Although I have no empirical knowledge of this, it has been noted that
the newer form of dICE has eliminated most of the softening process.
Having studied a few scans which were done with the original dICE and
the IR channel from that process, I concluded that the softening was a
result of two factors. One, the IR was capturing more than just surface
defects and dust. It seemed to also capture grain. I was never able to
determine if this grain was actual residual silver left in the film
after processing of color film, or just "grainlike", but either way,
dICE was subtracting some "substance" of the image, rather than just
surface defects. Further, since our eyes tend to look for distinct
edges, like dye clouds or grain, removal of these might tend to make the
image appear even more soft. Secondly, the process of replacing areas
with defects (obviously, the software needs to "reconstruct" chroma and
luma in areas where it has removed a defect), is an interpolative
function, and if there is a lot of "stuff" being subtracted from the
image then there has to be a lot of "stuff" being faked back into it.
If the newer dICE finessed an improvement by reducing what would be
discarded from the image as a result of the IR scan to just real
defects, this could explain an improvement is apparent resolution with
the newer version.
Art
Rob Geraghty wrote:
> It occurred to me to wonder whether some of the softness in the use of "ICE"
> or infra-red cleaning may be due to the difference in focus needed. If
> you take photos with IR film, you have to set the focus differently to visible
> light. If the difference is significant in a scanner light path, maybe
> this accounts for some of the softness when removing dust and scratches?
>
> Rob
>
>
> Rob Geraghty harper@wordweb.com
>
>
>
>
> .
>
>
