An 8.5x11 piece of paper should have 77% less pixels/dots on the 8.5" side. However, most consumer printers seem to print at "Dots per inch" (DPI) of 1200x600, 2400x600, 4800x1200, or even 4800x600 for the entire page.

This tells me that the dots printers print are often not square. Displays have a similar situation with pixel aspect ratios like 16:9.

However, 4800x600 seems greatly disproportionate. It's like a 8:1 aspect ratio. How do the images look okay with so few vertical dots and why do printers have so few vertical dots to begin with? Why don't printers print square pixels by default?

  • \$\begingroup\$ I’m voting to close this question because it has nothing to do with photography. \$\endgroup\$
    – Tetsujin
    May 6 at 17:09
  • \$\begingroup\$ You are confusing dots and pixels. An inkjet printer uses several dots to render a pixel. Plus the head move across the paper (timing belt) is likely more accurate than the paper move under the head (friction). \$\endgroup\$
    – xenoid
    May 6 at 22:04
  • 1
    \$\begingroup\$ Have you ever tried to spray a square pattern out of a garden hose nozzle? \$\endgroup\$
    – Michael C
    May 6 at 23:12

3 Answers 3


You are confusing a lot of things. Let me try to explain a bit.

  1. PPI and DPI are not the same units. A Pixel is an information unit. The values of color on one location of an image. And a Dot is... a dot. A tiny droplet of ink.

  2. Dots are not squares, rectangles, circles, or ellipses. They are droplets of random shapes dictated mainly by the absorption of the fibers of the paper.

  3. One reason for the printers having different DPI on the horizontal axis, vs the vertical axis is that they are controlled by two different mechanisms. The X-axis is controlled by a moving printer head on a lane, or a laser beam shoot by a rotating mirror. The Y axis is controlled by rollers or a rotating drum. If you wanted the same DPI on both axes you would get the smaller number, because one mechanism is rougher than the other (600x600 print instead of 1200x600 for example). But you want smaller dots by overall area so if you can make them smaller even in only one direction, do it. They are so small that you do not see the shape.

for the entire page

No, they are Per Inch. That is what the second P and the I mean. 1200 dots each inch.

Why don't printers print square pixels by default?

A printer will print a pixel on whatever aspect ratio the image has. It has nothing to do with how many DPI they print.

Imagine these images are 5x5 pixels. You can squish or squash the image, so the pixels are no longer square. The printer will print them accordingly.

enter image description here

Here is a simulation of different DPI Two sets of checkerboard of the same size printed in different dot sizes. Simply means, smaller dots or bigger dots. (In this simulation I have little squares, but they will be random shapes on a real print)

enter image description here

And here is a simulation of two different PPI using the same image.

The first image has 5 pixels on one inch. The second one has 2.5 pixels on one inch.

enter image description here

Here is another simulation

In order to clearly see it, you need to open it on a new tab at 100% so it is not scaled inside this post on your browser. The image is optimized to be viewed on a Full HD monitor.

  1. The first image has black and white dots of 1 pixel, simulating a print of "X DPI" Let me put an example number. 100 DPI.

  2. The second image has twice bigger dots. It has X/2 DPI or 50 DPI. Each square is 2x2 pixels.

  3. The third image has vertical rectangles as dots. It is simulating a resolution of, let's say 100x50 DPI. Each dot is 1x2 pixels.

enter image description here

If you move away from the screen, you will not really notice the rectangular dots, but you will appreciate the improvement in detail compared to image 2.



Same reason digitized old TV formats do not have square pixels, and scanners tend not to have square pixels: the majority of printers (including laser printers) work with a scanning process that inherently has a discrete number of vertical lines at fixed distance (laser scan distance, number of ink nozzles, density of scanning sensors) which are semi-continously served by moving across those lines and placing/reading dots to considerable degree based on timing. That doesn't mean that the actually placed/sensed dots/pixels/sensed entities have different dimension horizontally and vertically, just that the positioning horizontally can be done more fine-grained as a rule.

To what degree a given device utilized this finer horizontal grain for doing its own font rendering (with stuff like "resolution enhancement technology") or offering the resolution nominally to a user-programmed graphics renderer is manufacturer choice. That the device accepts information at that resolution does not mean that it can produce line patterns at this resolution: just that it will likely utilize some of the information for placing its darkening agents and thus avoid "staircasing" of shallow curves to some degree.

  • 1
    \$\begingroup\$ This is correct, but I would add (or rather emphasize) for clarity that this dense/precise horizontal dot positioning often results in overlapping physical dots (for printers). Such sub-pixel positioning doesn't mean that the dots are correspondingly small. \$\endgroup\$
    – Zeus
    May 8 at 1:07

Because inevitably ink and toner are not square, printed squares will have rounding errors.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.