How is the difference between the two modes stored in the file?

WEBP houses two different encoding methods under a single extension; which method was used is recorded inside the file and the viewer decodes accordingly. From the user's point of view, because both share the same extension, there is no way to tell which mode a file was produced in just by looking at it. This is as much WEBP's most confusing aspect as it is its most practical feature. When you receive a WEBP file, you can only tell whether its content was encoded lossily or losslessly by knowing how it was produced.

Does converting to lossless mode bring back quality that was already lost?

It does not. Lossless mode preserves the picture exactly as it stands at the moment of conversion; it does not regenerate information discarded during some earlier compression. If the file you have came from a JPG and carries blockiness, lossless WEBP will store those flaws faithfully. Losslessness is not a repair feature but a freezing feature; it stops losses from here on, it does not erase the ones in the past.

Is transparency preserved in both modes?

Yes, WEBP supports the alpha channel in both lossless and lossy mode, and this is one of the most important features separating it from JPG. In lossy mode the alpha channel is handled separately, meaning transparency information can be stored independently of color information. In practice this means you can store a shadowed logo in a small file in lossy mode without losing its transparency. There is no equivalent to this scenario in JPG; transparent regions are permanently filled in.

Which mode should I choose if I am not sure?

If the image contains text that has to be read, thin lines or sharp-edged graphics, choose lossless; on this content the degradation lossy compression creates is most conspicuous. If the content is photographic, choose lossy; lossless mode produces very large files on photographs and offers no visible gain in return. When you are torn, exporting in both modes and comparing the file sizes and the visual result takes a few minutes and gives you the right answer with certainty.

Lossless WEBP or Lossy WEBP? When to Choose Which

One format, two different logics

The most important thing separating WEBP from other common image formats is that it houses two completely different compression approaches under a single extension. JPG is lossy only, PNG is lossless only; WEBP can do both.

This flexibility is a major advantage, but it brings a responsibility with it: choosing the right mode is now your decision. The wrong mode choice produces either needlessly enormous files or text degraded past the point of readability.

In this article we explain what the two modes do, which one gives the right result on which content, and how transparency is handled in each.

What does lossless mode do?

In lossless mode, no pixel information is discarded. When you open the file, you get the picture exactly as it was at the moment you saved it, bit for bit. Compression is achieved not by discarding data but by packing it more efficiently.

WEBP achieves this with more advanced methods than PNG. Instead of writing a region from scratch, it predicts it from the surrounding regions and stores only the difference; it also handles color palettes and repeating patterns more flexibly. As a result, on graphic and text-heavy content it usually produces a smaller file than PNG — with no quality loss at all.

Lossless mode's limit is photographs. In a photograph, almost every pixel differs from its neighbor; prediction does not hold, there is no repetition to pack, and the file stays very large. This is not a flaw in the format but the nature of losslessness.

What does lossy mode do?

In lossy mode, WEBP selectively discards information the eye struggles to notice. It predicts part of the image from a similar part processed earlier and encodes the remaining difference with coarse rounding. The discarded information is not stored in the file and cannot be brought back.

This method is extremely effective on photographic content. Smooth color transitions, natural textures and complex scenes can be stored in far smaller files in a way that is visually indistinguishable at a reasonable quality setting. The reason lossy WEBP usually comes out smaller than JPG at the same apparent quality is precisely that this prediction logic is more advanced than JPG's.

Lossy mode's weak point is sharp transitions. Small-point text, thin lines, sharp-edged icons — these are the structures compression struggles with most, and at low quality settings you get halos around letterforms and softening on lines.

Transparency: present in both modes

One of WEBP's most valuable features is that it supports the alpha channel in both modes. This offers a combination that has no equivalent in other formats.

In JPG there is no transparency at all; transparent regions are permanently filled with a solid color. In PNG transparency exists, but because it works only losslessly, files get very large on photographic content. WEBP, by contrast, can store a shadowed logo or a product photo with the background removed in lossy mode — that is, in a small file — while preserving its transparency.

This resolves scenarios that were long stuck in the dilemma of "if I want transparency, I have to accept a large file."

Which one for which content?

Screenshots, interface images, documentation: Lossless. Interface text legibility is critical, and lossless mode already works efficiently on this content type.

Photographs, product shots, cover images: Lossy. Lossless mode produces very large files here and offers no visible gain in return.

Flat-color logos and icons: Lossless. Sharp edges must be preserved exactly, and on top of that the file stays small on this content anyway.

Shadowed, gradient logos: It depends. Lossless is the safe choice; if file size is a problem, lossy can be tried at a high quality setting, but the output must always be checked at magnification.

Infographics and diagrams: Lossless. On text- and line-heavy content, legibility takes priority.

Images mixing photography and text: Look at the dominant character. If photography covers most of the area, lossy; if flat background and text cover it, lossless.

If you cannot decide

The most reliable method is to try. Export from the same source in both modes, compare the file sizes and examine both at one hundred percent zoom. If you can see no degradation in the lossy output and the file is markedly smaller, lossy is the right choice. If you can see degradation or the size gain is small, stay lossless.

While making this comparison, watch for one thing: if your source file already comes from a lossy format, lossy WEBP adds a second round of compression. At reasonable settings this stays invisible, but the degradations of the two rounds can stack. Whenever possible, make sure your source is lossless when you convert PNG to WEBP.

In summary

WEBP's two modes exist for different jobs. Lossless mode discards no pixel information and usually gives a smaller file than PNG on graphic and text content; this is where screenshots, logos and diagrams belong. Lossy mode selectively discards information and is usually more efficient than JPG on photographic content. Transparency is preserved in both modes — this is what makes WEBP unique. When you cannot decide, exporting both and comparing takes a few minutes and gives a far more reliable answer than guessing.