What Is a TIFF File? The Scanner Format Lawyers Still Get
By Schema LawPublished Updated
A TIFF file is an image file that stores a picture of a page as a grid of pixels, along with tags that record its size, resolution and compression. It holds no text. Scanners, vendors and opposing counsel still send TIFF files because the format is old, stable and built into scanning and production routines.
For a lawyer the format raises practical questions more than technical ones. Is this file one page or forty? Why can nobody search it? Why does it open on one computer and show an error on another? Each has an answer in how the format is put together.
What is a TIFF file?
A TIFF file is a container for raster images, meaning pictures made of rows of pixels. The Library of Congress describes TIFF, Revision 6.0 as "a tag-based file format for storing and interchanging raster images." File names end in .tif or .tiff.
Three facts from that Library of Congress description explain much of how the format behaves in an office:
- It is old and has not moved. The description states that "the most recent version of TIFF is 6.0, published in 1992." A file made by a scanner this year follows a specification written more than thirty years ago.
- It is a wrapper. The same page says TIFF "serves as a wrapper for different bitstream encodings for bit-mapped (raster) images" and that the encodings "may represent different compression schemes". Two files with the same extension can hold pixels packed in quite different ways.
- The name is not settled. The page notes that the acronym "is variously glossed as Tagged Image File Format and Tag Image File Format". Both mean the same file.
Why the format is still in your inbox
TIFF arrives from three directions. A scanner or copier may have it as a save option beside PDF. A fax server or document management system may store incoming pages in it. And a production in litigation may be specified in it, page by page.
The Library of Congress entry for TIFF with Group 4 compression, the black and white variety, records that it "has also been widely used for scanning business documents into document management systems." Whether a production should be TIFF at all is a separate decision, weighed in the comparison of TIFF and PDF for document production.
What a TIFF is not
A TIFF is not a document in the way a word processor file or a PDF is. It has no bookmarks, no form fields and no layer of characters. Whatever you can read on screen, you are reading from a picture.
Single-page and multi-page TIFF: one file per page or per document
A single-page TIFF holds one image, so a forty-page document is forty files. A multi-page TIFF holds many images in one file, so the same document is one file. Both are valid, both use the same extension, and the file name does not tell you which you have.
The structure behind this is simple. A TIFF file contains one or more directories, each describing one image. The TIFF 6.0 specification says: "There may be more than one IFD in a TIFF file. Each IFD defines a subfile." It gives as an example "the pages of a facsimile transmission." The Library of Congress description adds that "there is no limit to the number of subfiles a TIFF image file may contain."
| Single-page TIFF | Multi-page TIFF | |
|---|---|---|
| Files for a 40-page document | 40 | 1 |
| Where a document begins and ends | Not recorded in the images; a separate list or the folder structure must say | The file is the document |
| Page order | Depends on file names sorting correctly | Fixed inside the file |
| Replacing one page | Swap one file | Rebuild the file |
| Risk when opening | Pages viewed out of order, or one overlooked | Only the first page displayed |
The first-page trap
The last row deserves attention. The specification states that "a Baseline TIFF reader is not required to read any IFDs beyond the first one." A basic image viewer may therefore show page one of a multi-page TIFF and offer no way to reach the rest, with no message that anything is hidden.
Consider a fictional example. Elena Marchetti receives Fax_0412.tif from a records custodian and opens it in the picture viewer on her laptop. She sees a one-page cover sheet and files a note that the custodian sent only the cover. The file is 1.9 MB, which is large for one black and white page. Opened in a program that shows a page count, it has 37 pages.
Page through a medical bill saved as TIFF before you judge its length. The first image is often a cover, and the total plus the party name sit further in. Once you add text, search the amount and any statute or regulation printed along the footer of the later sheets.
Three checks avoid that mistake:
- Look at the file size. One black and white page is small. A file many times larger than its neighbours holds more than one page.
- Open the file in a program that displays a page count or thumbnails.
- Ask the sender how many pages they sent, and compare.
Sorting single-page files
With single-page files the hazard is order. Names such as 1.tif, 2.tif, 10.tif sort as 1, 10, 2 in a plain alphabetical listing. Names padded with zeros, such as 0001.tif, sort correctly. Check that your file listing runs in the order of the printed page numbers before you combine or convert anything.
