Why Images Come Out Rotated in a PDF
An image rotated wrong in a PDF is almost never a damaged file. Your camera saved the frame the way the sensor read it and attached a small tag saying which way up it goes — photo apps obey that tag, and the converter that made your PDF did not.
Choose the sheet, the margin and how the image sits on it — and see the print resolution you get.
🔒 Private by design: your images are converted in this browser tab and never uploaded.
The quickest fix, if this is one photo
If you just need this one PDF to look right, fix the photo rather than hunt for a better converter. On Windows, open the picture in Photos, rotate it until it is upright, and save. On a Mac, open it in Preview, press Command-R until it looks right, then Command-S. Both apps usually record the turn as an EXIF orientation tag rather than moving the pixels, so the photo looks right in anything that reads the tag — but a converter that ignores the tag can still hand you the same sideways page.
Two things worth knowing. Some apps rotate a JPEG without touching the compressed data and some quietly re-encode it, so keep the original if the picture matters. And if the file looks correct in every app except your PDF, nothing is wrong with the file. Your converter is ignoring a tag it should be reading, and it will do the same thing to the next hundred photos.
What EXIF orientation actually is
A camera sensor is a rectangle bolted into the body one way round. Turn the phone sideways for a portrait shot and the sensor still reads out its own landscape frame. Rather than rotate several megabytes of pixel data on a battery, the camera writes the frame as-is and records a tiny EXIF field called Orientation — tag 0x0112 — that says how far the picture has to be turned to look right.
Galleries, browsers and photo viewers read that field and turn the image for you, which is why the photo looks perfect everywhere until the moment you convert it. A PDF page has no equivalent. It draws the image into a rectangle exactly as the bytes describe, so the converter has to look the tag up itself and act on it. Plenty of converters never look. That is the whole story behind a photo turning up sideways, or a PDF image upside down, after a conversion that seemed to work fine.
The eight values, and why four of them cannot be fixed by rotating
The tag holds a number from 1 to 8. Four of them describe a plain rotation. The other four describe a rotation plus a left-to-right flip — these turn up on some front-camera selfies, and no amount of rotating will correct them, because a mirror image is not a rotated image.
| Value | What it means | Fixable by rotation alone |
|---|---|---|
| 1 | Already upright, nothing to do | No work needed |
| 2 | Flipped left to right | No — mirrored |
| 3 | Upside down, needs 180 degrees | Yes |
| 4 | Upside down and flipped | No — mirrored |
| 5 | Quarter turn clockwise and flipped | No — mirrored |
| 6 | Needs a quarter turn clockwise (the common one: phone held upright) | Yes |
| 7 | Three quarter turns clockwise and flipped | No — mirrored |
| 8 | Needs a quarter turn anticlockwise | Yes |
Value 6 is what a phone held upright records, which is why it is the one behind most sideways pages.
How to check a file's orientation on Windows and Mac
Worth doing before you blame the converter, and it takes about ten seconds.
Windows
- Switch File Explorer to Details view.
- Right-click any column heading, choose More, and tick Orientation. The value appears as a column against each file.
- "Normal" means the tag is 1 and rotation is not your problem. Anything else is the instruction your converter skipped — but note that Windows names the values the opposite way round to the table above, so the common upright-phone photo shows as "Rotate 270 degrees", not 90.
Mac
- Open the image in Preview.
- Choose Tools, then Show Inspector — or press Command-I.
- Open the More Info tab and look at the TIFF section. Orientation is listed there as the number from the table above.
If there is no Orientation row at all, the file has no tag. That is normal for PNG, for most screenshots, and for images that have already been edited and re-exported by something that baked the rotation into the pixels.
How to fix an image rotated wrong in a PDF, for good
- Convert with a tool that reads the tag. This fixes every affected photo in the batch at once, with no editing and no re-saving of originals.
- Check the thumbnail before you convert. This tool shows a thumbnail of each image before you convert, so you can spot a sideways one — though TIFF, and HEIC outside Safari, show a labelled tile rather than a picture. If one is on its side there, it will be on its side in the PDF.
- Use a per-image rotate control for the leftovers. Files with no tag — screenshots, scans, anything already edited — cannot be fixed automatically, because there is nothing to read. Those need one manual turn.
- Only re-save the original as a last resort. It works, but it rewrites a file you may want to keep exactly as the camera made it.
What the converter on this page does with the tag
It reads EXIF tag 0x0112 out of every JPEG before anything else happens, then applies the rotation as a page transform in the PDF rather than by redrawing the picture. The original JPEG data is still embedded untouched, so the photo comes out upright and loses nothing on the way. The rotate button on each thumbnail works the same way — it composes with whatever EXIF already asked for, and turning a JPEG or PNG by hand does not re-compress it.
The mirrored values are the exception. A page rotation cannot express a flip, so orientation 2, 4, 5 and 7 are decoded and re-encoded once to come out the right way round, and the tool tells you when that has happened rather than leaving you to spot it. TIFF gets the same treatment through its own orientation tag, number 274 — useful if you are converting TIFF without Adobe. Formats that have to be decoded anyway, such as HEIC from an iPhone, WebP and AVIF, are decoded with the browser told to honour the file's own orientation, so they land upright too — though unlike a JPEG they are re-encoded once on the way in.
PNG only gained an optional Exif block late in its life, and screenshot tools do not write one, so the converter treats every PNG as upright. That is why screenshots almost never come out sideways, and why the rotate button is the fix when one does.
The other kind of sideways: a portrait photo on a landscape page
Sometimes the picture itself is upright and it is the page that is the wrong way round — a tall photo squeezed into a wide A4 sheet with white bars either side. That is not EXIF. That is the page setup, and it has its own control: set orientation to match the image and the page turns to suit each photo instead of forcing every one into the same shape.
If the page shape is right but the picture looks squashed, that is the fit setting rather than rotation. Fit inside the page keeps the proportions and leaves margins, fill the page crops the edges, and stretch distorts — which is the setting people usually turn on by accident. Colours shifting at the same time is a separate matter entirely, covered in why image colours look different in a PDF.
If you are doing this on an iPhone
The iOS app converts photos to PDF entirely on the device, with no internet connection needed, and keeps a history of everything you have made. It reorders pages with arrow buttons rather than dragging. One thing to know before you use it for rotation problems: the app resizes photos to 1200 px wide and re-encodes them at JPEG quality 0.8 to keep PDFs small. If the picture needs to keep its original quality, open this page in Safari and convert here instead, where JPEG and PNG data is embedded untouched.
