>Ken Durling writes ...
>...
> > I have a slide that I've been spending hours trying to yield what I
> > can see through the loupe on the light table, but it's evading me.
> > It's a very high contrast sunset shot taken on Velvia, with one side
> > very dark under dense clouds, and the opposite side has brilliant -
> > one might say "blown out" - area of sunlight. Along the bottom of the
> > photo is a lot of city detail, ... and I'd like to retrieve it.
> >
> > My main problem has been trying to bring out all the detail in the
> > city - which is in the relatively dark area of the photo.
> > Secondarily, the finding a contrast range that doesn't blow out the
> > sunlit areas too severely, while not darkeneing the shadows too much.
> > ...
I didn't see the original of this as many posts are still missing, but as I
am becoming a specialist at doing high-contrast scans (due to poor
photographic technique and/or choice of subject), I add my reply to MIchael's.
If you look at the
first and 3rd images show two recent photos of bushfires that seem the same
as you are trying to do. Half the sky is very dark under dense smoke, and
the other is looking into the sun, full brightness sky, and there is some
shadowed foreground as well (bush in my case, not cityscape).
Mine were taken on negs so the treatment will not be the same for you, and
here I would suggest that what Michael suggested probably won't be
sufficient with a slide using a consumer scanner - you just can't get the
range onto the one scan.
My suggestion - scan once for the bright stuff, and separately for the
shadowed info, then combine them in PS etc. If you don't have more than 8
bits on your scanner, you are more or less forced to do this anyway.
In case it is useful, here is the detail of what I did for the fire photos...
Scan once into 16bit file and import to Photoshop. Make sure that you get
the whole brightness range into the scan, so no clipping at all. You might
have to use Vuescan to achieve this. The output of this is a very
low-contrast image and would look appalling if printed like that.
Duplicate the image in 16-bit, then use levels/curves to optimise one for
bright areas and one for shadows. The bright areas of the 'shadow" image
will be completely burnt out, and the dark areas of the 'bright' image will
be completely black.
Convert both to 8-bit and layer one on top of the other. Create a layer
mask on the top image and paint on it till you have selected the required
areas of each image, then flatten and do final adjustments for sharpen etc.
In your case with the slide, you would do two different scans with
different exposures, and after optimising each in PS, you then have the
additional step of aligning the two images in case they are slightly offset
when scanned due to mechanical variations.
In the case of negs it is truly amazing what a brightness range you can get
onto one shot. Absolutely astounding, and it is here that the Nikon analog
gain is invaluable so you can tweak it till you get the bits you want. In
the shots described above, it must have been 10 stops or so of variation
between the dark areas and the brightness of the sky, which I can get onto
a single scan. Remembering that a slide has a range of only 6 or so stops,
you have the single great advantage of using negs.
PS my third photo also has foreground detail but I chose to leave it as
silhouette because it looked better.
HTH,
Julian
----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or
body
