Image quality is more than a number on a slider.

Lossy and lossless are different goals

Lossless formats preserve the stored pixel information when the image is encoded and decoded. PNG is a familiar example. Lossy formats such as JPG deliberately discard some information to reduce file size, aiming to remove details that are less noticeable.

Neither approach is automatically better. A photograph can often tolerate lossy compression extremely well and become much smaller. A screenshot with small text and sharp interface lines may reveal compression artifacts quickly and is often better suited to lossless storage.

What a quality slider changes

A quality control for JPG, WebP, AVIF, or another lossy output usually influences how aggressively the encoder compresses the image. A higher number generally spends more bytes preserving detail; a lower number generally produces a smaller file with a greater risk of visible artifacts.

The number is not an absolute quality measurement that can be compared perfectly between formats or encoders. “Quality 80” in one codec is not guaranteed to match “quality 80” in another. Treat the slider as a tuning control and inspect the result.

Repeated conversion can cause generation loss

If you open a JPG, save it as another JPG, then repeat the cycle again and again, each lossy re-encoding can introduce additional changes. This is sometimes called generation loss. It may be subtle at first but can accumulate around edges, texture, gradients, and fine detail.

That is why keeping a master matters. Edit from the best source you have, then export delivery copies for email, web, social media, documents, or other destinations.

Resizing can matter more than compression

One of the biggest mistakes in web image optimization is focusing only on format and quality while ignoring dimensions. A 5000-pixel-wide photo displayed in a 700-pixel content column contains far more pixels than the layout needs.

Reducing dimensions to match the intended display size can dramatically reduce file weight before compression is even considered. A well-sized JPG or WebP can outperform an oversized AVIF simply because fewer pixels need to be delivered.

Why tiny files can still look bad

A small file size is not automatically a success. Heavy compression may create block patterns, halos, banding, muddy texture, or color changes. The correct target is the smallest file that still looks appropriate for the intended use.

The viewing context matters. A thumbnail can tolerate more compression than a portfolio hero image. An image displayed on a high-density screen may need more detail than a small email preview. Technical optimization should follow the real user experience.

Why converting to PNG does not restore lost detail

If a JPG already contains compression artifacts, converting it to PNG will store those current pixels losslessly—but it cannot reconstruct the original detail that was discarded when the JPG was created. The PNG may become much larger while looking almost identical to the compressed source.

The useful reason to switch from a lossy source to a lossless working format is to avoid further lossy saves during editing, not to reverse past compression.

A practical quality workflow

• Keep the highest-quality original or master file.
• Resize a copy to the dimensions required by the destination.
• Choose a format based on content and compatibility.
• Start with a reasonably high quality setting for lossy output.
• Inspect fine details, edges, gradients, and text.
• Lower quality only until the file size improves without unacceptable visual damage.
• Avoid repeatedly recompressing an already compressed delivery file.