关于「实证闭环:品牌手写体规范、书写练习速度日志与证伪方案」,《书法运笔平滑度与分形维数》给出的定量结论是什么?
结论:第八章给出品牌手写体规范、书写练习速度日志的闭环、预注册方案与蒸馏小工具「笔势平滑度评分器」。书中以标注框区分已发表的证据、可运行的 Python 计算(21 段程序)、原典与作者的假说。
数据口径:书中轨迹、纸墨与评分参数均为示意;全部计算由书中程序生成;本书是非个人化的分析,不构成经营、设计或医疗建议。 第二章证明盒计数维数看不见运笔:同一条弧线以三种时序写出,维数都是 1.2251,平滑度却从 -7.653 落到 -11.140;示意的两组笔迹中,维数的效应量只有 -0.36(95 % 区间 -0.98 到 +0.27),平滑度的效应量是 3.74;随机游走的维数 1.704 已接近滴画研究报告的 1.72。
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: 21
| rate Hz | w | noise 0 | 0.02 mm | 0.05 mm | 0.1 mm |
| 100 | 1 | 100.0 | 81.6 | 59.9 | 42.7 |
| 100 | 5 | 100.0 | 99.9 | 99.5 | 98.3 |
| 100 | 9 | 100.0 | 100.0 | 99.9 | 99.9 |
| 100 | 21 | 100.0 | 100.0 | 100.0 | 99.9 |
| 200 | 1 | 100.0 | 33.2 | 21.3 | 12.2 |
| 200 | 5 | 100.0 | 96.4 | 82.5 | 66.9 |
| 200 | 9 | 100.0 | 99.6 | 97.5 | 92.2 |
| 200 | 21 | 100.0 | 100.0 | 99.9 | 99.5 |
| 500 | 1 | 100.0 | 8.1 | 2.9 | 1.9 |
| 500 | 5 | 100.0 | 39.1 | 24.6 | 17.5 |
| 500 | 9 | 100.0 | 66.3 | 43.6 | 31.9 |
| 500 | 21 | 100.0 | 97.0 | 85.5 | 69.1 |
| cost of smoothing: a real 9 Hz tremor of 0.4 mm at 200 Hz | |||||
| w = | 1 ( | 5 ms) | score 65.4 | ||
| w = | 5 ( 25 ms) | score 69.5 | |||
| w = | 9 ( 45 ms) | score 81.9 | |||
| w = 21 (105 ms) | score 100.0 |
查看完整验算程序
# Chapter 8. p83: sensor noise, sampling rate and the smoothing window
# Model: a finite-difference jerk multiplies position noise of size
# sigma by about sigma / h^3, so a cheap tablet can make a fluent
# stroke look shaky. One quintic sweep (150 mm in 0.5 s) is logged at
# 100, 200 and 500 Hz with Gaussian noise of 0 to 0.1 mm, smoothed by
# two passes of a moving average of w samples (w = 1 means none), and
# scored as in p81 against a reference put through the same pipeline.
# Each cell is the mean score of 20 noisy logs. Last block: how much
# smoothing hides a real tremor laid along the stroke. Noise levels
# and rates are illustrative; seed fixed.
import math
import random
def q(u):
if u <= 0.0:
return 0.0
if u >= 1.0:
return 1.0
return 10 * u ** 3 - 15 * u ** 4 + 6 * u ** 5
def smooth(v, w):
out = v
for _ in range(2):
half, src = w // 2, out
out = []
for i in range(len(src)):
lo, hi = max(0, i - half), min(len(src), i + half + 1)
out.append(math.fsum(src[lo:hi]) / (hi - lo))
return out
def dj(xs, h, w):
xs = smooth(xs, w)
n = len(xs)
sp = [0.0] * n
for i in range(1, n - 1):
sp[i] = abs(xs[i + 1] - xs[i - 1]) / (2 * h)
vmax = max(sp)
mov = [k for k in range(2, n - 2) if sp[k] >= 0.05 * vmax]
a, b = mov[0], mov[-1]
jj = [((xs[k + 2] - 2 * xs[k + 1] + 2 * xs[k - 1] - xs[k - 2])
/ (2 * h ** 3)) ** 2 for k in range(a, b + 1)]
dur, amp = (b - a + 1) * h, math.fsum(sp[a:b + 1]) * h
return math.fsum(jj) * h * dur ** 5 / amp ** 2
def log(rate, noise, rng):
n = int(0.7 * rate) + 1
return [150.0 * q((i / rate - 0.1) / 0.5) +
(rng.gauss(0, noise) if noise else 0.0) for i in range(n)]
if __name__ == "__main__":
rng = random.Random(20261113)
print("rate Hz w noise 0 0.02 mm 0.05 mm 0.1 mm")
for rate in (100, 200, 500):
for w in (1, 5, 9, 21):
ref = dj(log(rate, 0.0, rng), 1.0 / rate, w)
row = []
for noise in (0.0, 0.02, 0.05, 0.1):
reps = 1 if noise == 0.0 else 20
sc = [min(100.0, 100.0 * (ref / dj(log(rate, noise, rng),
1.0 / rate, w)) ** 0.25)
for _ in range(reps)]
row.append(math.fsum(sc) / reps)
print("{:>6d} {:>2d} ".format(rate, w) + "".join(
"{:>9.1f}".format(s) for s in row))
print("cost of smoothing: a real 9 Hz tremor of 0.4 mm at 200 Hz")
for w in (1, 5, 9, 21):
h = 1.0 / 200
ref = dj(log(200, 0.0, rng), h, w)
xs = log(200, 0.0, rng)
ts = [0.4 * math.sin(2 * math.pi * 9.0 * i * h)
for i in range(len(xs))]
xs = [a + b for a, b in zip(xs, ts)]
print(" w = {:>2d} ({:>3.0f} ms) score {:.1f}".format(
w, w * h * 1000, min(100.0, 100.0 * (ref / dj(xs, h, w))
** 0.25)))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《书法运笔最小加加速度轨迹、飞白墨量毛细流与笔势美感分形维数母典》第 8 章「实证闭环:品牌手写体规范、书写练习速度日志与证伪方案」(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-229 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026229,
author = {Fan, Yafu},
title = {Minimum-Jerk Brush Trajectories, Capillary Ink Flow in Flying-White Strokes and the Fractal Dimension of Stroke Beauty},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}