Lossy Compression
A compression method that permanently discards some data to achieve smaller file sizes — the original cannot be exactly reconstructed.
Last updated
What Is Lossy Compression?
Lossy compression permanently removes data from a file to achieve greater size reduction than lossless methods allow. The decompressed result is an approximation of the original, not an exact copy. The amount of data removed is controlled by a quality setting.
How It Works
Lossy algorithms exploit limitations of human perception:
JPEG: Uses Discrete Cosine Transform (DCT) to identify high-frequency detail (fine texture) that the human eye is less sensitive to, then discards that detail based on the quality setting.
MP3/AAC (audio): Uses psychoacoustic models to identify sounds masked by louder simultaneous sounds, then discards the masked data.
H.264/H.265 (video): Encodes differences between frames rather than full frames, and applies lossy compression to each frame similar to JPEG.
The Generation Loss Problem
Each time a lossy file is opened and re-saved, the compression algorithm runs again on the already-compressed data. Each cycle introduces additional loss. After multiple cycles, visible artefacts accumulate.
Rule: Never use a lossy format as an intermediate working copy. Edit from the original lossless source, export to lossy only for final delivery.
Quality Settings
All lossy formats offer a quality control:
| Setting | Description |
|---|---|
| 95–100 | Near-lossless; minimal visible artefacts |
| 80–85 | Recommended for most uses; imperceptible loss |
| 70–75 | Smaller file; artefacts visible at high zoom |
| Below 70 | Significant artefacts; acceptable only for thumbnails |