Dslr scanning conundrum

[lukitas]

I'm scanning my black and white negatives with a nikon D3100, the kit zoom and a macro ring, fixed on a bench, with the head of a Durst 670 enlarger as light table.

After a lot of going back and forth between the enlarger and the computer (and a few helpful hints from rff members), I determined that the best exposure for shooting the negs is about one and two thirds of a stop over.
If the matrix metering determines 1/60th, an exposure of 1/20th will give the most usable raw file. Even so, the flat on the right side of the histogram (the transparent parts of the negtive) is longer than the left side.

That is my first question : Why? It doesn't seem logical. The negative being a transparent medium, middle grey should be middle grey, not nearly two stops under. Is it something about fitting the density curve of a negative to the sensors'?

The second question is this. The histograms of the scans I make show flats on both sides: the darkest and the brightest densities are way into the grey.
Is it normal that a negative should show a smaller range of densities than the sensor is capable of? And if so, is there a way to force the camera to pull the white and black points out to the maximum?

Full disclosure : as the viewfinder on the dslr doesn't show full frame, I frame the scan to take in a little of the negative carrier, which is as black as it can get, and the gate of the negative carrier is a bit larger than the negative, so I have full white on the borders of the negative.
Would that influence the way the camera calculates exposure?

 

[mdarnton]

I can answer the second question, and that is that the density range of a B&W negative is quite a bit smaller than a DSLR is capable of showing, probably only about six stops.* So what you're seeing in the histogram is the 6-stop density range of the negative stuck in the middle of the 11-stop range that the camera could handle. Shoot a slide, with it's greater density range, and you'll see the histogram fill up.

*(Don't be confused about this: the whole 10 or 12 or 15 stop range of a normal real world subject gets compressed into the film's comfort range, and then a normal paper grade turns that compressed range into a black to white range on the paper (which still covers a smaller brightness range than the original subject, but looks good to us, nevertheless.)

If you use curves instead of levels to straighten out the tonal range, you'll see that the resulting curve is even stranger than levels lead you to believe--a really low dip. I don't really know what's going on, but I do know that exposing considerably "to the right" makes the correction curve much more normal-looking than putting the bulk of the histogram right in the center.

 

[f16sunshine]

I won't try and answer your question but ratger suggest that you tether the camera to Aperture3 or LR4/5 or the like.
You will have an much easier go at it .

Cheers!