关于「听力努力:卡尼曼的容量模型与瞳孔的倒 U」,《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:第六章以卡尼曼的努力模型与瞳孔文献说明信噪比与听力努力之间的倒 U:同样 5 dB 的改善可以让玩具模型中的努力增加 0.568。
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。
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: 0.568
| SNR dB | p(intel) | effort p0=0.15 | effort p0=0.40 | spare |
| -14 | 0.007 | 0.053 | 0.000 | 0.947 |
| -12 | 0.023 | 0.071 | 0.001 | 0.929 |
| -10 | 0.076 | 0.171 | 0.001 | 0.829 |
| -8 | 0.223 | 0.630 | 0.022 | 0.370 |
| -6 | 0.500 | 0.500 | 0.440 | 0.500 |
| -4 | 0.777 | 0.223 | 0.223 | 0.777 |
| -2 | 0.924 | 0.076 | 0.076 | 0.924 |
| 0 | 0.977 | 0.023 | 0.023 | 0.977 |
| 2 | 0.993 | 0.007 | 0.007 | 0.993 |
| 4 | 0.998 | 0.002 | 0.002 | 0.998 |
| 6 | 0.999 | 0.001 | 0.001 | 0.999 |
| p0 = 0.15: effort peaks at SNR -7.5 dB, p = 0.281 | ||||
| p0 = 0.40: effort peaks at SNR -5.9 dB, p = 0.516 | ||||
| SNR -12.0 -> | -7.0 dB: effort +0.568 (p0 0.15), +0.171 (p0 0.40) | |||
| SNR | -7.0 -> | -2.0 dB: effort -0.563 (p0 0.15), -0.096 (p0 0.40) | ||
| SNR | -2.0 -> | 3.0 dB: effort -0.072 (p0 0.15), -0.072 (p0 0.40) |
查看完整验算程序
# Chapter 6. p61: A capacity-and-engagement toy model of listening effort.
# Kahneman (1973): effort is allocated to meet demand, within a limited
# capacity; motivational-intensity reasoning adds that effort is
# withdrawn when success looks unlikely. Toy model (illustrative):
# intelligibility p(SNR) = 1 / (1 + exp(-(SNR - SRT) / s)),
# SRT = -6 dB, s = 1.6 dB;
# demand D = 1 - p; engagement g = 1 / (1 + exp(-(p - p0) / 0.05));
# effort E = D g; spare capacity = 1 - E.
# p0 is the success level below which the listener disengages: 0.15 for
# a motivated listener, 0.40 for a tired one. The output is a shape to
# be tested against pupil data, not a measurement.
import math
SRT, S = -6.0, 1.6
def intel(snr):
return 1 / (1 + math.exp(-(snr - SRT) / S))
def effort(snr, p0):
p = intel(snr)
g = 1 / (1 + math.exp(-(p - p0) / 0.05))
return (1 - p) * g
if __name__ == "__main__":
print(" SNR dB p(intel) effort p0=0.15 effort p0=0.40 spare")
for snr in range(-14, 7, 2):
e1, e2 = effort(snr, 0.15), effort(snr, 0.40)
print("{:>7} {:>9.3f} {:>15.3f} {:>15.3f} {:>6.3f}".format(
snr, intel(snr), e1, e2, 1 - e1))
grid = [x / 10 for x in range(-200, 101)]
for p0 in (0.15, 0.40):
pk = max(grid, key=lambda x: effort(x, p0))
print("p0 = {:.2f}: effort peaks at SNR {:.1f} dB, p = {:.3f}".format(
p0, pk, intel(pk)))
for a, b in ((-12.0, -7.0), (-7.0, -2.0), (-2.0, 3.0)):
print("SNR {:>5.1f} -> {:>5.1f} dB: effort {:+.3f} (p0 0.15),"
" {:+.3f} (p0 0.40)".format(
a, b, effort(b, 0.15) - effort(a, 0.15),
effort(b, 0.40) - effort(a, 0.40)))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 6 章「听力努力:卡尼曼的容量模型与瞳孔的倒 U」(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}
}