Lossy vs Lossless Compression: What It Means for Images, PDFs and Video
Every time you shrink a photo, a PDF or a video, a compression method is at work. Compression methods come in two families: lossless, which can be reversed perfectly, and lossy, which permanently throws away some data. Understanding the difference helps you choose settings that give small files without visible damage.
Lossless compression
Lossless compression finds patterns and stores the same information more efficiently, the way a ZIP file does. When you decompress, you get back exactly what you started with, bit for bit. PNG images and lossless WebP use this approach, as do ZIP archives. File size reductions are modest, typically 20-50% for images with flat colors, and small for noisy photographs.
Lossy compression
Lossy compression removes information that people are unlikely to notice: subtle color differences, very fine detail, sounds the ear can barely hear. JPG, lossy WebP, MP3 and most video codecs work this way. The savings are dramatic, often 80-95% for photos, but the removed data cannot be recovered.
What happens when quality is set too low
- Blocky squares around edges and in smooth gradients.
- Smudged text and "mosquito noise" around letters.
- Color banding in skies and shadows.
Generation loss
Each time you open a JPG, edit it and save it as JPG again, you apply another round of lossy compression, and the damage accumulates. Keep an original or edit a lossless copy, and export to JPG once at the end.
Choosing the right type
| Content | Recommended |
|---|---|
| Photographs for sharing | Lossy (JPG or WebP at about 80-90%) |
| Logos, icons, screenshots | Lossless (PNG) |
| Images you will edit again | Lossless master, lossy export |
| Video for sharing | Lossy (H.264 MP4) |
| Documents with text | Lossless or high-quality (avoid heavy compression) |
Where each format fits
Our guide PNG vs JPG vs WebP compares formats, and what is WebP explains a format that supports both modes.
How much quality do you need?
For most photos, a JPG quality between 75% and 90% is visually indistinguishable from the original at normal viewing sizes, while the file is a fraction of the size. Going below 60% usually shows artifacts. The image compressor offers four levels, and you can hit a precise file size with the exact size tool; see how to compress an image to 100KB for practical examples.
PDFs and video
PDFs combine lossless text with lossy images; compressing scanned PDFs mostly means lowering image quality, as in how to reduce PDF file size. Video is almost always lossy and benefits from choosing sensible resolution and bitrate, as in how to compress a video.
How JPEG compression works, in plain language
A JPG encoder first separates brightness from color. Because the eye is far more sensitive to brightness than to color, it can store color at lower detail without you noticing; this is called chroma subsampling. The picture is then divided into small blocks, and each block is converted into a set of patterns of increasing detail. The encoder keeps the broad patterns accurately and rounds away the fine, high-frequency ones. The "quality" slider controls how aggressive that rounding is. At moderate settings the discarded information is invisible; at low settings the blocks become visible as squares.
Why the same quality setting gives different file sizes
A photo of a clear sky has few details and compresses to a tiny file. A photo of foliage, fabric or a crowd contains fine detail everywhere and stays large even at the same quality number. That is why a quality number is not a promise about size, and why a target-size tool tests several settings instead of using a fixed one.
A sensible workflow for photos and graphics
- Keep a master copy in the original or a lossless format.
- Do all edits on a copy.
- Export once to the delivery format at a quality of about 80-90%.
- Never re-save a delivery JPG repeatedly; go back to the master instead.
The same idea in audio and video
MP3 and AAC audio and every mainstream video format work the same way: they discard what people are least likely to perceive. Lossless equivalents such as FLAC for audio exist for archiving. For video you will almost always deliver lossy files, so choose resolution and bitrate sensibly, as described in how to compress a video.
Quick reference
- Lossless: exact copy after decompression; ideal for masters, logos and screenshots.
- Lossy: much smaller files; ideal for delivery of photos and video.
- Do not re-save lossy files again and again.
- For photos, 80-90% quality is usually the best compromise.
- Larger savings come first from resizing, then from quality reduction.
The best compression decision is rarely about one file; it is about a workflow. Keep high-quality originals, derive lighter copies for each purpose, and label them so you always know which is which.
Frequently asked questions
Is lossy compression always bad?
No. At sensible settings it is the reason the modern web works. The key is not to over-compress and not to recompress repeatedly.
Can I restore quality after lossy compression?
No. Once detail is discarded it cannot be recovered, so keep your originals.
The tool runs in your browser. No upload, no sign-up, no limits.
Compress your imagesRelated guides
PNG vs JPG vs WebP: Which Image Format Should You Use?
Not sure whether to save an image as PNG, JPG or WebP? This guide compares quality, file size, transparency and compatibility so you can pick the right format every time.
ImagesWhat Is WebP? Benefits, Drawbacks and How to Convert It
WebP explained: what the format is, how much smaller it is than JPG and PNG, where it works, why some sites reject it and how to convert WebP to JPG or JPG to WebP.
ImagesHow to Compress an Image to 100KB (or 50KB, 20KB)
Forms asking for a photo under 100KB, 50KB or 20KB? Learn how to compress a JPG to an exact file size quickly, plus the right pixel sizes for passport and ID photos.