2° and 10° observers in video and display analysis
by Florian Friedrich | June 2023
After Display Week 2023 in Los Angeles, I was asked which observer is more relevant to video signal analysis and display measurement: the 2° or the 10° observer. The choice affects many of our measurement and analysis tasks.
Observer models provide a reference for quantifying colour perception. The CIE 2° and 10° observers cover different visual fields and have different applications.
What the 2° and 10° observers describe
The CIE 1931 2° observer is based on experiments with a small field in the centre of vision, near the fovea, which is important for detailed central vision. The 10° observer, introduced in 1964, covers a larger visual field.
Display measurements, observers and modern colorimeters
Central vision plays an important role in many colour-critical tasks. The 2° observer provides an established colorimetric reference for video signal analysis and display measurements.
Modern colorimeters use physical filters matched to the 2° observer functions. These functions are also used to convert spectroradiometer measurements to coordinates such as X,Y,Z, x,y,Y or u′,v′,Y. I frequently use the Konica Minolta CA-410 for display measurements, including in our InnoPQ measurement system. Its matched filters allow faster measurements than a spectroradiometer in many situations. At the time of writing in June 2023, I also planned to test the Admesy Prometheus and compare it with my experience of the CA-410.
As measurement technology develops, the observer used to define the colour space remains essential to consistent results.
Wider colour gamuts
DCI-P3 and BT.2020 allow more saturated colours. This raises the question of whether a 10° observer would be preferable for wider gamuts, since the two models differ particularly for highly saturated colours.
Many colour-space definitions, including sRGB, DCI-P3 and BT.2020, use the 2° observer. It is an established reference for many colorimetric applications.
For comparable results, the observer used to analyse video signals or measure displays must match the observer used in the colour-space definition.
Display characteristics, viewing conditions and individual differences in perception also affect the result. The observer model is one part of this assessment; it does not capture every difference between viewers.
Practical recommendation
For video signal analysis and display measurements, the 2° observer is an established reference. The 10° observer is relevant to other tasks and visual fields. The essential point is to match the measurement to the colour-space definition.
Even with wider gamuts, use the reference on which the relevant specification is based. For many video and display applications, I therefore continue to recommend the 2° observer to support comparable measurements and consistent colour assessment.