Why is colour.sd_to_XYZ definition returning different results from the instrument?

[honhon199]

Hi,

First time using this program and wanted to understand where I'm going wrong with this:
currently I have tested a panel on an instrument which gives me the spectral data (400-700nm, 10 interval) the XYZ and LAB

The values for XYZ and LAB that the instrument provides are: 86.63, 91.80, 95.58, 96.74, -0.76 and 1.95 respectively

Using the spectral data and the following code:

import colour
cmfs = MSDS_CMFS['CIE 1964 10 Degree Standard Observer']
illuminant = SDS_ILLUMINANTS["D65"]
shape = colour.SpectralShape(400, 700, 10)
obs=colour.CCS_ILLUMINANTS['CIE 1964 10 Degree Standard Observer']['D65']
data = np.array(
    [
        45.42,
        73.35,
        85.37,
        88.85,
        89.43,
        89.86,
        90.33,
        90.61,
        90.88,
        91.15,
        91.46,
        91.55,
        91.67,
        91.93,
        91.92,
        92.12,
        92.28,
        92.23,
        92.26,
        92.10,
        92.19,
        92.08,
        91.85,
        91.76,
        91.57,
        91.33,
        91.22,
        90.81,
        90.34,
        89.79,
        88.54
    ]
)
sd = colour.SpectralDistribution(data, shape)
XYZ=colour.sd_to_XYZ(sd, cmfs, illuminant)/100

LAB=colour.XYZ_to_Lab(XYZ/100,obs)

print(XYZ,LAB)

gives me the results:

[ 86.6043296 91.79963466 95.46418938] [ 96.73834113 -0.80271132 2.0215199 ]

which doesn't match with what the instrument give me (above)

going the other way so taking the LAB the instrument gives and using LAB_to_XYZ gives me what I expect.

import colour
obs=colour.CCS_ILLUMINANTS['CIE 1964 10 Degree Standard Observer']['D65']

LAB= [96.74,-0.76,1.95]
colour.Lab_to_XYZ(LAB,obs)

array([ 0.86631035, 0.91803687, 0.95574975])

I initially thought it was the MSDS_CMFS choice but I tried all of them and the 1964 was the closest and the illuminant is correct.

[KelSolaar]

Hello,

Without seeing how the instrument does the integration, it is hard to know where the difference is coming from. It seems like though that the spectral distribution is given without spectral bandpass correction, try this and it should get you closer:

import colour
import numpy as np

cmfs = colour.MSDS_CMFS["CIE 1964 10 Degree Standard Observer"]
illuminant = colour.SDS_ILLUMINANTS["D65"]
shape = colour.SpectralShape(400, 700, 10)
obs = colour.CCS_ILLUMINANTS["CIE 1964 10 Degree Standard Observer"]["D65"]
data = np.array(
    [
        45.42,
        73.35,
        85.37,
        88.85,
        89.43,
        89.86,
        90.33,
        90.61,
        90.88,
        91.15,
        91.46,
        91.55,
        91.67,
        91.93,
        91.92,
        92.12,
        92.28,
        92.23,
        92.26,
        92.10,
        92.19,
        92.08,
        91.85,
        91.76,
        91.57,
        91.33,
        91.22,
        90.81,
        90.34,
        89.79,
        88.54,
    ]
)

sd = colour.bandpass_correction(colour.SpectralDistribution(data, shape)) / 100
XYZ = colour.sd_to_XYZ(sd, cmfs, illuminant)

LAB = colour.XYZ_to_Lab(XYZ  / 100, obs)
# [ 86.62746005  91.80419271  95.56581136]
# [ 96.740207    -0.76756712   1.95650694]

# Reference
# 86.63, 91.80, 95.58
# 96.74, -0.76 1.95

print(XYZ)
print(LAB)

[honhon199]

Unfortunately the only information I can get from the instrument is that its a 10degree observant and D65 illuminant! However using the code you provided with other data sets does give better alignment. Many thanks this was very helpful!