Archived from groups: rec.photo.digital (
More info?)
"Dave Martindale" <davem@cs.ubc.ca> wrote in message
news:cteshr$inq$1@mughi.cs.ubc.ca...
> Ron Hunter <rphunter@charter.net> writes:
>
> >Digital prints will always look digital, if you magnify them enough
> >since the pixels are square (or rectangular). Grains in the film medium
> >are of a random shape, so you can always tell a digital photo, if you
> >can magnify it enough.
>
> That's not true, if the device making the print is working at high
> enough resolution. The pixels can be treated as point samples of a
> continuous image, and the image can be resampled to many more pixels all
> different (so there are no square or rectangular edges to see).
Quibble: the printer resolution doesn't matter: all that matters is that the
source file is intelligently upsampled to the native resolution of the
printer. At which point, someone taking a loupe to the print only sees the
printer technology, not the imaging technology.
> For this to work well, the native resolution of the output device has to
> be several times that of the input data. If the digital printer is
> operating at 300 or 400 PPI, you can see the individual pixels with
> magnification. But if the printer is operating at something like 2000
> PPI, you aren't going to see pixel edges.
>
> You just need an output device pixel density too small for you to see
> at whatever magnification you are using (or a blurry enough spot that
> you can't distinguish neighbour pixels from each other). You can't
> normally see pixel boundaries on television even when the source is a
> digital video recorder, for example.
Still, you've got the right idea here: there's no need to see the imaging
technology in the print. You can do the same thing with film. Scan at a
grossly high resolution, NeatImage heavily, downsample to a much lower
resolution. This gives an image with clean pixels indistinguishable from a
dSLR image. Now interpolate up to your heart's content. Look ma, no grain,
no pixels.
(Whether or not this is a good idea is an arguable point. There's probably a
slight loss of detail. I'd rather use a larger format than see grain, so the
above approach supports both the finely detailed smooth-toned lower
magnification enlargements I prefer as well as posters, but if you insist on
squeezing every last bit of detail from miniature formats, putting up with
the grain may make sense.)
The basic idea is to extract the image content and to represent that at the
native resolution of the printer. Then all you will see is the printing
technology, not the imaging technology, when you pull out your 10x loupe.
David J. Littleboy
Tokyo, Japan