DPI vs. pixels: why changing the number does not make art sharper

When I started selling Spanish classroom printables, I thought DPI was a quality switch. I would export a JPEG, open the file info, and change 72 DPI to 300 DPI. The picture on my screen looked the same, which should have been my clue. The print size label changed. The art did not get any new detail.

DPI means dots per inch. In a digital file it is a label that says how to spread the pixels across paper. It does not count new pixels, and it does not draw cleaner letters. Pixels are the actual grid of color. If the grid is small, the picture is small, no matter what number you type next to DPI.

The same picture, two labels

Take a file that is 1000 pixels wide and 1500 pixels tall. That grid does not change when you edit the DPI label.

If the label says 72 dots per inch, a program may describe the print as about 13.89 inches by 20.83 inches. That is 1000 divided by 72, and 1500 divided by 72. The math is only a label. The file still has one thousand dots across.

If you change the label to 300 dots per inch and save again, the same division gives about 3.33 inches by 5 inches. You did not shrink the art by painting it again. You told the next program to pack those same dots into fewer inches. Up close, each inch of paper now gets more of the dots you already had. The dots themselves were not improved.

This is why a buyer can say, "I set it to 300 DPI and it is still blurry." They changed the instruction. They did not add the missing picture.

What you need for a real size

If you want an 8 by 10 inch print at 300 dots in each inch, the grid has to be 8 times 300 by 10 times 300. That is 2400 by 3000 pixels. A 1000 by 1500 file cannot become that grid by a settings change. Something has to create the extra pixels. A plain resize copies and blends the old ones. An enlargement model guesses new ones. Neither one is the same as having drawn the poster at full size in the first place.

I use a short test on every new file. I ignore the DPI field for a moment. I read only the pixel width and height. Then I divide each number by 300. Those two answers are the inches I can print at the sharp target. If I like a bigger inch size, I am accepting fewer than 300 dots in each inch, and I say so to myself before I write the listing.

Here is the same test on a few sizes I actually sell:

• 8×10 inches needs 2400 × 3000 pixels at 300 per inch.
• 11×14 inches needs 3300 × 4200 pixels.
• 16×20 inches needs 4800 × 6000 pixels.
• 18×24 inches needs 5400 × 7200 pixels.
• 24×36 inches needs 7200 × 10800 pixels.

If my file matches the first line and not the last, I do not describe it as a sharp 24×36. The DPI menu will happily say 300 either way. The pixel count will not.

Where the label still matters

I am not saying you should delete the DPI setting. Printers and design apps use it to pick a starting size when they open the file. A JPEG can store that number. A PNG can store a similar density. Setting it to 300 is a courtesy when the pixel count really supports the inch size you claim. Setting it to 300 when the pixel count does not is how people talk past each other.

I also watch for apps that resize and relabel in one click. "Export for print at 300 DPI" sometimes means "scale this up until the math hits 300." That click manufactures pixels by stretching. You should see the new pixel dimensions after the export and decide if the letters survived. On my vocabulary posters, the little words under the pictures fail before the big title does.

A classroom example

Suppose I have a poster of classroom objects. The working file in my design app is large, but I once exported a "small preview" to send to a friend. That preview is 1200 by 1800 pixels. If I only change DPI from 72 to 300, the preview is still 1200 by 1800. Divided by 300, that is a 4 by 6 inch print at the sharp target. It is not an 18×24.

If I instead scale the preview up to 5400 × 7200 so the inches say 18×24 at 300, I have stretched it. The poster may still be usable if the shapes are simple and I check it. The Spanish labels on the objects may not be. I would rebuild those labels, or go back to the large working file and export the real pixels.

What I write in the listing

I tell buyers the pixel size, not only the inch size. "18×24 inches, 5400 × 7200 pixels" is a sentence a careful buyer can check. "High resolution 300 DPI" is not, because that phrase gets pasted onto files that were never that large.

When someone asks me to remake a poster bigger, I ask for the original file if they have one. If they only have a small export, I tell them the enlargement can help the shapes and still fail on the words. That expectation is kinder than a surprise after they pay a print shop.

The free check on this site does the division for you from the pixels in the file. It does not believe a DPI label that the pixels cannot support. You can use it before you decide whether a picture is ready for a frame size, and before you pay for a full-size file.