SumatraPDF Copied Text Is Garbled: How to Fix It

You select a readable paragraph in SumatraPDF, press Ctrl+C, and paste it into Notepad, Word, or a browser. Instead of the expected text, you get wrong characters, missing spaces, boxes, scrambled columns, or unreadable output. Although this can look like a display problem, garbled copied text usually points to the document's hidden text layer rather than the fonts displayed on screen.

The most likely causes are defective PDF character mappings, poor optical character recognition, custom embedded fonts, ligatures, an unusual page layout, or limitations in the document format. A damaged SumatraPDF configuration is possible, but it is less common. The tests below help you identify which category applies before you reset settings or reinstall anything.

1. Confirm the Symptom With a Simple Test

Begin by determining whether the problem follows SumatraPDF, the destination application, or one particular document. This distinction prevents unnecessary changes to Windows and SumatraPDF.

1.1 Copy from a known-good document

Open a simple, digitally created PDF containing ordinary selectable text. A PDF exported from a recent word processor is suitable. Copy one complete sentence and paste it into Notepad. Notepad is useful because it removes rich-text formatting from the test.

• If the sentence pastes correctly, SumatraPDF's basic clipboard function is working. The original document is probably the problem.
• If several unrelated, known-good PDFs paste incorrectly, investigate SumatraPDF's configuration, the application receiving the text, or clipboard software.
• If text works in Notepad but not in another program, the destination program or its formatting rules are involved.

Success means that ordinary letters, spaces, punctuation, and sentence order survive the copy-and-paste operation. Once a known-good file works, stop changing global SumatraPDF settings and concentrate on the failing document.

1.2 Compare the same page in another PDF viewer

Open the troublesome PDF in another reputable viewer, such as Microsoft Edge or Adobe Acrobat Reader. Copy exactly the same short passage and paste it into Notepad.

• If every viewer produces the same bad characters, the PDF's text layer is malformed or incomplete.
• If another viewer extracts the passage correctly, it may be interpreting a difficult font mapping, ligature, or layout differently.
• If the output is readable but the line order differs between viewers, the issue is probably page structure rather than character encoding.

This comparison is diagnostic, not proof that either viewer is universally better. PDF viewers can use different extraction heuristics. If one viewer handles this specific document correctly, using it for extraction is often the fastest practical fix.

2. Understand Why Readable PDF Text Can Copy Incorrectly

A PDF describes how marks should be placed on a page. It does not always store text in the same straightforward sequence used by a word-processing document. The visual page can therefore look perfect while its copied text is unusable.

2.1 PDF encoding and ToUnicode maps

A PDF font can assign internal character codes to glyphs, which are the visible shapes shown on the page. To copy those shapes as meaningful text, a viewer needs to determine which Unicode character corresponds to each code. PDFs commonly provide this information through a ToUnicode character map.

If the map is missing, incomplete, or wrong, a viewer may display the intended glyphs while copying unrelated letters or symbols. For example, the visible shape for an uppercase A might internally use a code that does not map to Unicode A. A viewer may attempt to infer the character, but reliable recovery is not always possible.

There is no general SumatraPDF setting that can reconstruct a missing character map. If multiple viewers fail in the same way, the durable solution is to obtain a better copy, export the document again from its source, or create a corrected searchable text layer through legitimate OCR software.

2.2 Ligatures and embedded custom fonts

Some fonts use ligatures, where combinations such as “fi” or “fl” are represented by one glyph. Correctly produced PDFs map that glyph back to the appropriate letters. Poorly produced files may omit or misstate the mapping, causing missing letters, unexpected symbols, or boxes when copied.

Custom subset fonts can create a similar problem. Font subsetting is normal and helps reduce file size, but the PDF still needs enough mapping information for extraction. Installing a visually similar Windows font generally does not repair a defective text layer because the issue is the relationship between internal codes and characters, not merely the availability of a display font.

2.3 Two-column pages and broken reading order

Text on a PDF page may be stored according to drawing operations rather than human reading order. On a two-column page, copied text might jump from the first line of the left column to the first line of the right column. Headers, footers, captions, and sidebars can also appear in unexpected positions.

