What's wrong with this picture?

[arnulf]

... Besides, I always thought that underexposure would lead to more black, not no black...

 

[Gabriel M.A.]

What is this? Anyone?

At first glance, and I'd be more certain if I saw the original scan file, it looks to me like you either set the sharpening on your scanning software way too high, and/or you scanned this with a color film profile rather than a B&W film profile which caused mayhem on high-contrast gradations from one edge to the other (if I knew the term for it, I'd tell you; sorry).

*and*, on top of that, looks like this particular negative suffers from reticulation with evident scanner noise. That'll make all of the above look worse.

I also suggest turning any ICE settings off (yes, including "shadow recovery"); I know you turned the "IR" ICE off (otherwise you'd have horrible highlights and higher midtones), but try turning the rest of it off.

 

[Gabriel M.A.]

It's actually more the upper part of the picture that disturbs me. Sure the light is a little low, but not that much, do you think. Look at this one. Much lower light, but the shadow areas are at least black, not noisy grey.

Oohh... you've got different moving ducks disturbing the pond...so to speak.

I think then you're a victim of the Autobrightness setting. "Properly" exposed, or frames exposed within a few stops fare very well, but if it's underexposed a lot, the software will unwittingly bring out a lot of sensor noise (also "masking" as grain) into your shadows.

Try doing a scan of your first shot above with manual brightness, and set it to normal (no + or - diff.); it may come out pretty dark compared to your first scan.

 

[arnulf]

Ahh, thanks a lot Gabriel! "Auto brightness", God dammit!!! I just had a look at the two baby pics in the scanner's preview, and sure enough: the second one is a little darker. Adjusting the brightness gives that noise like texture, more than when adjusting the curves. Hmmm, wonder what the difference between the two actually is...

Anyway, I think I got my answer here, Gabriel. Thank you very much!!

Arnulf

 

[Gabriel M.A.]

Anyway, I think I got my answer here, Gabriel. Thank you very much!!

No problem; you're welcome 🙂

I've run into all different walls you can think of, trust me. It took me almost a whole month to get my Coolscan 5000 to scan Kodachrome 64 correctly with IR ICE. People say it can't be done; oh, it can be.

The only one I haven't been able to tackle is using IR ICE with true B&W. I let Silverfast's pretty neat dust removal technology do its job, at the expense of some detail loss. It's not 100% proof, though. However, I hear some Vodka is.

 

[arnulf]

Well, now I also know about Silverfast. Downloading it now for closer inspection tomorrow.

Very excited about that Vodka. That'd be like two birds...

Thanks again

 

[Jonathan R]

I'll be fascinated to know whether this works. Arnulf, please keep us posted! Even if Gabriel is right about the autobrightness function, it seems to me there must be some difference between the two later baby negs, and that the Coolscan software is trying to compensate for it. If the second one was a properly exposed negative, the scanner with its automatic settings wouldn't be struggling to create tones in the way that it clearly is.

I still say you have to look directly at the negatives for your answer, not at the scans. One or two stops is not a small difference in exposure: it's a matter of a half or a quarter of the light hitting the film. In these indoor conditions I suspect that you'd been struggling to achieve a hand-held picture at all with Tri-X at 400 ISO. There surely can't have been the margin to trim off 2 stops of exposure? Anyway, you may not have consciously changed the exposure: someone could have stepped into the light behind you?

 

[Bobfrance]

I reached a similar conclusion to Gabriel and turned all my auto scanning assistance off. I found that the net effect of it's efforts to sharpen up my scans was that it made the grain much more pronounced. I now do all tweaking myself in Photoshop.