1. Home
  2. Guides
  3. Lossy vs Lossless

Lossy vs lossless:
what really happens.

One throws data away forever; the other just packs it tighter. Knowing which to use — and when — is most of image optimization.

100% private — no uploads Free forever Batch & ZIP Works offline

Two fundamentally different ideas

Lossless compression is packing: it finds repetition and stores it more cleverly, like writing "50 blue pixels" instead of listing blue fifty times. Unpack it and you get every original pixel back, bit-for-bit. PNG works this way. So does ZIP.

Lossy compression is triage: it ranks image information by how visible it is to human eyes and permanently discards the least visible part. JPG, lossy WebP, AVIF and HEIC all do this. The discarded data is unrecoverable — and that's fine, because it was chosen precisely for being imperceptible.

Lossless (PNG, WebP-lossless)Lossy (JPG, WebP, AVIF, HEIC)
Pixels after decodeIdentical to originalApproximation (tuned to look identical)
Typical photo savings10–30%60–95%
Quality dialNone — always perfectYes — you choose the trade-off
Repeated re-savingSafe foreverDegrades each generation
Best forMasters, graphics, text, archivesPhotos being viewed or shared

Why throwing data away is usually right

Human vision has well-mapped blind spots: we resolve brightness far better than color, notice smooth areas more than busy ones, and can't perceive high-frequency texture beyond a threshold. Lossy encoders are built directly on those blind spots — JPG, for example, stores color at half resolution before doing anything else, and you have never noticed.

That is why a 6 MB photo can become 500 KB with no visible change: the discarded 92% was mostly imperceptible detail and sensor noise. Keeping it has a real cost (bandwidth, storage, load time) and zero visible benefit.

The trap: generation loss

The one real danger of lossy formats is re-encoding the same image repeatedly. Each save discards a little more; ten rounds of edit-save-edit-save produce visible mush. The workflow that avoids it:

PixSquash never touches your originals — it produces new files, so your master stays pristine no matter how many variants you export. Try it: compress a JPG at different quality levels and compare.

Which should you use?

Photos being shared or displayed → lossy, quality 80–85. Logos, screenshots, text → lossless (or high-quality WebP). Archival masters → lossless, always. When a hard size limit rules (forms, portals), lossy is the only tool that can hit an exact number — that's what our compress-to-size presets automate.

FAQ

Frequently asked questions

Is lossy or lossless compression better?

Neither — they solve different problems. Lossy is right for photos being viewed (huge savings, invisible cost); lossless is right for masters, graphics and anything you will edit again.

Can I recover quality lost to compression?

No. Discarded data is gone permanently — AI upscalers can hallucinate plausible detail but cannot restore the original. This is why keeping a master copy matters.

Is PNG always lossless?

Yes, standard PNG is always lossless. (Some tools quantize colors before saving PNG, which loses data before encoding — but the PNG format itself is lossless.)

How many times can I re-save a JPG?

Once or twice is harmless; visible degradation accumulates over repeated saves. Edit from an original and export once rather than re-saving exports.

Is WebP lossy or lossless?

Both — it has separate lossy and lossless modes. Lossy WebP replaces JPG; lossless WebP replaces PNG at smaller sizes.

Keep going

Related tools & guides