RE: filmscanners: Scanner resolution (was: BWP seeks scanner)

["Austin Franklin" <darkroom@ix.netcom.com>]

> >> The red channel, OTOH, is the one with the best contrast,
> >> almost always.
> >
> >I believe the green has the best contrast, and is the sharpest.
> Even on the
> >snippet you sent me, I believe green has far better contrast, and is the
> >sharpest.  Blue was muddy, and red is not near as sharp.
>
>
> Of your paragraph above (in general, not specifically
> that JPG "snippet") I would agree that blue is usually
> the muddy channel.  I can't agree on the rest of it
> though -- it's contrary to my own experience and to what
> I've read.

Hell, Rafe, don't believe what you read in books!  Half of it's wrong
anyway...

> Well, Monroy/Moody agree with you that the green
> channel has the best sharpness...

...not that half though ;-)

> but they also
> claim that red has the best contrast, which lines
> up with my own observation.

Contrast is VERY image dependant, where sharpness really isn't.  In certain
images, red may have better contrast...BUT you can't really use that
contrast because it's so damn fuzzy!  You can't separate those two from one
image plane.  Yes, you can apply this and that to the image to increase the
sharpness...but I believe it degrades the data...but that's another topic.

> Quite often, I'll apply USM to R and G, and then
> do Gaussian blur (say, 0.5 - 1.5 pixels radius)
> on the B channel, to soften the noise a bit.
>
>
> >> According to Biedney/Monroy/Moody ("Photoshop Channel
> >> Chops") the conversion to grayscale in Photoshop uses
> >> the following proportions:  30% red, 59% green, 11% blue.
> >
> >And that gives you a fuzzy conversion...
>
> So the wizards at Adobe got it wrong? <g>

I wouldn't call it 'wrong', I'd just say it isn't the best way of doing it.
Which brings me back to....well, you know what it brings me back to ;-)