Rongnongno Loc: FL

These are important when creating and printing a digital image, regardless of folk's mantra: "A pixel has no dimension, you can ignore it".

SPI DPI and PPI relate to the size of Sensors, Pixels and Dots that fit within a set distance, in inches (I). When you see 3,000x2,000 px, it means that the sensor array creating the image had 3,000 sensors on one side and 2,000 sensors on the other. Divide this by the sensor array physical size, and you have what is called SPI, which is also later referred to as DPI and PPI.

SPI, PPI and DPI essentially refer to the same thing.

When working with a digital image, pixel 'size' is somewhat irrelevant, or so it appears, hence the mantra. In reality, you are dealing with a pixel array.

Now use your display to view a 400x400 pixel image on different screens with a different PPI resolution, and you will see that the visual size changes. The same goes when printing. The printing area changes depending on the pixel numbers AND the printer PPI.

The result? You have to plan your post-processing on the final media used. Using a printer, extrapolation or interpolation capabilities is a mistake if you want to print the best image possible. This time the DPI becomes important. The same goes if an image is simply displayed on a screen (PPI).

So...
SPI is important when creating an image. (sensor per inch)
PPI is important when displaying an image. (pixel per inch)
DPI is important when printing. (dot per inch)

When editing? To know what the final product is is important in order to exploit one's processing skills correctly. If you want to print an image with different sizes, you must send as many images as you need, one for each size. If not, you let someone decide for you what color is best, relinquishing your quality control. Images can become blurry, muddy, have some banding, pixelated, name it, all issues that exist when compressing or resizing w/o care.

Sure printers do a decent job to start with but decent does not apply if you want 'a perfect rendition of your work'.

You cannot control what another person uses as a display, so a big WHO CARES enters.

ELNikkor

Oh.

Longshadow Loc: Audubon, PA, United States

DPI is "dots (of ink) per inch" in a printer.
Won't a printer set at 300 DPI spit out 300 dots per inch no matter how many pixels are sent to it?
Doesn't the printer driver take the information it receives and convert it to whatever it needs to spit out the appropriate 300 dots per inch?

CHG_CANON Loc: the Windy City

This post should be deleted as misleading gibberish. There is not one single dot anywhere inside of any pixel-based images. Because if there were, you could very easily answer the basic question: how many pixels are in a dot? DPI is nothing more than a text-based "tag" added to pixel based images, for historical reasons, like the equally useless human appendix. It can have any whole-number value greater than zero. It impacts the image just the same as the file name, as is: none. It's just a text value, unrelated to any technical aspect of the image.

Longshadow Loc: Audubon, PA, United States

PPI and DPI are not directly related.

CHG_CANON Loc: the Windy City

You're right, one is a human-calculated value based on the specifics of one unique pixel-based image file and the measurement of the target print size. The other is a magic, non-zero, piece of text.

Human Calculated - there is no PPI value in an image file. Look up, down, left, right, back, front, anywhere within the file. PPI (pixels per inch) is a calculated value and relates simply to the print resolution. How many pixels on one side of the image divided by the print size expressed in inches, hence pixels per inch. PPI is relevant for printed images, irrelevant for image display. And, 'ppi' can be determined only when you know the intended print size, not before.

As a human-calculated value based on the specific pixel resolution of a specific image file and the target print size, PPI is not interchangeable with a random piece of text (DPI) placed as a tag into a file.

Longshadow Loc: Audubon, PA, United States

If a Printer Driver could talk: "Oh, there's a picture, let me see what I need to do to it to print it at X DPI.".

Rongnongno Loc: FL

Yes, and that is the issue. "Whatever it needs to spit out" (I really love your choice of words because it really exposes the issue at hand.)

If you do not prepare your image BEFORE printing, you let the printer driver/profile do it for you. Not a great idea if you care about the final print. That is one of my points.

Printers do a decent job, if you control the pixel array BEFORE sending the image to the printer, you also control what the printer needs to do.

The dots printed are where they are supposed to be w/o modification, assuming the printer color space matches the image color space.

As stated: One needs to know what the final product is in order to achieve optimal printing result. This prevents the device used to do "Whatever it needs to spit out".

There is no argument as to the so-called physical size of a pixel, as it is a mathematic formula. The size issue enters only when using a device to display the image. In effect, the device used to view/print used is controlling the pixel array, giving it a physical size. When scanning, this is the reverse, a sensor array has a physical size and so does the individual sensor.

When you look at the image this way, the number of pixels to use when post-processing is determined by the device PPI/DPI array.

This reality is obfuscated by the mantra 'a pixel has no size'.

burkphoto Loc: High Point, NC

The appendix analogy is pretty appropriate here.

'dpi,' decades ago, referred to the number of "cells" a scanner recorded. It still does. If you have a flatbed scanner, and set the driver for 300 dpi, it records 300 individual samples per inch, horizontally and vertically, which it then converts to pixels. If you set the driver to 200%, it records four pixels for every dot. If the driver is set for 100%, it saves one pixel per dot. But either way, if the driver is set for 300 dpi, the header value of the file remains the same! But the dimensions stored in the file DO change to reflect 200%. The total number of pixels is quadrupled for 200%.

The 'dpi' header in the metadata of the file is simply there to tell ancient page layout software how big to make the image when it is imported onto ("flowed" onto) a page. Aldus (later Adobe) Pagemaker used that header.

dbrugger25 Loc: Raleigh, NC

Many years ago, in printer terms, I had an HP color printer that produced images in 75 DPI. There were no sharp edges in images. Everything it printed looked a little blurry. It was a state-of-the-art priner in its day. Today's printers produce extremely sharp, crisp images when not viewed in high magninication.

In that old printer, I could see the pixels so I can attest that they were there. From a distance away, they became less visible.

In modern, quality printers, the pixels are still there but, because they are much smaller, and the interpolation of colors is better, they are nor visible to the unaided eye. They are still there.

Longshadow Loc: Audubon, PA, United States

And who will know, other than the photographer?????

selmslie Loc: Fernandina Beach, FL, USA

Yes, it is absolute meaningless nonsense that serves no useful purpose.

That is total nonsense!

How can you "prepare your image BEFORE printing" and avoid the "printer driver/profile"?

That driver is the only thing that can convert the information in a pixel (a trio of three integer numeric values for red, green and blue) into an array of dots that look like the intended color.

Longshadow Loc: Audubon, PA, United States

And the printer driver cannot be over-ridden. It simple reacts differently to a different image it has to process.

Rongnongno Loc: FL

The usual push back from folks who are just not paying attention and love GES.

Longshadow Loc: Audubon, PA, United States

Or they understand how it works and are not concerned.....