In a TIFF, the metadata is the file
Every other format bolts metadata on: a JPEG puts a TIFF block inside an APP1 segment, a HEIC stores one as an item, a PNG invented chunks. TIFF has no segment because it needs none — the file opens with an 8-byte header pointing at an Image File Directory, and that directory is simultaneously the tag list and the index to the pixels. When you read EXIF anywhere else, you are reading a TIFF fragment that was copied into a different container.
Byte order is declared, not assumed
A TIFF starts with either II or MM — Intel or Motorola, little-endian or big-endian — followed by the number 42 written in that order as a self-check. Everything after it, every offset and every integer, is read that way. This is why the same tag can be stored in two different byte layouts and why a truncated TIFF fails so completely: lose the header and there is no way to interpret a single offset in the file.
One file, several directories
IFDs chain: each one ends with the offset of the next, or with zero. A multi-page scan uses that chain for its pages, and a thumbnail commonly sits in a second directory. EXIF and GPS are sub-directories hung off the first one by pointer tags. So the tag table below may be showing you the first page of several — and a raw DNG, which is TIFF with extra tags, has the same shape.
Where a TIFF keeps things
- II / MM
- Byte order, in the first two bytes. Everything else is read through it.
- IFD0
- The main directory: image tags, make, model, orientation, resolution.
- IFD1
- A second directory, conventionally the thumbnail or the next page.
- ExifIFD
- Exposure, lens and capture time. A sub-directory, reached by a pointer tag.
- GPSIFD
- Latitude, longitude, altitude. Another sub-directory.
- StripOffsets
- Where the pixels are. A tag like any other — which is the point.
Questions about TIFF
Why can I read the tags when my browser will not show the image?Open
Because reading tags means walking directories and offsets, which is arithmetic, while displaying the image needs a decoder for whatever compression the file uses. No mainstream browser ships a TIFF decoder. The preview will say it is unavailable; the tag table will be complete.
Is a DNG the same as a TIFF?Open
Structurally, near enough. DNG is built on TIFF/EP, so it uses the same header, the same directories and the same 12-byte tag entries, with additional tags for raw sensor data. What you learn here about IFDs applies to it.
My TIFF is 40 pages. Which one am I looking at?Open
The first. The directories are a linked list and the workspace reads the one the header points at, which is page one. Later pages carry their own tags and are not shown.
How is a TIFF cleaned?Open
Through Re-encode Cleanup, which exports a JPEG. Since a TIFF's tags and its pixel index live in the same structure, there is no way to drop the tags and keep the file a TIFF in a browser — and no browser can write one anyway.
What it can do
- Remove EXIF data
- Drop a photo and download a copy with its embedded metadata removed. JPEG loses its EXIF segment losslessly; HEIC, PNG, WebP and the rest are re-encoded. Nothing is uploaded.
- Remove GPS location
- See the coordinates a photo is carrying, then download a copy without them. JPEG loses the location tags and keeps everything else. Nothing is uploaded.
- Edit JPEG EXIF
- Change the title, author, description, copyright, software and capture time on a JPEG, then download the edited copy. Runs locally; your original file is never modified.
- XMP, IPTC and ICC viewer
- Read every metadata group a file carries — EXIF, XMP, IPTC, ICC, maker notes — grouped and exportable as JSON. For asset QA, forensics and debugging a pipeline. Nothing is uploaded.
- Open several photos
- Open up to fifty photos, see which of them carry a location, and take the whole set out as a spreadsheet. Nothing is uploaded, and the originals are never written to.