🚀 go-pugleaf

I hate group-think�

If we can�t understand the relationship between a sensor element and a sensor site, we can�t understand the question. Mixing up the sensor terms is like mixing PPI and DPI. People do it all the time. That�s why some folks can�t understand the basics of printing. Group-think leads to parroting out of context.

The filter over a sensor element determines the color (RGB) recorded, but a photodiode (sensor element) records luminance the value only. Color = wavelength and luminance. This recorded element data is gray scale. True, but this obscures the meaning of the data. Each channel represents a unique color (wavelength).

The dictionary defines mosaic as (abbreviated): a picture made with small bits of colored tiles. Demosaic is not a word in the dictionary. We can only infer the meaning. In a Bayer pattern the four elements (tiles) are demosiaced to a single tone. This tone includes missing colors such as yellow. The resulting tone is then encoded with three channels, RGB. The result is a pixel, from a single sensor site. Before this, they are sensor elements, not pixels. The accepted and better terms are �sel� for sensor elements and �pel� for printer elements. Demosiacing is not performed on pixels.

Chris said that the raw file �contains individual sensor element data, before demosaicing the color pattern�. Thank you Chris. That answered my question. Authoritatively and clearly. Thank you again.

I still did not understand Chris� comments about the number of elements per site. I emphatically believe these comments were incorrect.

The actual number of elements per site and how they contribute to the formation of a single tone is a function of the digital sensor design. Ultimately this design contributes to the range and fidelity of the tones. The Bayer pattern is not the only kid on the block. If the Fuji S2 has five elements I expect to see five channels in the raw data. If the Fuji S3 has six, I expect to see six. The Fuji S20 already has the new sensor similar to what will be in the S3, but it is not on the ACR supported cameras list (yet).

Just for grins, it would be possible and maybe even beneficial for some manufacturer to build a digital sensor that records only two metrics of light, wavelength and intensity. Such a device already exists. It is called a spectrometer. It is used in astronomy to examine the composition of distant stars. I used one for a high school project in 1959.

I�m outta here.