关于「实践与证伪方案:攻心文案检查器与预注册检验」,《耳内降噪、堵耳效应与听力努力》给出的定量结论是什么?

结论:第八章先对照降噪的宣传数字与用户整体听到的变化,再给出攻心文案检查器、经营者执行清单与预注册检验方案。书中以标注框区分已发表的证据、可运行的 Python 计算(20 段程序)、原典与作者的假说。

数据口径:书中的耳道尺寸、噪声谱、环路增益与认知参数均为示意;全部计算由书中程序生成;本书不是医疗建议,不宣称任何产品或聆听方式对听力、耳鸣或健康有任何益处。 第一章在合成噪声上运行最小均方与滤波 x 最小均方算法,说明教科书步长界 2/(L·P) 对高斯输入偏松:16 抽头时精确界是 0.1111,而次级通路模型的相位误差越过 90 度时算法发散。

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: 20

test dz n@80% n@90% sim@80%
H4 OE shallow - deep, own voice0.750 16 21 0.846
H5 SRT gain, ANC vs passive0.667 20 26 0.843
H6 pupil, mid - low intelligibility0.500 34 44 0.828
plan: recruit the n@90% column plus 15% for dropouts:
H4 OE shallow - deep, own voice25
H5 SRT gain, ANC vs passive30
H6 pupil, mid - low intelligibility51
查看完整验算程序
# Chapter 8. p82: Sample sizes for the pre-registered tests of this book.
# Paired designs, two-sided alpha = 0.05. Normal approximation with the
# small-sample correction n = ((z_a + z_b) / dz)^2 + z_a^2 / 2, where dz
# = mean difference / SD of differences. Then a fixed-seed Monte Carlo
# (4000 simulated studies, normal differences, z test on the sample SD)
# checks the power reached at that n. Effect sizes are the author's
# planning assumptions, stated before any data exist.
import math
import random
import statistics

ND = statistics.NormalDist()
TESTS = (
    ("H4 OE shallow - deep, own voice", 6.0, 8.0),
    ("H5 SRT gain, ANC vs passive", 2.0, 3.0),
    ("H6 pupil, mid - low intelligibility", 0.05, 0.10),
)


def n_for(dz, power, alpha=0.05):
    za = ND.inv_cdf(1 - alpha / 2)
    zb = ND.inv_cdf(power)
    return math.ceil(((za + zb) / dz) ** 2 + za * za / 2)


def sim_power(n, mean, sd, rng, reps=4000):
    za = ND.inv_cdf(0.975)
    hit = 0
    for _ in range(reps):
        xs = [rng.gauss(mean, sd) for _ in range(n)]
        m = statistics.fmean(xs)
        s = statistics.stdev(xs)
        if abs(m / (s / math.sqrt(n))) > za:
            hit += 1
    return hit / reps


if __name__ == "__main__":
    rng = random.Random(20261112)
    print("test                                   dz   n@80%  n@90%  sim@80%")
    for name, mean, sd in TESTS:
        dz = mean / sd
        n80, n90 = n_for(dz, 0.80), n_for(dz, 0.90)
        print("{:<36} {:>5.3f}  {:>5}  {:>5}  {:>7.3f}".format(
            name, dz, n80, n90, sim_power(n80, mean, sd, rng)))
    print("plan: recruit the n@90% column plus 15% for dropouts:")
    for name, mean, sd in TESTS:
        n90 = n_for(mean / sd, 0.90)
        print("  {:<36} {:>4}".format(name, math.ceil(n90 * 1.15)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《耳内主动降噪最小均方自适应滤波、堵耳效应低频骨导增益与信噪比认知恢复》第 8 章「实践与证伪方案:攻心文案检查器与预注册检验」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-164 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026164,
  author    = {Fan, Yafu},
  title     = {LMS Adaptive Filtering in In-Ear Active Noise Cancellation, Occlusion-Effect Bone-Conduction Gain and SNR-Driven Cognitive Recovery},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论