Archived from groups: comp.periphs.printers (
More info?)
In article <7v4Zd.9934$C47.9877@newssvr14.news.prodigy.com>, measekite
<measekite@yahoo.com> writes
>Anyone (who uses Photoshop) and has a good understanding of the
>difference between pixels and dots per inch. I know that printers
>print in dots per inch and monitors display in pixels per inch.
>
And images are made from pixels.
>If you have a photo that has a resolution of 400 pixels per inch (using
>1024x768) how can you tell the dots per inch that the printer will
>print at?
>
It will print at whatever dots per inch you tell it to print at - it has
nothing at all to do with the pixels.
>How can you tell and equate the levels of quality to dots per inch.
The more dots per inch the better. The printer can only print a dot of
ink or not, and it only has ink in a few colours. So, when it tries to
reproduce a particular colour of, say, a pixel in an image it does this
by varying the number of dots it prints of each of the colours that it
has. This process is called dithering. Some inkjet printers might only
have 4 colours: cyan, magenta, yellow and black, others have more, such
as light cyan and light magenta. Printers with more ink colours require
less dithering, which means that they can either produce finer colour
variations or they require less dots per inch to achieve the same colour
variation.
To make this even more complicated, the dither pattern does not need to
occur entirely within an individual pixel. In other words, it doesn't
matter if the colour of each pixel is accurate - as long as the average
colour is. So the printer uses a complicated algorithm to print each
pixel on the image, measuring the actual colour that it creates from the
dots it prints and then calculating the difference between that and the
colour that the pixel should be. This colour error is then shared
between the adjacent pixels that have not been printed yet, so that they
compensate for the error. The process of error compensation continues
to the edge of the page.
The advantage of this is that for large areas with a specific colour or
a gradually changing colour, the actual colour is determined by the dots
of ink that are placed across the entire area, not just in the
individual pixel. So the colour can be controlled very precisely.
However, when there is a sudden change of colour between pixels, such as
on an edge, the same algorithm can reproduce that edge with pixel
accuracy.
Using some of these advance algorithms you can readily print an image
with a few colours of ink where only one or two dots of ink are placed
per pixel and still get both the fine colour tonality and the pixel
resolution in your image.
> My printer can print a resolution of over 1200 dots per inch. Taking
>my 400 pixel photo, what will it really print out at and how do you
>calculate it..
>
You don't - the printer driver does it all for you, completely
transparently. You just tell it what dpi you want it to place the ink
dots at, based on the printer's capabilities, and the printer driver
decides where the dots should be placed, how often and what colours. The
dpi you choose is determined by the media you are using and the quality
of the print you want - not the ppi of the image. With high quality
paper you should always use the highest dot resolution that the printer
offers. With poorer paper, the ink bleeds into the paper and spreads so
there is little advantage to using the finest dot resolution of the
printer and a lower resolution is likely to print faster - but it
probably won't save any ink, unless you printer offers a draft mode,
where the image can be lighter than the final version.
So, forget about the printer dpi other than matching it to your paper
stock. What you really need to worry about is your image ppi. To all
intents and purposes, you can't see a significant benefit by printing at
more than 300ppi unless you want to view the image under magnification
(and there are some cases where you might want to do that - such as a
sheet of "contact" prints). However if you get below about 200ppi then
the individual pixels in the image become quite visible, and by 100ppi
are quite objectionable even from a reasonable distance. So your
400pixel image would look as good as a photo booth print if you print it
no more than 1.5inches wide/high, and will progressively become more
visibly pixelated as you increase the size. Of course, if you view it
from a sufficient distance the pixelation will be less visible, but you
will see it close up with sizes of 3inch or more on a decent photo
quality printer. (This, of course, assumes you really mean a 400pixel
wide/high image, and not a 400ppi image as you referred to earlier - a
1024x768 pixel image can easily be printed at 5x3.5inches without too
much visible pixelation.)
Your printer may offer a special mode of printing which reduces the
visibility of the pixels by interpolating between them, but you will
have to enable this in the driver option page. On Epson printers it is
usually called "DCC" or "digital camera correction" or something
similar. If the printer does not have this option then you can upsample
the original image in Photoshop or other software to get sufficient
pixels to make them indiscernable on the page - aim for 200ppi as a
minimum with a photo quality printer. This is exactly the same as the
printer is doing in its special, low resolution image, mode but it is a
little slower because you are sending the printer more data. The upside
is that most software packages offer several upsampling interpolation
algorithms, some of which are likely to be better quality than the
printer's simple version.
A final issue is the printer's native image resolution. This is not the
ink dots per inch, but the pixels per inch that the printer driver uses
when the image is rasterised. For all Epson desktop printers this is
currently 720ppi, but other manufacturers vary - and it is not a number
that they generally publish, you either get it from someone who knows or
work it out using devious test images, like I did! ;-)
The optimum image resolution to sent to the printer should *never*
exceed this native resolution (because the driver will just ignore the
extra pixels and you have no control over how it does that) and should
be an integer division of it, so that each pixel in the image is
resampled to the same number of pixels in the printer driver
rasterisation. So, while the optimum for HP printers is often quoted as
being 300ppi, the optimum for Epson is 360ppi or, the next one down,
240ppi. In practice, the degradation caused by being off optimum is
negligible for most photographic images, but for certain images with a
lot of repetitive fine detail, for example, a picket fence or bricks in
a wall, it can be quite noticeable.
Hope that helps.
--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a ah heck when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when replying)
Facebook
Twitter
Instagram