Subject: Foveon X3 sensor response, metamerism indices and matrix-compliance
Following-up an earlier discussion on ‘sensor design’, I came across this enlightening article on this general subject:
“The spectral characteristics of the X3 sensor response are similar in breadth to the human cone sensitivities, and have been shown to be an extremely accurate linear transform of these fundamentals. … The predicted color accuracy is represented by a metamerism index. With the selected color transformation matrix, the metamerism index is among the best of the collection of sensors sampled. The experimental results (of rendered, sRGB test images), the noise and color accuracy compare very well with similarly configured sensor systems.”
It’s not that I’d have any bias towards Sigma; however, this article also communicates the industry’s goals as well as related problems with surprising openness. For example, it seems that the experimental results (see Figure 9) do not perfectly support the top metamerism index calculated from the spectral characteristics (see page 4). Many colors are finally very well in the ballpark, whereas others show a noticeable dEab – upon using a ‘best fit’ matrix space for description.
As far as I understand (and highly appreciated from my perspective as a customer) the overall direction seems to be: color accuracy first; thus, laying the ground for a pleasing rendition via an appropriate tone curve (to compensate for dyn. range compression) and maybe a chroma boost depending on taste…
View All Messages in adobe.photoshop.camera.raw
Copyright © 2006 WatermarkFactory.com. All Rights Reserved.