Tại sao tệp PDF của tôi lại quá lớn? 7 nguyên nhân và cách khắc phục

PDF files can balloon to enormous sizes for reasons that aren't immediately obvious. A 2-page letter shouldn't be 45 MB — but it often is. Here's exactly why, and what you can do about it.

The 7 Most Common Causes of Large PDFs

1. High-Resolution Embedded Images

This is the #1 culprit. When you embed a 12-megapixel photo taken on a modern smartphone, it arrives in your PDF at full resolution — often 5–15 MB per image. A report with 10 such images is immediately a 50–150 MB file.

The fix: Compress images to web resolution (72–150 DPI) when screen viewing is the goal, or print resolution (300 DPI) when printing matters. The WhiskFile Compress PDF tool automatically downsamples images using Ghostscript's /ebook setting, which produces excellent quality at 60–80% smaller file sizes.

Expected reduction: 50–80% for image-heavy PDFs.

2. Unsubsetted Font Embedding

When a PDF embeds a font, it can embed the entire font file — including every character you didn't use. A full Unicode font like Noto Sans can be 1–2 MB by itself. A document using 5 such fonts adds 5–10 MB before a single word of content.

The fix: Use a PDF compressor that subsets fonts (only embeds the characters actually used). Most modern PDF generators do this correctly, but older tools often don't. Ghostscript with -dSubsetFonts=true handles this correctly.

Expected reduction: 20–50% for text-heavy documents with multiple fonts.

3. Redundant or Duplicate Resources

If your PDF was assembled by merging multiple documents, identical resources (images, fonts, color profiles) may be embedded multiple times. A company logo appearing on 50 pages might be stored 50 times instead of referenced once.

The fix: Run the PDF through a PDF optimizer that deduplicates resources. The Compress PDF tool applies this deduplication automatically.

Expected reduction: 10–40% for merged PDFs.

4. Embedded Thumbnails and Previews

Some PDF creators embed page thumbnail images to enable fast preview in certain PDF readers. These thumbnails are additional images stored inside the file, doubling the space needed to represent each page.

The fix: Remove embedded thumbnails using Ghostscript's -dCreateJobTicket=false and related optimization flags. Our compressor does this.

Expected reduction: 5–20%.

5. Uncompressed Content Streams

Page content in a PDF (vectors, text rendering instructions) is stored in content streams. These should be compressed with Flate (zlib/deflate) compression, but older PDF creators sometimes leave them uncompressed.

The fix: Re-export through any modern PDF pipeline (including Ghostscript or pdf-lib's optimization pass) to apply stream compression.

Expected reduction: 10–30% for vector-heavy PDFs.

6. Metadata and Hidden Data

PDFs can contain revision history, embedded XML metadata, document information dictionaries, and even deleted but not fully removed content. A document with 50 revision cycles can carry significant hidden overhead.

The fix: Flatten the PDF through a compression pipeline to strip revision history. Also consider using the Edit PDF Metadata tool to remove excessive metadata.

Expected reduction: Varies — usually 5–15%.

7. Color Profiles

Embedded ICC color profiles (for print-accurate color management) can be 500 KB–2 MB each. A PDF prepared for professional printing often carries full CMYK profiles that are completely unnecessary for digital distribution.

The fix: When distributing for screen viewing, use Ghostscript's /screen or /ebook PDF settings, which strip color profiles and convert to sRGB.

Expected reduction: 5–15% for documents with embedded profiles.

PDF File Size Benchmarks

Based on analyzing 10,000 PDFs processed through WhiskFile in June 2026:

| Document Type | Typical Unoptimized Size | After Compression | Reduction | |---|---|---|---| | 10-page text report | 500 KB – 2 MB | 200 – 800 KB | 40–60% | | 20-page photo brochure | 20 – 80 MB | 4 – 16 MB | 70–80% | | Scanned letter (300 DPI) | 2 – 5 MB | 600 KB – 2 MB | 60–70% | | Merged invoice pack (20 files) | 40 – 120 MB | 8 – 25 MB | 65–80% | | Presentation (50 slides, images) | 100 – 300 MB | 15 – 50 MB | 75–85% |

How to Check What's Making Your PDF Large

Before compressing, it helps to know what's inside your PDF:

1. Open in Adobe Acrobat Reader → File → Properties → Description tab shows embedded fonts.
2. Use WhiskFile's PDF to Text tool to check if the PDF has a text layer (if not, it's likely a scanned image PDF with no compression benefit from font optimization).
3. Check image count: if your file is mostly images, compression will give the biggest gains.

The Right Compression Level for Your Use Case

| Use Case | Recommended Setting | Expected Size | |---|---|---| | Email attachment | Small / Screen | Typically under 5 MB | | Web embed / download | Balanced / Ebook | Good quality, 50–75% smaller | | Print-on-demand | Quality / Printer | Minimal loss, 20–40% smaller | | Archival | None (keep original) | Full quality preserved |

Rule of thumb: For anything that will only be read on a screen, /ebook compression is the right choice. It produces files that look identical to the original at screen resolutions while reducing size by 60–80%.

Quick Answers

Why is my 2-page letter 45 MB? Almost certainly because it was created from a scanned image at very high DPI (300–600 DPI), or contains one or more high-resolution photographs. Compress it with the /screen setting to reduce it to under 1 MB for email.

Will compression reduce PDF quality? For screen viewing: no noticeable difference. For print: use the "Quality" setting to minimize any quality loss. The Balanced setting is the right default for 95% of use cases.

Can I compress a password-protected PDF? Not directly — you need to remove the password first using the Unlock PDF tool, then compress, then optionally re-protect it.