关于「一张脸上的两种色温:混合光的色度、白平衡偏差的量化与校色纸的代价」,《布光平方反比与色温混合》给出的定量结论是什么?

结论:第六章计算两种色温在一张脸上的混合:日光份额一半时相关色温为 4142 K,不是开尔文平均的 4400 K,混合点落在普朗克轨迹下方;三比一时把白平衡设在 3809 K 可使亮暗两面偏差相等。

数据口径:书中参数为示意;全部计算由书中程序生成;本书属于非个人化经营建议。 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。

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: 4142, 4400

share of 5600 K light in a 3200 K + 5600 K mix
wMcCamynearest-BBmired-meanDuv
0.00323532003200+0.0000
0.10339233563343-0.0009
0.25365436193584-0.0019
0.50417241424073-0.0026
0.75481347934716-0.0019
0.90527152565209-0.0009
1.00561256005600+0.0000
arithmetic mean of the kelvins at w = 0.5: 4400 K
查看完整验算程序
# Chapter 6. p61: two colour temperatures on one face.
# Tristimulus values add, so a mix of a 3200 K and a 5600 K blackbody
# with illuminance shares (1 - w, w) is X = (1-w) X1 + w X2 etc., each
# source scaled to Y = 1 first. Chromaticity via the fitted CIE 1931
# functions of p51; CCT by McCamy; Duv = signed distance in CIE 1960
# (u, v) to the nearest blackbody (golden-section search in mired),
# positive above the locus (towards green). Mired mean for comparison.
import math

LOBES = {"x": ((0.362, 442.0, 0.0624, 0.0374),
               (1.056, 599.8, 0.0264, 0.0323),
               (-0.065, 501.1, 0.0490, 0.0382)),
         "y": ((0.821, 568.8, 0.0213, 0.0247),
               (0.286, 530.9, 0.0613, 0.0322)),
         "z": ((1.217, 437.0, 0.0845, 0.0278),
               (0.681, 459.0, 0.0385, 0.0725))}
C2 = 1.4388e-2
WL = [360.0 + i for i in range(471)]


def cmf(key, lam):
    out = []
    for a, b, g, d in LOBES[key]:
        t = (lam - b) * (g if lam < b else d)
        out.append(a * math.exp(-0.5 * t * t))
    return math.fsum(out)


BAR = {k: [cmf(k, w) for w in WL] for k in ("x", "y", "z")}


def planck_xyz(temp):
    sp = [(w * 1e-9) ** -5 / math.expm1(C2 / (w * 1e-9 * temp)) for w in WL]
    xyz = [math.fsum(s * b for s, b in zip(sp, BAR[k])) for k in "xyz"]
    return [v / xyz[1] for v in xyz]


def uv(xyz):
    d = xyz[0] + 15.0 * xyz[1] + 3.0 * xyz[2]
    return 4.0 * xyz[0] / d, 6.0 * xyz[1] / d


def mccamy_xyz(xyz):
    t = math.fsum(xyz)
    x, y = xyz[0] / t, xyz[1] / t
    n = (x - 0.3320) / (y - 0.1858)
    return -449 * n ** 3 + 3525 * n ** 2 - 6823.3 * n + 5520.33


def duv(xyz):
    u, v = uv(xyz)

    def dist(m):
        pu, pv = uv(planck_xyz(1e6 / m))
        return math.hypot(u - pu, v - pv)
    a, b = 50.0, 600.0
    g = (math.sqrt(5.0) - 1.0) / 2.0
    for _ in range(60):
        c, d = b - g * (b - a), a + g * (b - a)
        if dist(c) < dist(d):
            b = d
        else:
            a = c
    m = 0.5 * (a + b)
    pu, pv = uv(planck_xyz(1e6 / m))
    sign = 1.0 if v > pv else -1.0
    dv = sign * math.hypot(u - pu, v - pv)
    return (0.0 if abs(dv) < 5e-5 else dv), 1e6 / m


def mix(w, t1=3200, t2=5600):
    a, b = planck_xyz(t1), planck_xyz(t2)
    return [(1 - w) * p + w * q for p, q in zip(a, b)]


if __name__ == "__main__":
    print("share of 5600 K light in a 3200 K + 5600 K mix")
    print("   w     McCamy   nearest-BB   mired-mean    Duv")
    for w in (0.0, 0.1, 0.25, 0.5, 0.75, 0.9, 1.0):
        xyz = mix(w)
        dv, tb = duv(xyz)
        mm = 1e6 / ((1 - w) * 1e6 / 3200 + w * 1e6 / 5600)
        print("  {:4.2f}   {:6.0f}     {:6.0f}       {:6.0f}     {:+.4f}"
              .format(w, mccamy_xyz(xyz), tb, mm, dv))
    print("arithmetic mean of the kelvins at w = 0.5: {:.0f} K".format(
        0.5 * (3200 + 5600)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《摄影棚布光平方反比与面光源衰减、色温混合与光比代数母典》第 6 章「一张脸上的两种色温:混合光的色度、白平衡偏差的量化与校色纸的代价」(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-325 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026325,
  author    = {Fan, Yafu},
  title     = {Inverse-Square Lighting and Area-Source Falloff in the Studio: Mixed Colour Temperatures and the Algebra of Lighting Ratios},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论