Try selecting one column or one paragraph at a time instead of dragging across the full page. If smaller selections paste correctly, the characters are intact and only the reading order is ambiguous. In that case, settings resets and font changes will not help. Use narrow selections, another viewer with a different layout heuristic, or return to the source document when accurate structural extraction matters.

2.4 Scanned pages and unreliable OCR

A scanned PDF is primarily a collection of page images. It may have no selectable text, or it may contain an invisible OCR layer behind each image. If that OCR layer contains mistakes, SumatraPDF copies those mistakes even though the scanned words look correct to a person.

Typical OCR symptoms include confusion between O and 0, l and 1, missing spaces, incorrect punctuation, or words from adjacent columns being mixed together. Zooming the page or changing rendering options does not improve the hidden OCR results.

Success requires a new or corrected OCR text layer. If you own the document or have permission to process it, run OCR using a trusted document application and choose the correct language. Then test a sentence with names, punctuation, and numbers. Do not attempt to bypass copying restrictions, DRM, passwords, or permissions imposed by the document owner.

3. Check SumatraPDF and Format-Specific Factors

SumatraPDF supports several document families, but copying behavior depends on the information available in each format. A PDF text-layer problem should not automatically be treated like a CHM, DjVu, XPS, comic, or image problem.

3.1 Verify which application and file type are open

Confirm the title bar shows SumatraPDF and inspect the file extension in File Explorer. A Windows file association decides which application opens a file by default, but it does not rewrite the document's character encoding. Correcting an association helps only when the file has been opening in a different application or the wrong program has been blamed for the result.

For images and image-only comics, there may be no text layer to copy. DjVu and XPS files can have their own text extraction characteristics. CHM content may depend on the text and layout embedded in the help file. Test a regular PDF separately before concluding that SumatraPDF copying is broken across all formats.

3.2 Distinguish rendering from extraction

If letters look wrong on screen and also copy incorrectly, there may be both rendering and encoding issues. If the page looks correct but pasted text is wrong, focus first on extraction. Display quality options, zoom levels, color settings, and printer settings usually do not repair PDF character mappings.

Likewise, Ghostscript is relevant to certain PostScript workflows and external conversions, but it is not a routine fix for copying text from an ordinary PDF in SumatraPDF. Codec packs and printer drivers are also unlikely to affect direct clipboard extraction. Investigate them only if the failure occurs during a specific conversion, print-to-PDF, or image-decoding workflow.

3.3 Check document permissions without bypassing them

Some PDFs declare restrictions on copying or content extraction. A viewer may limit copying or provide no usable output. Check the document properties in a viewer that displays security information, or ask the document provider for an accessible copy.

Do not use tools intended to defeat restrictions. If copying is intentionally disabled, the appropriate fix is permission from the owner, an accessible source file, or an authorized alternative version.

3.4 Test the original source when available

If the PDF was generated from Word, a browser, desktop-publishing software, or a reporting system, compare it with the source document. Copying directly from the source often preserves paragraphs, spaces, lists, and reading order more accurately.

If you control PDF creation, export it again with fonts embedded and accessibility or tagging options enabled where available. Test the resulting PDF in more than one viewer. Success means both the visual page and copied text match the source. Printing the malformed PDF to another PDF may preserve its appearance, but it can flatten text into images or retain poor mappings, so it is not a dependable repair method.

4. Inspect SumatraPDF-settings.txt Safely

SumatraPDF stores preferences in a text configuration file named SumatraPDF-settings.txt. An installed copy normally keeps user settings under the current Windows user's local application data, while portable operation may keep settings with the executable. The exact active location depends on how SumatraPDF is being run.

There is no standard setting that fixes a PDF with a broken ToUnicode map. Inspecting or resetting the file is appropriate only when known-good documents fail, settings do not persist, or installed and portable copies behave differently.

4.1 Back up the settings file

1. Close all SumatraPDF windows.
2. Locate SumatraPDF-settings.txt in the SumatraPDF folder under your local application data or beside the portable executable.
3. Copy the file to a safe location or rename it to SumatraPDF-settings-backup.txt.
4. Do not edit the original while SumatraPDF is running because the application may overwrite changes when it closes.

