Thank you, I only understood the wavelengths. Instead of 0, 1, 2 at the end the original images have some two-digit numbers, different for different sets. I had no idea what they meant, but I had to standardize the names to simplify the task, so I just replaced them with sequential numbers. I think you are right and it has something to do with the light source, it never occurred to me before. As far as I understand, for a rough surface changing the direction and the apparent size of the light source would highlight different details, this can be very useful.
The original TIFFs are 16bits per channel, but only something like 11 bits contain useful information (the range of values is from 0 to 2^11 - 1), if I remember it correctly. The processing code takes the cropped part, sets the darkest pixel to black and the brightest to white, applies gamma of 2.0 and then linearly squeezes the result into 8 bits. This normalization works fine for most images, but for the edges of the folios when parts of the crop contain bright or black objects from the area around the book, the page itself could become low contrast.
The original TIFFs are 16bits per channel, but only something like 11 bits contain useful information (the range of values is from 0 to 2^11 - 1), if I remember it correctly. The processing code takes the cropped part, sets the darkest pixel to black and the brightest to white, applies gamma of 2.0 and then linearly squeezes the result into 8 bits. This normalization works fine for most images, but for the edges of the folios when parts of the crop contain bright or black objects from the area around the book, the page itself could become low contrast.