How does the Y component in CIE XYZ relate to physical luminance?

Asked 2/26/2020

5 views

2 answers

0

I understand that in CIE XYZ and xyY, the Y component is often described as luminance or relative luminance. What I’m trying to understand is how that relates to luminance in the photometric sense.

If I measure the reflected spectral radiant power from a surface and compute its luminance, then photograph that surface and convert the image data to XYZ, how should I interpret the resulting Y value? Is Y directly proportional to physical luminance, or only a relative value unless the camera/image pipeline is calibrated? In other words, when do the XYZ values retain real physical units, and when are they just normalized color coordinates?

Originally by Photography Stack Exchange contributor. Source · Licensed CC BY-SA 4.0

Photography Stack Exchange contributor

6y ago

2 Answers

2

Luminance is basically Radiance weighted by the Human Visual System (HVS), i.e. Power per Unit Area per Unit Solid Angle. Whether it is given in relative or absolute units depends on how the values were measured, absolute values require a calibrated reference light source and/or measuring device.

If you have some absolute spectral radiometric measurements and need the corresponding absolute photometric quantity, you need to integrate them to CIE XYZ tristimulus values while using a normalization factor k equal to Km = 683lm.w-1, the maximum spectral luminous efficacy of the HVS. 1 watt of light at 555nm equals to 683 lumens. Y in that context would be an absolute measure of Luminance.

Colorimetry 4th Edition has most of the relevant details, the 3rd edition can be found online and has very similar content.

Originally by user20809. Source · Licensed CC BY-SA 4.0

user20809

6y ago

0

AI Answer

Generated from our catalog & community — verify before relying on it.

The connection is real: in CIE XYZ, the Y tristimulus value is defined to track photometric luminance because it uses the human visual response weighting.

If you start from absolute spectral radiometric data, you can compute XYZ by integrating the spectrum against the CIE color-matching functions. With the proper normalization factor (including the photometric constant 683 lm/W at 555 nm), the resulting Y can represent absolute luminance.

But in most images, XYZ values are not absolute. A camera image usually has unknown exposure, tone mapping, white balance, and scaling, so converting that image to XYZ gives a relative Y, not luminance in cd/m².

So:

  • From calibrated spectral measurements: Y can correspond to physical luminance.
  • From a normal photo converted to XYZ: Y is usually only proportional to scene luminance up to an unknown scale, if at all.

The physics do not disappear; the units are lost only when the capture/reproduction chain is not radiometrically or colorimetrically calibrated.

UniqueBot

AI

6y ago

Your Answer