How to Compress a PDF (Without Wrecking the Text)

"This PDF is too big to email" is one of the most common file-format complaints. The reason it happens, and the reason there's no one-click fix, is that PDFs can contain wildly different content — text, vector graphics, raster images, embedded fonts, attachments, metadata. Different compression strategies attack different parts. Each strategy has trade-offs.

This guide walks through the four real strategies, when each one's appropriate, and what Dropvert's PDF Optimizer actually does.

Why PDFs get huge

Three usual suspects:

1. Embedded raster images at higher resolution than needed. A scan at 600 DPI is 4× the bytes of one at 300 DPI, with no visible benefit on screen.
2. Embedded font subsets you don't need. Some PDF generators embed the entire Unicode glyph table "just in case" — 5-50 MB of unused glyphs.
3. Multiple revision histories or layers. PDFs created by repeatedly editing in Acrobat can accumulate "deletion as overlay" — the old content is hidden, not removed.

The two reliable size-reduction strategies attack #1 and #2. The other two attack the structure of the document itself.

Strategy 1 — Downsample images

The biggest single win for image-heavy PDFs (scans, reports with charts, photos). The tool walks every embedded image, recompresses it at a lower DPI / quality, and writes the result back into the PDF stream.

Effect on file size: 40-90% smaller for image-heavy PDFs. Effect on text: none — text remains selectable and searchable.

This is what most "compress my PDF" tools do under the hood. Adobe Acrobat Pro's "Reduce File Size" function uses this strategy.

Strategy 2 — Rebuild the PDF as flattened pages

Render each page to a JPEG, then rebuild a new PDF with those JPEGs as the page content. Aggressive — typically 60-80% smaller on any input — but text loses selectability. The page is now an image of text, not text.

This is what Dropvert's PDF Optimizer does. It's the simplest reliable approach, and the right trade-off when:

• You're sending the PDF to be viewed, not edited or searched.
• The PDF has lots of complex elements (gradients, transparency, layers) that resist per-stream compression.
• You need a hard size cap and don't have time to fiddle with per-image quality.

If you need to keep text selectable, this isn't the right tool — Acrobat or a per-image-downsample path is.

Strategy 3 — Subset fonts

Strip embedded fonts down to only the glyphs actually used in the document. Effect: 5-30% smaller, no visible change.

Most modern PDF generators already subset fonts correctly. This strategy mostly helps with PDFs from older office suites or template-heavy "Save as PDF" exports. Dropvert doesn't offer font subsetting today; for that workflow, Ghostscript's command-line tool is the most reliable free option.

Strategy 4 — Remove non-essential structure

Drop metadata, annotations, form fields, attached files, JavaScript, embedded thumbnails. Effect: usually 1-5% smaller on its own, but can matter on PDFs that have been through many edits.

Combined with #1 or #2, you sometimes squeeze out another few percent.

Which one should you pick?

Your PDF	Best strategy
Image-heavy report and text must stay selectable	Image downsample (Acrobat Pro, or a third-party tool with that flow)
Image-heavy report and viewing only	PDF Optimizer — fastest path
Mostly text, large because of fonts	Font subset (Ghostscript)
Came from Acrobat, you've edited it many times	"Reduced Size PDF" save in Acrobat — redoes structure
You're going to print it	Don't compress — printers want the raw 300 DPI images

How to compress with Dropvert

1. Open PDF Optimizer.
2. Drop the PDF.
3. Pick a quality preset (Standard / Smaller / Smallest).
4. Click Optimize.

Standard preset: ~50% smaller for most input. Smaller: ~70%. Smallest: ~85% with visible JPEG artefacts on detailed images.

Output is downloadable as soon as it finishes — the file never leaves your browser.

FAQ

Will the text be searchable after Dropvert's optimizer? No. The flatten-to-JPEG strategy removes the text layer. If text-searchability matters, run PDF OCR on the optimised file afterward to bake a new text layer on top — the optimised JPEG pages get OCR'd and the resulting layer is invisible to viewers but searchable.

Why is my optimised PDF actually larger? Two reasons: (a) the source already had heavily compressed JPEG images, so re-flattening adds JPEG-on-JPEG artefact tax without saving bytes, or (b) the page count grew because the source had pages stored as one giant image that got split. For (a), skip optimisation; for (b), check the source generator settings.

Does it work on password-protected PDFs? No. Strip the password first (Acrobat → File → Properties → Security → No Security). The browser-side PDF library doesn't decrypt encrypted content streams.