FluffyBench

What does an image quality number actually mean?

That Quality slider can shrink a photo without changing its dimensions. Here's what the number controls and how to decide whether the smaller copy still looks right.

FluffyBench · 5 min read · Measured image quality and choosing a saved copy

What does the image Quality slider control?

Think of Quality as a dial for how closely a saved picture follows its source. Turn it down and the software can simplify more detail to save space. Turn it up and you're asking it to keep a closer match. The number isn't a percentage of the picture that survives.

Quality describes one save in one format. Matching JPG and WebP numbers aren't the same promise about detail or file size.

Conceptual Quality dial: lower settings allow more simplification; higher settings ask for a closer match. JPG and WebP quality scales differ.
A direction to try, not a prediction of bytes or visible damage. How we measured and reviewed

JPG uses lossy compression: it stores an approximation of your picture to reduce the file size. That trade can be useful when you're sending a photo or putting it on a website. The question is how much simplification your picture can take before you mind the difference.

How did one photo change at JPG quality 50?

Saving a lily photo that was already a JPG at quality 50 made it about 53% smaller, with the same pixel dimensions. The marked petal patch lets you compare texture in the source and saved copy.

Whole lily photograph with the compared petal detail marked.

Original JPG

Dark speckles and fine grain in the original lily petal.

112.0 KB

Quality 50

The same lily speckles and grain after saving at quality 50.

53.0 KB

Compare the grain around the dark speckles; sizes are for complete JPGs. Your photo's texture and starting compression affect the trade. · Photo: Rosendahl; comparison assets made by FluffyBench, Public domain, released by the author, via Wikimedia Commons and FluffyBench.

Need to rotate or flip an image before reencoding?

FluffyBench image crop tool

FluffyBench's image crop tool handles a rotation or flip when the photo's orientation also needs changing.

How do you decide whether a saved JPG looks right?

The detail that matters depends on your image. These clean and affected illustrations show changes worth looking for in your source and downloaded copy.

Clean and affected illustrations: smooth versus banded shading, fine versus smeared texture, readable versus blurred text, clean versus rough edges. Conceptual illustrations, not measured outputs.
Look for broken-up shading, smeared texture, hard-to-read letters or rough edges.

Review the downloaded JPG outside FluffyBench's JPG converter. These pictured checks use the measured lily pair and an illustrated page layout.

Review steps: compare source and copy at matching zoom; view at intended display size; check destination format and size rules. If detail suffers, make a higher quality candidate from the source. If it looks right and fits, use the copy.
Choose the copy when its look and size work for where you're using it. · Photo: Rosendahl, Public domain, released by the author, via Wikimedia Commons.

Why can't another JPG save restore lost detail?

The next save starts with the copy you have now. It's like making a photocopy of a photocopy: the new pass works with the detail left in its input.

Conceptual detail across saves: the source has fine marks, saving simplifies them, and another save keeps the simplified pattern instead of restoring fine marks.
The marks illustrate detail being simplified during a save. A higher quality setting or a different format can't bring the missing marks back; this is a conceptual illustration.

Choosing JPG quality in FluffyBench

FluffyBench's JPG converter lets you make a candidate with a chosen Quality setting. Here's the job from source photo to downloaded copy; the lily's setting is an example, not a recommended number for every image.

Keep the format, encoder and pixel dimensions fixed while you try a quality setting. That makes the comparison about Quality.

  1. JPG converter intake with the Choose images button.
    JPG converter intake with the Choose images button.
    Step 1. Choose images: start from your best source, and keep the original.
    1. Choose images opens the file picker.
  2. The toolbar gear used to open Settings.
    The toolbar gear used to open Settings.
    Step 2. Open Settings with the toolbar gear.
    1. The gear is at the top right of the workspace.
  3. JPG Quality control set to 50.
    JPG Quality control set to 50.
    Step 3. Choose Quality 50 for a candidate copy.
    1. Quality trades picture detail for file size.
  4. Size control with Original selected.
    Size control with Original selected.
    Step 4. Keep Size at Original so resizing does not change the comparison.
    1. Original keeps the pixel dimensions.
  5. Encoder control with Standard selected.
    Encoder control with Standard selected.
    Step 5. Keep Encoder at Standard to match the browser-encoder example.
    1. Standard uses the browser JPG encoder.

With the settings chosen, return to the photo to make and download your candidate.

  1. The toolbar Settings toggle while settings are open.
    The toolbar Settings toggle while settings are open.
    Step 1. Close Settings with the same toolbar gear.
    1. Your chosen settings stay in place.
  2. The enabled Convert button in the JPG converter.
    The enabled Convert button in the JPG converter.
    Step 2. Press Convert to make the JPG with these settings.
    1. This candidate uses JPG quality 50.
  3. Completed JPG batch size summary, showing input and output sizes.
    Completed JPG batch size summary, showing input and output sizes.
    Step 3. Check the completed size summary.
    1. The finished batch shows input and JPG file sizes.
  4. Download button for the completed JPG.
    Download button for the completed JPG.
    Step 4. Download the JPG as a separate copy for comparison with your source.
    1. Download saves the JPG to your device.

Need exact pixels for a graphic? FluffyBench's WebP converter offers Lossless compression. Our lossy and lossless WebP comparison shows that choice in practice.

How we reviewed image quality

We measured this JPG save on October 15, 2026, at the source dimensions with resizing off. The color map records changes in the same petal patch; a changed color value isn't a measure of visible damage.

Source JPG

An enlarged crop of dark lily speckles and petal grain from the source JPG.

x=896 px, y=640 px; same 37 × 64 pixel region.

37 × 64 px; 8× magnification.

Original JPG input; the crop is enlarged without resizing it first.

112.0 KB111,968 bytes

JPG quality 50

The same enlarged lily crop after JPG quality 50 encoding.

x=896 px, y=640 px; same 37 × 64 pixel region.

37 × 64 px; 8× magnification.

Browser JPG encoder at quality 50; the crop is enlarged without resizing it first.

53.0 KB52,950 bytes

Color changes

A color map showing how much the red, green, or blue value changed at each pixel in the same lily crop, from zero to 26 channel levels.

37 × 64 px; 8× magnification.

A color map showing how much the red, green, or blue value changed at each pixel in the same lily crop, from zero to 26 channel levels.

Maximum channel difference 026

Decoded-pixel difference: 1,190,376 of 1,228,800 pixel positions (96.9%) across the full 1280 × 960 px image.

The map shows the largest red, green or blue change at each position in the crop. · Photo: Rosendahl, Public domain, released by the author, via Wikimedia Commons.
Search FluffyBench