why matlab calls (pixel per inch "PPI") a (dot per inch "DPI") ?

why matlab uses the DPI abbreviation for a something called the screen pixels per inch unit,

the DPI term indicates to a something related to printing terminology which is dot ink per inch unit,

is it a mistake or there is an explaination for this?

[Bjorn Gustavsson]

Just a notation quirk. What problems do you have/anticipate with this discrepancy?

[Osama Alkurdi]

i dont know, do they use dpi to indicate that the dot on your screen may be a bunch of pixels so they represent a light dot, because the documentation says that the dpi is determined by your system resolution not by physical pixels number on your screen and the physical real screen size in inches.

[Bjorn Gustavsson]

I'm happy enough to work predominantly on Linux-systems, so I get the resolution "determined by your system resolution"...

...I will not be forward enough to tell you what to do with regards to this, but I'm happy with my choise.

[Osama Alkurdi]

ok thats not helping me :(, and what is "system resolution" mean, i found the sentence misleading, do they mean you can specify your dpi at linux and matlab works with that value, and if so how you can change something fixed like dpi which depends on number on pixels which is constant and and screen dimensions which is also constant.

Answer 1

[Walter Roberson]

the DPI term indicates to a something related to printing terminology which is dot ink per inch unit,

Although it did certainly have that meaning, historically most electronic displays were CRT (Cathode Ray Tubes), and historically those used round-ish pixels created by steering an electron beam, so historically it was dots per inch for displays.

Technology for displays now varies quite a bit. Quite a lot of the technologies use dots of various forms. There are not a lot of display technologies around for which "dot" is technically wrong.

Dots Per Inch is mostly used to refer to physical limitations, whereas Pixels Per Inch is mostly used to refer to logical arrangements. For example my physical display is normally 2560 x 1440, but I can configure it for 1600 x 900 pixels per inch, with the graphics card automatically scaling . As that is a 1.6 ratio, the implication is that the hardware might have subpixels that are not user addressible. Hypothetically individual sub-pixels on the hardware might not have enough brightness or clarity, or there might be too much leak between adjacent wells to make the sub-pixels useful, but having the same level going to multiple sub-pixels in an area might overcome those problems, allowing a variety of pixel sizels (visible units) to be overlayed on hardware dots.

[Osama Alkurdi]

thanks for the historical info, i really appreciated.

do you mean that computers can change the pixels size by forming a bigger pixels from smaller ones and that new pixel will be treated as one unit, and like that we can get different resolution and different dpi, and matlab do that to graphical object like line plot for example so he can maintain 96 dpi no matter the monitor display dpi or resolution.

[Walter Roberson]

Modern displays get sent digital signals about pixel values. How displays translate that into hardware is a black box. (Though, I suppose, considering it is a display, it wouldn't exactly be black...) Certainly it is beyond MATLAB's control.

It is not possible to maintain 96 dpi no matter what the monitor display dpi, considering that some displays do not have 96 dpi.

Consider that you could be using a monitor that allows "pincushion" control or vertical or horizontal size controls at the monitor level (though these controls are getting rare for non-CRTs). Or you could be using a magnifying lens over your monitor. Or you could be using an EMF Shield over your monitor that had the effect of changing the optical size.

Therefore it is not possible for a computer to know what physical size an object will show up as.

The best you can do is to calibrate.

https://www.mathworks.com/help/matlab/creating_guis/dpi-aware-behavior-in-matlab.html

[Osama Alkurdi]

thats not helping me :(, i am so confused.

[Walter Roberson]

A) It is not a mistake.

B) The explanation is at https://www.mathworks.com/help/matlab/creating_guis/dpi-aware-behavior-in-matlab.html

C) Using "DPI" is technically correct in most hardware situations. It is not always true, but you are unlikely to be using a retro-computing Tektronix 4014 https://en.wikipedia.org/wiki/Tektronix_4010 (early 1980s) or one of the highly specialized displays they use for radar tracking. (At one point in early 1980s I was part of an airplane tracking software team, but I was never permitted near a real tracking system, only a mock-up.)

D) "DPI" usually refers more to hardware, and PPI (Pixels Per Inch) usually refers more to logical resolutions overlayed on hardware

E) The reasons for taking this approach is that even as resolutions increase, people need pretty much fixed-physical-width text, and trying to mix angular resolutions (readability) with plots and images (which can often do nicely with same number of denser pixels => smaller physical size) is messy. https://www.mathworks.com/matlabcentral/answers/753864-matlab-gives-wrong-graphical-display-info#comment_1349719