How to Scan Legal Size Documents Without Cut-Off Text

To scan legal size documents without losing text, set the paper size to legal (8.5 by 14 inches) before you scan, turn off any fit-to-page scaling, and feed the stack through the document feeder if the glass is only letter length. Then confirm that the last line of every page is in the file.

How do you scan a legal size page without losing the bottom of it?

Tell the scanner the page is 14 inches long, scan at full size, and check the foot of the result. A legal sheet is the same width as letter and three inches longer, so a scanner left on its letter setting captures the top eleven inches and stops.

The procedure

1. Measure a page. Legal size is 8.5 by 14 inches. An old file can also hold sheets of other lengths, such as 8.5 by 13 inches, and a ruler settles which you have.
2. Choose the feeder or the glass. Look at the scale printed along the edge of the glass. If it ends near 11.7 inches, a legal page cannot lie on it whole, and the feeder is the route. If the glass is marked to 14 inches, either works.
3. Set the original size to Legal. On a copier panel this is under a setting named something like Original Size or Scan Size. In desktop scanning software it is the paper size or document size list. Do not leave it on Letter, and treat Auto with suspicion until you have tested it.
4. Set scaling to 100%. Switch off Fit to Page, Reduce to Letter and any output size that differs from the original size.
5. Set resolution and color mode as you would for any page you plan to convert. The guide to scanner settings that make OCR work has the full list.
6. Adjust the feeder guides to the paper and extend the output tray, so long sheets do not curl back into the exit.
7. Scan one page and open it. Check the page dimensions in your PDF viewer's document properties. You are looking for 8.5 by 14 inches. Then look at the bottom edge: the last line, the footer and the page number should all be there with white space under them.
8. Scan the rest, then count pages and check the foot of every tenth page.

Where legal size turns up

For a paper brief in a federal court of appeals, the size is fixed by rule. Federal Rule of Appellate Procedure 32(a)(4) says: "The brief must be on 8 1/2 by 11 inch paper." Other courts set their own paper rules, and a file is more than its court papers. Look for the longer sheet among deeds and mortgages, closing documents, older pleadings, wills, leases, insurance forms and anything a client kept from years back. One legal page inside a letter stack is enough to lose a signature block.

Those longer sheets are often a mortgage or a homestead form from a closing. The signature, the notary, and a joint tenancy choice sit in the bottom three inches, which is the strip a letter-size setting cuts away. Open that edge before you send the file on.

After the scan

A page that was captured whole converts like any other. Schema Law takes a PDF, PNG or JPEG of up to 20 MB and reads the pages into text and tables. The steps from upload to export are on the home page. What no converter can do is read three inches that the scanner never recorded.

Why does a legal size page come out cropped or shrunk?

A legal page is cropped when the scanner records a letter-length area and ignores the rest. It is shrunk when the device captures the whole sheet and then scales the picture down to fit a letter-size output. Both come from a letter default somewhere between the paper and the saved file.

Letter defaults

A scanner, a copier profile and a scanning program each store a default paper size. In an office that scans letter paper all day, all three may sit on Letter. Changing one and not the others gives odd results, such as a full-length capture saved onto a letter-size PDF page. Check the size at each step you control: the device panel, the scan profile and the save or export dialog.

Auto size

Automatic size detection measures the sheet with sensors in the feeder or on the glass. It can misjudge a sheet that is folded, curled, dark at the edges or fed crooked. It also may choose from a short list of sizes that does not include yours. Auto is convenient for a mixed stack and should be tested on your own machine before you rely on it. Scan a letter page and a legal page together and look at the dimensions of each in the result.

Later steps that crop

A good scan can still lose its foot afterward. Printing a legal PDF to a letter-size "print to PDF" driver, faxing it, pasting pages into a letter-size document, or running a file through a tool that normalizes page size can each crop or shrink. If a file reaches you already damaged, the question is which step did it, and whether someone upstream holds the full-length version.

How much is lost

Three inches of a fourteen-inch page is a little over a fifth of the sheet. At six lines to the inch, that is about eighteen lines of single-spaced text. On a form it is the part of the page where signatures, dates, notary blocks and footers sit.

Does shrinking a legal size page to letter size hurt OCR?

It can. Fitting 14 inches into 11 scales the page to about 79% of its size, so every letter is drawn with fewer pixels. Body text in an ordinary size survives. Footnotes, form labels and fine print on a legal-size form may drop below what recognition can read reliably.

The arithmetic

The scale factor is 11 divided by 14, or 0.786. Type shrinks by the same factor, and so does the resolution measured against the original page.

To keep 300 dpi worth of detail on a page that will be reduced, the scan would need about 382 dpi, which means choosing 400 on a scanner that offers fixed steps. The post on the best DPI for OCR explains how many dots a letter needs and why small print is the first to fail.

Reduced on paper versus reduced in software

The two cases are not equal.

• A photocopy reduced onto letter paper has lost the detail on paper. Scanning the copy at a higher resolution records the small letters more finely but cannot restore strokes the copier blurred. If the original exists, scan that.
• A full-length scan scaled down by software may still exist at full size. Look for the original scan file before the reduced one, and ask the person who scanned it whether the setting can be turned off and the job rerun.

When shrinking is acceptable

A reduced page is whole, and a whole page with small type is a better starting point than a cropped page with large type. If a reduced copy is all you have, convert it, and then proofread the smallest text on the page against the image: footnotes, initials lines, form numbers in the corner and any fine-print disclosure.

