关于「实践与预注册试验」,《声音说服力与人际同步》给出的定量结论是什么?

结论:在这段虚构的录音中,9 个陈述句里有 6 个以超过 1 个半音的下降结尾,另外 3 个以平调或轻微上扬结尾;唯一的问句以上扬结尾,这是恰当的。5 个关键点中只有 3 个之前有 0.3 秒以上的停顿。平均语速约为每秒 5.5 个音节(每分钟约 330 个),句子之间的变化系数约为 11%,变化不大。麦克风从 40 厘米移到 10 厘米后,人声电平提高 12 dB,信噪比从 34 dB 提高到 46 dB。

数据口径:书中以标注框区分已发表的证据、可运行的 Python 计算(14 段程序,全部核对;书中的声音均为程序合成)、原典与作者的假说。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.3, 10, 11, 34, 40, 46

1 final f0 falls: 6/9 declaratives end with a fall > 1 semitone (median fall 2.0 st); the question ends rising, as it should
2 key points with a silent pause >= 0.3 s before them: 3/5
3 syllable rate: mean 5.5 syl/s (332 per minute), coefficient of variation 11% across phrases
4 microphone at 40 cm: speech -24.0 dBFS, noise floor -58.0 dBFS -> signal-to-noise 34 dB
4 microphone at 10 cm: speech -12.0 dBFS, noise floor -58.0 dBFS -> signal-to-noise 46 dB
查看完整验算程序
# The four-step acoustic checklist applied to a (fictional) 2-minute rehearsal recording, annotated by
# hand or with a tool such as Praat: one row per phrase.
#   (end of phrase: 'decl' declarative / 'q' genuine question; f0 at last peak and at end, Hz;
#    syllables; phrase duration s; is a key point; silent pause before it, s)
import math, statistics
phrases = [("decl", 132, 118, 14, 2.6, False, 0.0), ("decl", 128, 131, 11, 1.9, True, 0.1),
           ("decl", 135, 120, 16, 3.0, False, 0.4), ("q", 125, 150, 8, 1.6, False, 0.3),
           ("decl", 140, 122, 12, 2.1, True, 0.7), ("decl", 130, 134, 18, 2.8, False, 0.2),
           ("decl", 138, 119, 10, 2.0, True, 0.6), ("decl", 127, 129, 15, 2.2, True, 0.0),
           ("decl", 133, 117, 13, 2.5, False, 0.3), ("decl", 136, 120, 9, 1.9, True, 0.8)]
st = lambda a, b: 12 * math.log2(a / b)
decl = [p for p in phrases if p[0] == "decl"]
falls = [st(p[1], p[2]) for p in decl]
print(f"1 final f0 falls: {sum(f > 1.0 for f in falls)}/{len(decl)} declaratives end with a fall > 1 semitone "
      f"(median fall {statistics.median(falls):.1f} st); the question ends rising, as it should")
keys = [p for p in phrases if p[5]]
print(f"2 key points with a silent pause >= 0.3 s before them: {sum(p[6] >= 0.3 for p in keys)}/{len(keys)}")
rates = [p[3] / p[4] for p in phrases]
print(f"3 syllable rate: mean {statistics.mean(rates):.1f} syl/s ({statistics.mean(rates) * 60:.0f} per minute), "
      f"coefficient of variation {statistics.pstdev(rates) / statistics.mean(rates):.0%} across phrases")
# 4 microphone distance and noise floor: speech peaks and noise floor read from the recording (dBFS)
for dist_cm, speech_dbfs, floor_dbfs in ((40, -24.0, -58.0), (10, -12.0, -58.0)):
    print(f"4 microphone at {dist_cm} cm: speech {speech_dbfs} dBFS, noise floor {floor_dbfs} dBFS"
          f" -> signal-to-noise {speech_dbfs - floor_dbfs:.0f} dB")

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《声音说服力、韵律信息论与人际生理同步》第 8 章「实践与预注册试验」(版本 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.23753267

DOI: 10.5281/zenodo.23753267 · 阅读本章所在专著 · 5DL-B-018 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026018,
  author    = {Fan, Yafu},
  title     = {Acoustic Persuasion, Prosodic Information Theory and Interpersonal Physiological Synchrony},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  doi       = {10.5281/zenodo.23753267},
  url       = {https://doi.org/10.5281/zenodo.23753267}
}

本书其他章节结论