Bake-off, round 2 ...

[dmr]

Here's another comparison between scanning a B&W negative (again, old Panatomic-X) as a B&W negative and a color positive.

On the left are the as-is and the cropped slightly and adjusted one scanned as a B&W negative. On the right, inverted-only and then adjusted one scanned as a color positive. Yes, they are visibly different.

Both have highlights close to absolute white and shadows close to absolute black.

By bringing them up side by side and adjusting only the center level sliders, I can get them to look very very close to each other, but not exactly. I have a feeling that the real difference is due mostly to linearity.

Comments?

 

[ed1k]

I have a feeling that the real difference is due mostly to linearity.

Yes, you're right. For me, autoexposure sometimes cuts off highlights and shadow details too severe from the histogram, but works reasonably well almost all the times. That's the only reason I use manual exposure mode with scanner. It shouldn't really matter when and which software inverts image to positive. You can't probably control over linearity (or characteristic curve) of CCD while scanning, that's why it's worth to keep 16 bit deep sampling. Later in photoshop you can use "curves" to adjust the image to the look in a way you like. Yes, that's what I said - your eye (and monitor calibration... if you're going to send images to printer you're going to learn a lot more) is a judge for final look and linearity. Make whatever you think is best to express the subject (in a wet darkroom you'd play with papers of different grades or filters for multigrade paper and different paper developers and different paper types and zillion of different things to get that result).
I hope you learnt quite a bit and thanks for bringing this to the discussion.
Eduard.