OmniKit
OMNIKIT · PRACTICAL REFERENCE

When PDF compression makes a document larger

Download a text PDF and an image-only PDF, compare measured reference outputs, and see why searchable text matters.

Choose the workflow before choosing a preset

OmniKit’s PDF compressor renders each page, encodes that rendering as a JPEG, and creates a new PDF from those page images. This can help a scan. It can also turn an efficient text document into a larger file with no selectable text. “Compression complete” tells you that processing finished, not that the new document is better.

The following pair of original one-page documents makes the distinction visible. Both are 540 × 720 PDF points. The scan contains only an image; the text document contains searchable text. Neither includes private information or customer content.

What happened in the reference run

DocumentOriginal bytesOutput bytesSelectable text
Image-only page195,313144,075Absent before and after
Text page1,48762,166Present before; absent after

The reference uses the tool’s Balanced settings: 132 DPI and JPEG quality 0.75. The scan became smaller, while the text document became much larger. Both output files retain one page and the same page dimensions. Searching for OK-2026-0905 works in the text original and not in its image-only output.

Repeat it in the actual tool

  1. Open PDF compression and load the scan original. Select Balanced and process it. Record both sizes and check small text at a readable zoom.
  2. Load the text original. The content analysis should detect searchable text. Read the flattening notice before confirming that an image-only copy is acceptable.
  3. Process the text document. Download it and try selecting text or searching for the reference code.
  4. Keep each original beside its output. Choose the original whenever the output is larger or loses a capability you need.

A useful decision before uploading a document

What you needRecommended next step
Smaller scan for sharingTry compression, then inspect all pages
Selectable text or copyable tablesPreserve the source; this compressor does not preserve the text layer
Reorder or remove pagesUse PDF organization and inspect the exported file
Editable text from a scanUse OCR and check names, amounts, punctuation, and reading order
Forms, links, or certificate signaturesRetain the original; image-only compression flattens these features

OCR can recover a draft of visible text, but it does not restore the original form fields, links, or certificate signature. A tiny file is not a useful result if it breaks the task the recipient needs to complete. Download a new copy and retain the original until the result has been checked.

Continue reading

Explore the other practical guides →