TIFF tags: the resolution, page size and software a file records
Every image in a TIFF file carries tags, which are small labelled values describing that image. Tags record the pixel dimensions, the resolution, the compression used, and optionally the software, device and date that produced the image. Read together they tell you the physical page size and something about where the file came from.
The Library of Congress keeps a reference list of tags for TIFF and related specifications. The ones worth knowing for a scanned page:
| Tag | What it records |
|---|---|
| ImageWidth | Pixels across |
| ImageLength | Pixels down |
| XResolution, YResolution | Pixels per unit, horizontally and vertically |
| ResolutionUnit | The unit for those two: inch, centimeter, or none |
| Compression | A number identifying how the pixels are packed |
| Make, Model | The scanner's manufacturer and model |
| Software | The program that created the image |
| DateTime | When the image was created |
| PageNumber | The page's position in a multi-page document |
Working out the page size
Divide pixels by resolution. An image 2,550 pixels wide with an XResolution of 300 per inch is 8.5 inches wide. If its ImageLength is 4,200 at the same resolution, the page is 14 inches long: legal size, not letter. That calculation tells you whether the foot of a legal-size page was cropped to letter length before you ever see the paper.
Two cautions come straight from the TIFF 6.0 specification. It notes "that the X and Y resolutions may be unequal", so a page can look stretched in a viewer that assumes square pixels. And ResolutionUnit can be set to a value the specification defines as "No absolute unit of measurement", in which case the file does not record a physical size at all.
What the date and software tags are worth
The specification defines DateTime as "Date and time of image creation" in the form "YYYY:MM:DD HH:MM:SS", and Software as "Name and version number of the software package(s) used to create the image." PageNumber is "the page number of the page from which this image was scanned", and the specification adds that "the first page is numbered 0 (zero)."
Read those definitions narrowly:
- DateTime is when the image was made, by the clock of the device that made it. It is not the date of the document, and it is only as right as that clock.
- Software and DateTime are optional. Their absence proves nothing.
- Any program that later re-saves the file can rewrite them. A TIFF converted from another format carries the converter's values.
Treat tags as a lead to follow up with the person who made the scan, not as a finding. The same reasoning applies to the properties of a scanned PDF, which the post on scanned PDF metadata goes through field by field.
To see the tags, open the file's properties in your operating system and look at the details view, or use the document information panel in an image program. Viewers differ in how many tags they show.
Can a TIFF file hold searchable text the way a PDF does?
A TIFF file does not hold searchable text the way a PDF does. The format stores images. It has no equivalent of the invisible text layer a searchable PDF places behind each page, so the words recognized from a TIFF have to live somewhere else: a separate text file, a database, or a PDF made from the image.
The TIFF 6.0 specification does define text-valued tags, but none is a text layer. ImageDescription, for instance, is "a string that describes the subject of the image", with the example of a comment such as "1988 company picnic". That is a caption attached to a picture. It has no positions tying words to places on the page, and ordinary search tools do not look inside it.
So recognized text for a TIFF is kept beside it:
| Arrangement | How image and text are tied together | Weak point |
|---|---|---|
A .txt file with the same base name as the image | By name: ABC0001.tif and ABC0001.txt | Renaming either breaks the link |
| A document management system's index | By the system | The text does not leave with an exported image |
| A searchable PDF built from the images | The text sits inside the PDF | It is a new file; keep the TIFF as received |
When someone sends you TIFF images and you cannot search them, the first question is whether text files exist and were left out, not whether you need to run recognition again. In a document production the text files and the list that ties them to the images follow the production specification, which the comparison of TIFF and PDF productions covers under where the OCR text lives in each production format.
If you do need to recognize the pages yourself, check what your tool accepts. Schema Law takes PDF, PNG and JPEG uploads of up to 20 MB, and TIFF is not among them. The section on TIFF, HEIC and other image files that need converting first explains what to do with a format a tool does not accept. Once the pages are read, the text can be exported in any of the formats listed on the home page, including searchable PDF and plain text.
Why does a TIFF open in one viewer and fail in another?
A TIFF fails in one viewer and opens in another because the format allows many ways of packing pixels and a viewer is only required to understand a few. A file using a compression scheme or feature outside that basic set is valid, and a simple viewer may still be unable to draw it.
