What lossy compression removes
Encoders exploit how vision works. Contrast and color differences in busy areas are harder to see, so the algorithm simplifies those regions first and spends its bits on edges and faces that viewers notice.
The discarded data cannot be restored. Opening a lossy file, editing it, and saving it again applies the process a second time, so quality drifts downward with every round trip.
How settings affect the result
Higher quality settings keep more data and produce larger files, while lower settings push compression harder. The useful range is where the image still looks right at its final display size.
Dimension reduction often helps more than squeezing quality. A photo shown at 800 pixels wide does not need its original 4,000-pixel data, and trimming the excess shrinks the file sharply with little visible loss.
- JPEG photographs for the web and email
- WebP files when a smaller modern format is acceptable
- Video and audio, which rely on the same principles
- Any file where a set size ceiling must be met
Lossy compression in Filekind
The Filekind compressor works this way. It takes a JPEG or PNG and searches for a JPEG output at or below the KB or MB limit you set, lowering quality or dimensions as needed to reach it.
Because the output is JPEG, transparency in a PNG becomes white and fine detail may soften. The final size and dimensions are shown before downloading so you can judge the trade yourself.