Pixels per inch

Archived from groups: alt.comp.periphs.scanner (More info?)

On Wed, 27 Apr 2005 21:08:26 GMT, "N! Xau"
<nxauBASTACONLOSPAM@hotmail.com> wrote:

>Hi,
>
>I want to scan some color pictures, size 4x6 inches, printed by a lab.
>My scanner can be set up to 9600 pixels per inch, but the default (and also
>automatic setting) is 200 ppi.
>
>What is your suggestion about it? It may seem a dumb question but I see the
>file size increase a lot increasing the ppi #. Also I get larger pictures,
>but at a certain point (I would say between 200ppi and 300ppi but my eye's
>not an expert one) I start seeing losses of quality in the pics.
>
>File sizes are:
>200ppi: 287 Kb
>300ppi: 698 Kb
>600ppi: 2,1 Mb
>
>Let's say I don't care how the file size is, at the moment. What could I do
>with a larger picture if the quality is proportionally lower?
>I mean, if I scan at 200ppi, then I resize the pic up to the dimension of
>the file I get at 300ppi, do I get the same quality? A lower one? A higher
>one?
>
>Please help me to clear the clouds.
>thanks

Scan at the scanner's optical resolution. For most modern scanners,
that's 300dpi or on some of the higher end ones 600dpi -- check the
manual. Any higher than optical resolution is pointless because
you're not getting any more information, the driver code is just
interpolating the pixels in between.

Doing that gives you the best image quality and speed. You can always
resample it to a lower resolution if you need to.

If you intend to print them, start with the scanner's optical
resolution, and then resample the image down to the printer's native
resolution (usually 150-300dpi). Most printer drivers will have
options to do this for you, but they're slower, and some create weird
patterns in the final print. Doing it in a photo editor program
before printing is better.
---------------------------------------------

MCheu

Archived from groups: alt.comp.periphs.scanner (More info?)

In article <ebTbe.68392$IN.1185633@twister2.libero.it>, N! Xau
<nxauBASTACONLOSPAM@hotmail.com> writes
>Hi,
>
>I want to scan some color pictures, size 4x6 inches, printed by a lab.
>My scanner can be set up to 9600 pixels per inch, but the default (and also
>automatic setting) is 200 ppi.
>
>What is your suggestion about it? It may seem a dumb question but I see the
>file size increase a lot increasing the ppi #. Also I get larger pictures,
>but at a certain point (I would say between 200ppi and 300ppi but my eye's
>not an expert one) I start seeing losses of quality in the pics.
>
>File sizes are:
>200ppi: 287 Kb
>300ppi: 698 Kb
>600ppi: 2,1 Mb
>
>Let's say I don't care how the file size is, at the moment. What could I do
>with a larger picture if the quality is proportionally lower?
>I mean, if I scan at 200ppi, then I resize the pic up to the dimension of
>the file I get at 300ppi, do I get the same quality? A lower one? A higher
>one?
>
>Please help me to clear the clouds.
>thanks

Everything you ask above, and more, is answered at:
http://www.scantips.com/

In particular, read the page http://www.scantips.com/basics08.html
which addresses your particular issue in detail, and explains why the
experience you relate above is typical and expected.

Note that Wayne misuses dpi for ppi throughout and expects the reader,
however new to the topic, to be able to decipher the distinction. He
does try to justify this confusion with some very poor logic early in
his book, but apart from that most of the information he gives is a good
start for anyone beginning an interest in the topic.
--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when replying)

Archived from groups: alt.comp.periphs.scanner (More info?)

In article <d4rgm9$gli@dispatch.concentric.net>, Ron
<eagle102308@yahoo.com> writes
>Considering that the best Photographic Print Media maxes out at 16 LPMM,

Citation please - you will see why this is necessary below.

>lines per millimeter extrapolate from there. 25.4 mm/inch X 16 = 406 LPInch.
>any thing over that is wasted. 300 or 600 will do just fine.
>
Unfortunately you have made a fairly common mistake of confusing lines
per millimetre and line pairs (or cycles) per millimetre. This is
common because the normal test pattern for continuous media is black
lines on a white background and often only the black lines are counted,
ignoring the equally important white lines between them. A black and
white line pair is a complete cycle that is repeated throughout the test
area, and you need a minimum of two pixels per cycle to reproduce that
digitally. Consequently you need TWICE as many pixels as the lines per
mm that is often quoted.

Black and white print media can easily require more than 400ppi to get
everything off the print and, even accounting for your 2:1 error, 800ppi
can struggle.

Wayne Fulton goes to great lengths on his site to explain that his "rule
of thumb" of 300ppi is only really relevant to average colour print
media and is not relevant to professionally produced colour or black and
white prints. In particular, it is definitely far too low for black and
white contact prints.
--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when replying)