The TIFF 6.0 specification is explicit about the split. It defines a core called Baseline TIFF and a second part of extensions, and says: "TIFF Extensions are TIFF features that may not be supported by all TIFF readers." It also says "a Baseline TIFF 6.0 reader is not required to support any extensions."
What sits on each side of that line may surprise you:
| Feature | Where the specification puts it |
|---|---|
| No compression | Baseline |
| CCITT Group 3 one-dimensional (Modified Huffman) | Baseline |
| PackBits | Baseline |
| CCITT T.4 and T.6 encodings, known as Group 3 and Group 4 fax compression | Extension (Section 11) |
| LZW compression | Extension (Section 13) |
| Tiled images | Extension (Section 15) |
| JPEG compression | Extension (Section 22) |
| More than one image in a file | Allowed, but a baseline reader need not read past the first |
For the three baseline schemes the specification says "Baseline TIFF readers must handle all three compression schemes." Everything else is optional for a reader. Group 4, the scheme named in black and white scanning and production routines, is on the optional side, as is JPEG, which a scanner may use for color pages. On top of the published extensions, the specification reserves tag numbers "32768 or higher" for private use by individual organizations, so a file can also carry values that only its maker's software understands.
How each compression scheme treats the pixels, and which ones discard detail, is covered in the comparison of lossy and lossless compression in scanned files.
Diagnosing a TIFF that will not open
| What you see | What to suspect | What to do |
|---|---|---|
| An error naming an unsupported format or compression | An extension the viewer does not implement | Try a second program; ask the sender which compression was used |
| One page where you expected many | A viewer that reads only the first image | Open in a program that shows a page count |
| A page stretched tall or wide | Unequal X and Y resolution | Try a viewer that honors the resolution tags |
| A solid black or inverted page | The viewer misreading which value means white | Try a second program before assuming the scan is bad |
| The page opens but colors are wrong | A color scheme the viewer handles poorly | Ask for the page again as PDF |
| Nothing opens anywhere | A damaged or incomplete file | Compare the file size with the sender's; ask for it again |
When you ask a sender to resend, be specific. "Please send it as a PDF" gets you a file you can open. "Please confirm the compression and whether the file is multi-page" gets you an explanation. Do not convert and discard the file you were sent. Keep it as received and work on a copy.
TIFF, PDF, JPEG and PNG as containers for one scanned page
The same scanned page can be saved as TIFF, PDF, JPEG or PNG, and the pixels may be identical. The containers differ in how many pages one file can hold, whether text can travel inside the file, and whether the compression can discard detail.
| TIFF | JPEG | PNG | ||
|---|---|---|---|---|
| Pages per file | One or many | One or many | One | One |
| Text inside the file | No text layer | Yes, as a layer behind the image or as real text | No | No |
| Compression | Several schemes, some keeping every pixel and some not | Several schemes, some keeping every pixel and some not | Discards detail to save space | Keeps every pixel |
| Suits black and white text pages | Yes | Yes | Poorly; edges of letters blur at small file sizes | Yes |
| Suits color photographs | Yes, at large file sizes | Yes | Yes | Yes, at large file sizes |
| Records resolution | In tags | In the embedded image and the page size | Optional | Optional |
| Opens without special software | Depends on the features used | In any PDF reader and in web browsers | In any image viewer and browser | In any image viewer and browser |
| Accepted as a Schema Law upload | No | Yes | Yes | Yes |
Choosing among them
- You are scanning paper yourself and want to search it. Save as PDF. One file per document, and recognition can add text inside it.
- You have been handed TIFF files. Keep them as received. Make a PDF working copy if you need to read, search or upload.
- Someone asks you for TIFF. Ask which kind: single-page or multi-page, which compression, and where the text should go. Those three answers define the delivery.
- You have one page as a phone photo. Check which format the phone saved. A JPEG trades detail for size, so confirm the print is legible at full zoom before relying on it.
- You need a screenshot or a page with sharp line work kept exactly. PNG keeps every pixel.
Whatever the container, recognition quality depends on the pixels inside it: enough resolution, a straight page and clean contrast. Changing a file from one container to another does not add detail the scan never captured.