“Lossless” and “lossy” are two of the most commonly used but least clearly explained terms in image and document formats. PNG is lossless, JPG is lossy, and most explanations stop there, as if that alone tells you what to actually do. Here is what the distinction really means, in practice, for the files you work with every day.
The core idea in one sentence
Lossless means a file can be compressed and then perfectly restored to exactly its original data, with nothing discarded; lossy means some data is permanently discarded to achieve a smaller file, and that data cannot be recovered afterwards, even if you decompress the result. Both approaches are legitimate and useful — they are simply solving the size problem in two fundamentally different ways.
Why lossy compression can still look identical to the eye
The reason lossy formats like JPG remain so widely used, despite discarding data, is that they are specifically designed to discard the kind of detail the human eye is least likely to notice — extremely subtle colour variations, fine texture in areas the eye tends to skim over. At a sensible compression setting, the discarded information is genuinely imperceptible in normal viewing, even though it is technically, permanently gone from the file. This is why a well-compressed JPG can look identical to its source despite being a fraction of the size.
Why lossless matters for some content and not others
- Text and sharp lines suffer visibly under lossy compression, since the eye is very sensitive to the crispness of an edge or a letterform — this is why screenshots and scanned documents can look fuzzy if compressed too aggressively.
- Photographs are the ideal case for lossy compression, since they are full of exactly the kind of continuous tone and fine detail that lossy formats are built to simplify with minimal visible cost.
- Anything you will edit repeatedly benefits from staying lossless throughout editing, since each lossy save discards a little more data, and those losses compound with every additional save.
The compounding effect that makes repeated lossy saves risky
A single lossy compression pass at a reasonable setting is usually harmless to look at. The risk comes from repeating it: each additional lossy save works from an already-degraded copy, discarding a little more real information each time, and the effects accumulate visibly after enough passes. This is covered in more detail in our guide on why recompressing a JPEG doesn't help — the underlying reason is exactly this lossy, cumulative data loss.
How this applies to formats beyond just images
PDF compression, as covered throughout this blog, generally works by applying lossy compression specifically to the embedded images within a document, while leaving the actual text layer completely untouched — text in a PDF is not stored as pixels the way an image is, so there is no lossy/lossless distinction to apply to it the same way. This is exactly why a heavily compressed PDF can still have perfectly sharp, fully selectable text even while its photos have visibly softened.
A simple way to decide which approach you need
Ask whether the content is something you might need to reproduce or edit exactly later, or whether it only needs to look right to a viewer right now. A master copy of an important document, a logo you will keep editing, an archival scan — these benefit from staying lossless. A photo being shared once, a document being sent for someone to simply read — these are typically fine with lossy compression, since imperceptible visual quality is functionally identical to perfect quality for that purpose.
Why file size differs so much between the two approaches
Lossless compression can only shrink a file by finding and removing genuine redundancy without changing anything a viewer would perceive, which puts a hard ceiling on how much it can achieve for content that does not have much redundancy to begin with, like a detailed photograph. Lossy compression has no such ceiling in the same way, because it is willing to discard real information, not just redundancy — which is exactly why a lossy JPG of a photo is typically so much smaller than a lossless PNG of the same image, often by several times over.
Applying this the next time you have to choose a format
Rather than memorising which specific formats are lossy or lossless, it helps to think about the underlying trade-off directly: are you willing to permanently discard some detail in exchange for a meaningfully smaller file, for content where that detail will not be missed? If yes, a lossy approach like JPG is the right tool. If the content specifically needs to stay exact — sharp text, a logo, something you will edit further — lossless is worth the larger file size.
Why this distinction rarely needs to change your everyday habits
Understanding the mechanism is useful, but it does not usually mean you need to overhaul how you work day to day. Most default choices — JPG for photos, PNG for graphics and screenshots — already reflect this trade-off sensibly, because they were designed with exactly this reasoning in mind. Where it does become worth actively thinking about is the handful of situations already covered elsewhere on this blog: choosing a lighter compression setting because a document will be read closely, or keeping a lossless master while distributing lossy working copies of the same content.
Frequently asked questions
Is PNG always a safe lossless choice?
Yes, PNG does not discard data when saved, which is exactly why it is the standard choice for content that needs to stay perfectly crisp, like screenshots, logos and line art.
Is WebP lossy or lossless?
WebP can actually do both, offering a lossless mode as well as a lossy one, which is part of why it has become a flexible, widely adopted format.
Does lossless mean the file cannot be compressed much at all?
It means compression is limited to removing genuine redundancy; some content compresses well losslessly (flat graphics, simple icons) while other content (detailed photographs) does not shrink much without lossy compression.
If I only compress a JPEG once, is the quality loss something I would actually notice?
At a reasonable compression setting, generally no — a single well-chosen lossy pass is designed to be imperceptible for typical photographic content.
Does converting a lossless file to a lossy format and back recover anything?
No, once lossy compression has discarded data, converting back to a lossless format afterwards only losslessly preserves whatever remains — it cannot restore what was already discarded.
Which should I use for a document scan — JPG or PNG?
JPG generally, for most everyday scanned text documents, since the size saving is considerable and text remains legible at a sensible compression level; reserve PNG for scans with fine detail that specifically needs to stay crisp.
Is compressing a file, either way, free with your tools?
Yes. No sign-up, no watermark, and files are removed from the server automatically about an hour after processing.