Reading the field is easy, trusting it is the hard part

This one has a short answer and the short answer is genuinely short. Right-click on Windows, the Inspector on a Mac, the image size dialog in an editor. Every page ranking for this gives you those three and stops, and for a lot of readers that is all they needed.

The problem is what happens next. You read the field, it says a number, and you take the number as a fact about the picture. It is not always a fact. The image formats most photographs arrive in are allowed to carry no physical density at all, and when a reader finds nothing there it shows you something anyway, because a blank box looks like a bug.

So this page does both halves. Where to look, which takes a paragraph, and then how to tell whether what you are looking at was recorded in the file or supplied by the program you opened it with. If what you actually want is to change the value rather than read it, the sibling guide on raising DPI is the one to read and this page will hand you over at the end.

Where the field lives, on each thing you might be using

All of these read the same stored field. None of them inspects the picture, so they agree with each other whenever the field is present and disagree in an interesting way when it is not.

Windows: right-click the file, Properties, then the Details tab. It is listed as Horizontal resolution and Vertical resolution, which is a confusing name for the same value you are after, and the two are nearly always identical. You can add resolution as a column in File Explorer if you need to scan a folder of files rather than one.
macOS: open the image in Preview, then Tools and Show Inspector, and read the DPI line under the general information tab. Get Info in Finder shows dimensions but not density, which catches people out.
Photoshop: Image, then Image Size. The resolution field is the value, and the units dropdown next to it matters, since pixels per centimetre and pixels per inch are different numbers for the same file.
InDesign: the Links panel, where you get two values rather than one. Actual PPI is the density stored in the file. Effective PPI is the density the placed copy will really print at after you scaled it on the page, and effective is the one that decides whether the print is sharp.
An online checker: fine for a quick read, and worth preferring one that reports an absent field as absent instead of filling it in. That distinction is the subject of the next section and it is the only reason to care which tool you use.

The file is allowed to carry no density at all

This is the part the other results leave out, and it is not an edge case. It is the normal state of a great many screenshots, exports and files that have been through a web pipeline.

In a JPEG, the density lives in a short header alongside a units byte that says how to read it. The units byte has three states: dots per inch, dots per centimetre, or no units at all. In that third state the two density numbers are not a density, they describe the pixel aspect ratio, and a file saying one by one in that mode is telling you its pixels are square rather than that it is one dot per inch.

In a PNG it is simpler and more absolute. The chunk that holds physical dimensions is optional, so a PNG can be written without it, and the specification says that when it is missing the pixels are assumed square and their physical size is unspecified. There is no default hiding anywhere in the format for a reader to fall back on.

Which leaves every program that opens such a file with a gap to fill and no correct value to fill it with. So each one substitutes its own convention, which is why two readers can look at one file and report different densities without either of them being broken.

A JPEG whose units byte says no units has no density, whatever the two numbers next to it say.
A PNG with no physical-dimensions chunk has no density, and the specification declines to supply one.
What a reader shows for either of those is the reader’s own convention, not a measurement of your file.
Image tooling has had real bugs in exactly this corner, including a case where an aspect-ratio pair was read as a density, so a surprising value is worth a second reader before you act on it.

How to tell whether the number came from the file

There is no indicator for this in any of the dialogues above, which is the whole difficulty. What you can do instead is look for the signature of a substituted value, and it is usually obvious once you know to look.

Open the same file in two different readers. If they agree, the field is almost certainly present and populated, because two programs do not invent the same wrong answer. If they disagree, at least one of them is filling a gap, and the honest conclusion is that the file does not say.

The second check is the history of the file. A camera or a phone writes a density because its software has an opinion about print size. A screenshot, a web export, a file saved by a browser or a thumbnail out of an email frequently does not. If a file from the second group reports a tidy round screen-era figure, that is the shape of a default rather than of a measurement.

