Re: MCN-L Digitization procedures

[Roy Berns]

I have been following this thread with interest. I sense that resolution is being defined by file size. Have I doubled my sensor resolution by interpolating to increase file size, etc.? The answer is no. Suppose I have two image files, 100 MB and 200 MB. The camera was out of focus with the 200MB image. Its effective resolution is probably lower, despite the larger file size. If I store an image as 16 bit rather than 8 bit per pixel per channel, the file size is larger. This doesn't increase resolution. As so forth. Resolution needs to be determined based on measurement, for example, determining a imaging system's MTF (equivalent to slanted edge SFR). If my resolution increases at a cost of increasing image noise, how does this factor in? Evaluating a camera system requires analysis, not just reading specs and looking at file size. Although file size and resolution may correlate, all other things being equal except the number of pixels on the sensor, it is easy to forget this caveat. Furthermore, this caveat is never true in practice. This can only be confirmed by experimentation. After all, new sensors have more pixels for the same sensor area; this is not equivalent to more pixels over a larger sensor area.
<x-tad-bigger>
Roy

</x-tad-bigger>Roy S. Berns, PhD
R. S. Hunter Professor in Color Science, Appearance, and Technology
Center for Imaging Science - Color Science Building 18
Munsell Color Science Laboratory
Rochester Institute of Technology
54 Lomb Memorial Drive
Rochester, NY 14623
585-475-2230 Fax: 585-475-4444 Mobile: 585-719-5621
http://www.cis.rit.edu/people/faculty/berns
http://www.cis.rit.edu/mcsl/