Mixed letter and legal pages in one feeder stack

A stack that mixes letter and legal sheets has three workable routes: a scanner mode that sizes each page separately, one pass with everything set to Legal, or two sorted batches. The one setting to avoid is Letter, which crops every legal page in the stack.

Three ways to handle it

For a file where order is evidence, such as a document produced as kept, scanning everything as Legal is the safest of the three. An empty strip under a letter page harms nothing. A cropped signature page does.

If you sort

1. Before pulling anything, number the pages lightly on a removable flag, or slip a colored sheet into the stack wherever a legal page comes out.
2. Scan the two batches.
3. Reassemble the PDF using the flags or slip sheets, and remove the slip sheet images.
4. Compare the page count of the PDF with the count of the paper.
5. Read across each join: the last line of one page should run into the first line of the next.

Feeding problems with mixed sizes

Feeder guides are set to one width. Letter and legal share a width, so they feed together well. Add a narrower sheet, such as a receipt or a half-page form, and it can swing sideways and come out tilted. A long sheet that is tugged or held back as it feeds moves unevenly past the sensor, and the image smears. The section on what makes a flatbed or sheet-fed scan come out blurred covers that, and the comparison of a flatbed and a sheet-fed scanner covers what each kind of hardware does to a page.

Two-sided legal pages add one more check. Confirm that the back of each long sheet was captured at full length too; the post on what duplex scanning is explains how back sides get lost.

Ledger sheets, long receipts and other pages that do not fit the scanner glass

A page wider or longer than the scanner can take has four routes: a larger scanner, a folded pass through a carrier sheet, a long-page feeder mode, or overlapping partial scans. Choose by the shape of the sheet, and keep every part of the page in at least one image.

Carrier sheets

Some sheet-fed scanners take a carrier sheet that holds a folded page. The ScanSnap help page for one model, the iX1300, on scanning documents larger than A4 or letter size with the carrier sheet says: "Documents larger than A4/Letter size (such as A3, 11 × 17 in. and B4) can be scanned with the optional Carrier Sheet. By folding the document in half and scanning in duplex mode, the two sides are merged into a single image." Check your own scanner's manual for whether it has an equivalent and which sizes it accepts.

Two cautions apply to any fold-and-merge method. The fold runs through the middle of the page, so read every word and figure along the join. And folding changes the paper, which matters if the sheet is an original someone will want to inspect.

Overlapping partial scans

When the page is longer than the glass, scan it in parts.

1. Place the top of the page at the top of the glass and scan.
2. Slide the page so the bottom is on the glass, with at least an inch of text repeated from the first scan.
3. Scan again at the same settings.
4. Keep both images, in order, and name them as parts of one page: p14-top, p14-bottom.
5. Convert both. In the working text, delete the lines that appear twice, after confirming the overlap matches.

Do not stitch the two pictures into one in an image editor unless you also keep the two untouched scans. A stitched image is something you made, and the join is a place where a line can be doubled or lost.

Wide tables bring a second problem after capture: columns. A ledger sheet with twenty narrow columns needs resolution for small digits and care with column boundaries. The guide to OCR on exhibits with tables covers that.

Long receipts

A receipt several feet long will not fit any standard page setting. If your sheet-fed scanner has a long-page or long-document mode, test it on a receipt you do not need. If not, scan the receipt in sections on the glass with overlap, and keep the sections in order. A very long image can also produce a large file. Schema Law accepts files up to 20 MB, and the post on converting a large scanned PDF online covers splitting a file that is over a limit.

Camera capture is another route for sheets no scanner in the office can take; the guide on how to OCR a photo of a document covers it.

Spotting cut-off text in a scan you were sent

You cannot see what a crop removed, so look for what should be at the edges and is not. A page that ends without a footer, a sentence that does not continue overleaf, or a form whose signature lines never appear each point to a missing strip.

Signs that an edge is missing

• Page dimensions. Open the document properties. A deed or mortgage form reported as 8.5 by 11 inches deserves a closer look. So does a file in which every page is letter length but the type looks small.
• No page numbers or footers. If the document has a footer on some pages and none on others, the others may be cropped.
• Broken sentences. Read the last line of each page and the first line of the next. A gap in sense between them means lines are missing.
• Skipped paragraph or line numbers. Section 14 followed by section 16. Line numbers in a pleading margin that stop at 22 on one page and at 28 on another.
• Text touching the edge. A last line with no white space under it, or letters sliced through horizontally.
• Missing signature and notary blocks. A signed agreement with no signature page content, where the text plainly leads up to one.
• Missing stamps and numbers. Bates numbers, recording stamps or exhibit stickers that appear on some pages and not on others in the same position.
• A side cut. Punch holes or a margin missing on one side only, with line ends clipped. The page was fed off-center or the width was set too narrow.

A quick check on a long file

1. Note the page size the viewer reports for the first, middle and last pages.
2. Scroll the file in a two-page or thumbnail view and look only at the bottom edges.
3. Stop at any page with no bottom margin and read across the page break.
4. List the affected pages by number.

Search will not help with this. Converted text contains only what the image contains, so a cropped clause is absent from the text without any gap to mark it. Schema Law shows the result page by page on screen, so you can read across each page break there, but the absence is yours to notice.

What to do about it

Ask the sender for a rescan of the listed pages at full length, and say why: "pages 6, 9 and 14 appear to be legal size and are cut off at eleven inches." A specific request is easy to act on. Keep the cropped file as received, and record the request.

A cropped edge is different from a page that is absent altogether. If whole pages seem to be gone, the guide to missing pages in a scanned contract sets out how to test whether a scan is complete.