer Formant Clustering, Glottal Closure Quotients and Acoustic Authority Psychophysics: Subglottal Pressure Dynamics, Pitch-Floor Stability and Pause-to-Phoneme Persuasion Ratios [歌手共振峰簇、声门闭合商与声学权威度心理物理学] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090258 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).