Image guidesPractical guides

Why repeated JPG exports get blurry: keep a master and export fresh copies

A photo is compressed for a colleague, edited with a caption, compressed again and downloaded for another resize. Small text eventually develops rough edges. Start by tracing the files and export steps: moving the quality slider cannot tell you where detail was lost.

person editing photo on computer
Illustrative photo: Glenn Carstens-Peters · Unsplash License

The practical answer

Keep the earliest source and export each delivery version from that source or an editable master. Copying or renaming a file differs from encoding it again. Converting a damaged JPG to PNG does not restore missing detail. For a new size, return to the master and create a fresh copy with Image Conversion.

Identify copying, resizing and encoding separately

Open the source, the received copy and the current file. Compare their pixel dimensions first. Smaller dimensions indicate resizing; unchanged dimensions do not rule out a new encode. A smaller file alone does not establish equal visual quality or an untouched source.

MDN’s image-format guide identifies JPEG as lossy and PNG as lossless. Our editorial recommendation is to keep continuing-edit files separate from delivery copies. The visible damage depends on the input, encoder and settings; there is no universal number of saves that predicts it.

Copying a JPG in a file manager does not recalculate its pixels. An editor’s export, a conversion website or a platform-generated download may create a new encoded version. These actions should not all be described simply as saving.

Different actions in an image delivery chain
ActionNecessarily re-encoded?What to check
Copy or renameNo new encode requiredThe copied file should retain the same content
Export JPG from an editorIt may be re-encodedReview export settings and compare with the source
Resize and exportNew pixels and a new fileInspect dimensions and edges
Convert existing JPG to PNGStores the decoded imageEarlier lost detail is not recovered

A product-photo example with two delivery versions

This is an illustrative workflow, not a KitNelo test. Suppose a 2400 × 1600 product photo must become a website image and a captioned social image. Keep the source and save an editable master. Produce both delivery sizes from that master instead of building the social image from the smaller website JPG.

Names such as product-master, product-web-1200 and product-social make the roles explicit. Record dimensions and intended use. Return to the master when changing the caption. If your earliest source is already a JPG, keep it, but do not describe it as a lossless original merely because you have no better copy.

One master, independent exports

A fragile export chain

Source → compressed JPG → caption → smaller JPG → enlargement

Suggested delivery chain

Source → editable master → website JPG → social PNG/JPG

Both delivery files start from the same master. Choose PNG according to text edges, transparency and recipient requirements. Larger files are not automatically clearer.

Export one version before processing the batch

  1. Import the earliest source or confirmed master into Image Conversion. Note its dimensions. Decide the subject and final aspect ratio before cropping.
  2. Choose JPG, PNG or WebP for the recipient. Change one important size or quality setting at a time so you can identify its effect.
  3. Download one copy. Inspect small text, thin lines and high-contrast edges at actual pixel scale, then at the intended display size. A thumbnail is insufficient.
  4. Once acceptable, process the batch and keep the master. Generate later revisions from that master rather than from a delivery copy.

For a 200 KB upload requirement, use the size and clarity compression guide. Confirm dimensions and appearance, then verify the final byte size rather than relying on a quality percentage.

Acceptance checks and limits of conversion

Inspect edges against light and dark backgrounds. Review text, color and cropping on the recipient’s actual page, because the platform may compress again. Compare files at the same zoom level: different viewer scaling can look like an image-quality change.

A tiny source, missed focus or detail already discarded by compression cannot be repaired by changing formats. Find the camera original or rebuild the text layer when possible. PNG may be larger; JPG cannot preserve transparency. Review the PNG, JPG and WebP use cases before choosing.

Browser conversion is not a print color-management workflow, professional restoration service or lossless JPEG transcoder. Confirm exact color, animation and unusual bit-depth requirements in appropriate software, and retain a traceable source.

Record before handing off

  • The source and editable master remain available
  • Dimensions and format match the destination
  • Actual pixels and intended display size were inspected
  • Transparency and text have no unexpected changes
  • The recipient’s processed upload was checked

Common questions

Does converting a blurry JPG to PNG make it sharp?

It does not recover missing information. PNG can avoid another lossy encode while preserving the current image, including existing artifacts.

Does JPG quality 100 mean no loss?

No such guarantee follows. Quality is an encoder setting, not a recovery capability or a universal score across software. Compare with the source.

Does downloading or copying reduce quality?

Ordinary byte copying does not encode again. A platform may offer a processed download, so confirm whether it is the original file or a generated copy.

Further reading & sources

Put the guide into practice

Open the relevant tool, preview the result and check your exported file.

Newer articles on this topic

Continue with a newer article that refers to this guide or explainer.

Related articles