[scribus] DPI/PPI

Peter Nermander peter at nermander.se
Wed Jun 8 06:25:58 UTC 2011

I think we need to sort things out here. (Note the note at the end of
the post:-)

DPI and PPI are physical measures. A digital image in the computer can
not have a DPI or PPI because it does not have any physical
dimensions. (How many inches wide is a JPEG file?)

Most image file formats have a field in the file for a DPI value, but
that is just a number. Changing the number does not affect anything
else. In the Scribus forum someone posted a link to a web page
explaining this well (with pictures showing the differences, or rather
no difference). http://www.rideau-info.com/photos/mythdpi.html

This number is only used to tell the "scale" of the picture. If you
print it at "100%" size it will be scaled to the size given by the
number of pixels and the DPI figure. But when doing graphical design
you normally never work that way unless it is important that the
printed result has a correct scale (for example a diagram where you
want to be able to measure something using a ruler).

A digital image in the computer has pixels, for example 2000x3000
pixels. If that image is printed at 2x3 inches it will be printed at
1000 PPI. If the same image gets printed at 20x30 inches it will be
printed at 100 PPI. Note here that I use the term PPI, because the
digital image has pixels, not dots.

Now, when the image is printed it will, in almost all cases, become rasterized.

The raster consists of small dots (there they are, the dots). Each dot
can have colour, or not. (I think one exception is colour laser
printers, they mix different amounts of toner for each dot so they can
print a continuous colour range in each dot.)

Since most (office/inkjet) printers only use four colours, Cyan,
Magenta, Yellow and blacK, using one dot per pixel would only give us
like 16 different colour tones for each pixel (in reality a bit more
because the printer can mix two or more colours in the same dot). That
is a bit less than the 16 million possible colours our 24-bit image
can contain... So to get a full colour tone range, each Pixel in the
image will be represented by a small group of Dots, for example 4x4
tiny Dots will be used to represent one Pixel.

This means that to get full resolution an image that is to be printed
at 300 PPI would need at least a 1200 DPI printer (300x4=1200).

So, you need not only know the number of pixels in the image and and
physical printed size, you also need to know how the printer
rasterizes each pixel.

This is also why the "resolution" stated by printer manufacturers (for
example a 2400 DPI inkjet printer) is useless to know, unless you also
know how many of those dots are needed to represent a pixel.

A 1200 DPI printer using a 4x4 dots (16 dots per pixel) raster prints
a higher pixel density (300 PPI) than a 2400 DPI printer using a 12x12
raster (200 PPI, but 144 dots per pixel), but the latter can print a
finer colour tone range.

For printing presses things work a bit different, since rasters are
angled and dots can have different sizes. But the essence is that PPI
is the "pixel density" and the DPI is the "raster resolution". And
none of the exists without a physical dimension.

Note, this is in some ways simplified, many printers today use
different methods to increase the pixel density using fewer dots for
each pixel, for example by using more than 4 inks, varying the size of
the ink drops etc.


More information about the scribus mailing list