Skip to content
JPG Smaller
Tools Compress Resize Convert Crop Utilities
Compress a JPG
Home / Research / JPG compression quality test

JPG compression quality test

People open a generic “compress JPG” tool expecting a smaller file. We ran the live Compress JPG tool on eight already-encoded photographs. 8 of 8 outputs were larger than the source.

Ran: 5 October 2026 Tool: compress-jpg Engine: ImageProcessor (GD)

Share this research

Bar comparison of original JPEG bytes versus quality-first compressed output for eight test photographs
Original file weight versus compress-jpg output on the same eight photographs. Values are measured bytes.

The question this page answers

If a photograph is already a web JPEG, does a quality-first compressor reliably reduce bytes while keeping the same pixels? That is a different question from “can I get under 100KB”. The 100KB question has its own experiment.

Method

Each source JPG was processed once with the public compress-jpg settings. Dimensions were left unchanged. The encoder writes a high-quality JPEG. We recorded original bytes from disk and output bytes from the generated file. No rows were discarded.

Subjects: alpine landscape, coast, forest, city, ceramic still life, books, flower, watch. Long edge about 2400 pixels. These files were already JPEG-encoded in the project library, not camera RAW.

Results

SubjectOriginalOutputPixels inPixels outChange
Alpine503 KB644 KB2400×16002400×1600+28% larger
Coast448 KB608 KB2400×15952400×1595+35.7% larger
Forest1.09 MB1.49 MB2400×15982400×1598+35.8% larger
City730 KB927 KB2400×16002400×1600+26.9% larger
Ceramic293 KB402 KB2400×16002400×1600+37.4% larger
Books406 KB549 KB2400×15512400×1551+35.5% larger
Flower502 KB743 KB2400×16012400×1601+47.9% larger
Watch106 KB188 KB2400×17442400×1744+78.4% larger

What the numbers say

Every photograph in this set grew. City is the clearest example: a 730 KB source became 927 KB at the same 2400×1600 pixels. Forest went from 1.09 MB to 1.49 MB. Watch, the lightest source at 106 KB, still grew to 188 KB.

That is expected when an encoder rewrites an already-optimized JPEG at a high quality setting. The source had already spent its compression budget. The new file stores more residual detail (and JPEG artifacts of a different pass) than the compact original.

Practical conclusion

Use Compress JPG when the file is obviously heavy and you want a quality-first rewrite at full size. Use a ceiling tool such as Compress JPG to 100KB when a form names a limit. Do not treat “compress” as a promise of a smaller file.

Limits of this test

The set is eight still photographs. It does not include screenshots, documents, or illustrations. We did not sweep quality sliders. We did not score visual difference with a lab metric. A different encoder or a lower default quality would produce different bytes. The finding is about this engine, on this set, on 5 October 2026.

Technical reference

JPEG on the web is typically a JFIF file containing a lossy DCT image. MDN’s format guide is a clear overview of image/jpeg and when a lossless format is a better fit: MDN image file types.

Related

  • 100KB JPG compression test
  • JPG to WebP file size test
  • How we test
  • Compress JPG
JPG Smaller

Image tools for real upload limits: measured compression, exact resize, format conversion, crop, and inspection.

Common tools

  • JPG Compressor
  • Compress JPG to 50KB
  • Compress JPG to 100KB
  • PNG Compressor
  • Reduce Image Size
  • Image Resizer
  • PNG to JPG
  • JPG to WebP

Categories

  • Image Compression
  • Image Resize
  • Image Converter
  • Image Crop
  • Utilities
  • All tools

Company

  • About
  • How we test
  • Research
  • Contact
  • [email protected]
  • Privacy
  • Terms
  • Sitemap

Share this page

© 2026 JPG Smaller. Temporary processing, files are not kept as a permanent library.

Home Tools Compress Convert

More

  • Resize
  • Crop
  • Utilities
  • Categories
  • About
  • How we test
  • Research
  • Contact