[filmscanners] RE: 8 bit versus 16

["Paul D. DeRocco" <pderocco@ix.netcom.com>]

> From: Preston Earle
>
> I don't believe I'm confusing bit depth and resolution, I'm probably
> just not explaining myself very well. I'm trying to say that if a
> scanner can't (or at least "doesn't") scan adjacent pixels of uniform
> color as identical values in 8-bit precision, it doesn't matter what the
> other eight high-bits are. I don't have a tool to report 16-bit pixel
> values, but the 1x1-pixel point-source eyedropper in Photoshop shows
> *no* identical 4-pixel squares in a 2820ppi scan from my ScanDual II.
> For example, in an area where the pixels should be the same color, four
> adjacent pixels have values of 222r201g178b, 220r200g175b, 222r201g176b,
> and 200r200g175b. When these "combine" to make a print dot (or some
> other visible whatever) the average is 221r200(or201)g176b. If we knew
> the decimal values represented by the high bits, it wouldn't change the
> average. Thus it is irrelevant what the high-bit values are.

That's right.

> Maybe at some point in the process the 8th (and 9th, 10th, even 11th or
> 12th) bit *is* significant, and there is some "noise" added at some
> stage of conversion which eliminates the significance of the high bits.
> If so, it's valid to use high-bit data before that conversion. After
> that conversion, however, high-bit data is irrelevant. It's
> less-than-half-cent data in a dollar world.

As I said in another post, it appears to me, at least when scanning slides,
that film grain is the main source of noise, so there's no way around that.

--

Ciao,               Paul D. DeRocco
Paul                mailto:pderocco@ix.netcom.com

----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body