The third is the most reliable and takes no tools. Divide the pixel width by the width in inches you intend to print at, and that is the density your print will actually get. It cannot disagree with itself, it does not consult the stored field, and it is what a print shop checker is doing when it rejects a file whose header claims to be fine.

What the number is for, and what it never decides

The stored density is a default placement instruction. Its job is to tell a printer or a page layout how large to put the picture when nothing else has said, and anything downstream that specifies a size overrides it completely.

So the field cannot tell you whether an image is good, sharp or usable, and reading it as a quality score is the mistake that brings most people to this search. Two files with identical pixels can carry different values in that field, and the one claiming the higher number is not the better picture. It is the same picture with a different note attached.

What does decide a print is the pixel count against the physical size, which is the division in the previous section. The guide on raising DPI works through that arithmetic properly, including why re-tagging a file leaves every pixel byte for byte identical.

What to do with the answer you now have

If you read the field because a form or a print shop demanded a specific value and you wanted to know where you stood, do the division before anything else. Files from a recent phone or camera usually clear a normal photographic print size comfortably, in which case there is nothing wrong and the only outstanding job is satisfying a checker that reads the header.

If the division comes back short, no amount of looking at the field will help, because the shortfall is in the pixels. Your options are to find a better original, which is nearly always the better answer when one exists, or to add pixels with an upscaler, which proposes detail rather than recovering it.

And if the field was simply absent, that is information rather than a fault. Nothing is wrong with the file. It was written by something that had no opinion about print size, and the picture is exactly as detailed as it was before you went looking.

Where Unblur Photo fits here

Unblur Photo does not read or write that field, and it is the wrong tool for a portal rejecting your header. Setting a density is a metadata edit that any image editor does in one dialogue, and it does not touch the picture.

What it is for is the case the division exposes: a file that genuinely does not carry enough detail for what you want to do with it. It enlarges and sharpens, and the detail it adds is inferred from everything it has seen rather than recovered from your photograph, which matters a great deal for a face and very little for a wall or a landscape.

The guides on improving photo quality and on unpixelating an image cover where that inference holds up and where it quietly invents. Read one of them before you commit to an enlargement that has to establish a fact rather than fill a frame.

More guides

Questions

How do I check the DPI of an image on Windows?

Right-click the file, choose Properties, and open the Details tab. The value is listed as Horizontal resolution and Vertical resolution, which is the same thing under an awkward name, and the two are normally identical. If you need to check a whole folder, add resolution as a column in File Explorer rather than opening each file.

How do I check the DPI of an image on a Mac?

Open it in Preview, then choose Tools and Show Inspector, and the DPI appears under the general information tab. Get Info in Finder is the common wrong turn here: it reports pixel dimensions but not density, so people conclude the information is unavailable when it is one menu away.

Why do two programs show different DPI for the same image?

Because the file probably does not state one. JPEG can record its density numbers as a pixel aspect ratio rather than a physical density, and PNG can omit physical dimensions entirely, in which case the specification leaves the size unspecified. Each program then shows its own default, so the disagreement is the clearest signal you will get that the field is empty.

My image shows no DPI at all. Is something wrong with it?

No. Screenshots, web exports and files saved out of a browser are routinely written with no density, because whatever produced them had no view about print size. The picture is unaffected. Work out what you need by dividing the pixel width by the print width in inches, which gives you the density the print will really get.

Does checking DPI tell me if an image is high quality?

No, and this is the most expensive misreading of that field. It is a default placement instruction, overridden by any size specified downstream, and two files with identical pixels can carry different values in it. Quality comes from how many pixels you have and how much real detail they hold, neither of which the field reports.

What should the DPI of an image be?

The field has no correct value on its own, because density only means something once a physical size is chosen. Ask instead whether your pixel count covers the print you want: divide the pixel width by the intended width in inches and compare the answer with what you were asked for. If a form insists on a specific figure, set it, and understand that you have edited a label rather than the picture.

Sources

Restore a photo with Unblur Photo