If you are uncertain which file is active, change a harmless preference in SumatraPDF, close the application, and check which settings file received a new modification time. Restore the preference afterward.

4.2 Reset only as a controlled test

With SumatraPDF closed and the settings file backed up, rename the active settings file. Start SumatraPDF again so it can create a clean configuration, then test one known-good PDF and the troublesome PDF.

• If known-good files now copy correctly, the previous configuration or execution environment was involved.
• If only the troublesome PDF still fails, restore your preferred settings and treat the document as the cause.
• If nothing changes, repeated settings deletion is unlikely to help.

Success after a reset means the same test sentence that previously failed now pastes accurately. If the malformed document remains malformed across clean settings and other viewers, stop resetting SumatraPDF.

5. Run a Clean Temporary Test Before Reinstalling

Reinstallation often leaves per-user settings untouched, so it is not the best first diagnostic step. A clean portable test provides a clearer comparison without disturbing your normal setup.

1. Download SumatraPDF only from its official website.
2. Place the portable build in a new, writable folder.
3. Do not copy your existing SumatraPDF-settings.txt into that folder.
4. Open a known-good PDF directly from the temporary copy.
5. Paste one sentence into Notepad.
6. Repeat the test with the failing document and compare it in another viewer.

If the portable copy handles known-good documents correctly, your regular installation, shortcut, settings location, or surrounding clipboard software deserves attention. Verify that old shortcuts do not launch another copy of SumatraPDF from an unexpected folder.

If installed and portable copies produce identical bad output from one PDF, reinstalling is unlikely to repair that file. If every document fails only in SumatraPDF, obtain the current supported release from the official site and retest before reporting the issue. Keep a small non-confidential sample PDF that reproduces the behavior if you need to ask for technical help.

6. Quick Fix Checklist

1. Paste into Notepad to exclude rich-text formatting problems.
2. Copy from a simple known-good PDF.
3. Test the same passage in another reputable viewer.
4. Select one column or paragraph at a time.
5. Determine whether the page is scanned and relies on OCR.
6. Check whether copying is restricted by document permissions.
7. Use the source document when accurate structure is essential.
8. Re-export or legitimately OCR a malformed document when authorized.
9. Back up SumatraPDF-settings.txt before resetting it.
10. Run a clean portable test before reinstalling.

Stop changing SumatraPDF settings when known-good documents copy properly and the same PDF fails in multiple viewers. At that point, the evidence points to the document's text layer, OCR, or page structure. The practical resolution is a better source file, corrected export, authorized OCR repair, or a viewer that can interpret that particular file more successfully.

7. Frequently Asked Questions

7.1 Why does the PDF look correct but paste as nonsense?

The viewer can draw visible glyphs without knowing their correct Unicode identities. A missing or incorrect ToUnicode map may allow perfect visual output while making reliable text extraction impossible.

7.2 Can installing the document's font fix copied text?

Usually not. Embedded and subset fonts can display correctly without an installed Windows equivalent. Garbled copying more often reflects missing character mappings inside the PDF. Installing fonts may help a genuine display substitution problem, but it cannot generally rebuild the text layer.

7.3 Why are spaces missing or words joined together?

PDFs can position each word or glyph geometrically without storing ordinary space characters. A viewer must infer spaces from distances. Tight kerning, custom fonts, OCR errors, and unusual text placement can make that inference unreliable.

7.4 Why does copied text jump between columns?

The page may not contain a defined logical reading order. Try selecting one column at a time. For repeated extraction or accessibility needs, obtain a properly tagged PDF or return to the structured source document.

7.5 Will reinstalling SumatraPDF fix garbled copied text?

Reinstallation is unlikely to fix one badly encoded PDF. First test known-good documents, another viewer, clean settings, and a portable copy. Reinstall only if failures consistently follow the regular SumatraPDF installation rather than the source file.

7.6 What is the realistic fix for a malformed PDF?

The best fix is a corrected PDF exported from the original source with valid text mappings. For authorized scanned documents, accurate OCR can create a new searchable layer. If neither option is available, extraction may remain imperfect, and manual correction or use of another viewer may be necessary.