Why doesn't moving to a lossless format improve quality?

Because losslessness is a promise about preserving data, not about regenerating it. When an AVIF is saved in lossy mode, some information is removed from the file entirely; that information no longer exists anywhere. When PNG opens the image, all it sees are the pixels that remain, and it stores them without discarding anything. In other words, PNG does not compensate for past loss, it only prevents future loss.

Are semi-transparent areas preserved too?

Yes. The alpha channel in AVIF and PNG does not offer a binary choice between fully transparent and fully opaque; it carries every intermediate value as well. This matters for soft shadows, glow effects and logos with anti-aliased edges. Those intermediate values are preserved during conversion, so the image's edges do not look jagged when placed on a dark background. JPG offers no such capability at all.

Can I make the PNG file smaller?

Somewhat, yes. PNG optimisation tools can shrink a file by reducing the colour palette or making compression parameters more aggressive. But palette reduction can create banding in smooth gradients, so you need to check the result visually. The deeper point is that on photographic content no amount of optimisation will bring PNG down to JPG or AVIF levels; if size is your main criterion, it would be better to reconsider your format choice from the start.

Why should I prefer PNG if I am going to edit the image?

Because in lossy formats every save adds a new generation of compression and degradation accumulates. When you open an image, crop it and save it, then open it again, colour-correct it and save it again, a lossy format loses a little more detail at every step. In PNG that accumulation does not happen; you could open and save the file a hundred times and the data would stay the same. That is why using a lossless format for the working file and producing the final delivery copy in a lossy format is the standard workflow.

What Really Happens to Transparency and Quality When You Convert AVIF to PNG?

A misunderstood conversion

"I converted my AVIF to PNG and the quality didn't improve" is the most frequently voiced disappointment about image conversion tools. The expectation is understandable: you have heard that PNG is lossless, and the word lossless sounds like "perfect quality".

The truth is a little more subtle. PNG is lossless, but that is a promise of preservation, not a promise of repair. In this article we explain exactly what is preserved in this conversion, what cannot be preserved, and what the real benefit of moving to PNG is.

What do lossy and lossless actually mean?

An image is a numeric table made up of pixels. There are two ways to write that table into a file.

The lossless approach preserves every number in the table. Compression is achieved by expressing repeating patterns more briefly; when you open the file you get the entire original table back. This is how PNG works.

The lossy approach deliberately alters the table. It calculates and discards the details the human visual system struggles to notice, and stores the rest at a coarser precision. When you open the file you get back not the original but a reconstruction that closely resembles it. This is how AVIF and JPG work.

The critical word here is "discards". Discarded data is not in the file. It is not stored anywhere, it does not sit there in compressed form — it has been deleted.

What happens when you go from AVIF to PNG?

The conversion has two stages.

First the AVIF is decoded. A pixel table is reconstructed from the compact description the encoder stored. That table is not the original photograph's table — it is a close approximation of it. Because AVIF is a very good codec, the approximation is usually good enough to be visually indistinguishable, but it is mathematically different.

Then that table is written out as PNG. PNG discards nothing; it saves the table it sees exactly as it is.

So what PNG stores is not the original photograph but the reconstruction AVIF produced. The softening AVIF left behind, the slight blurring in flat areas, the micro-distortions along sharp edges — all of these are now the image's actual pixels, and PNG preserves them faithfully.

This is not a flaw in PNG, it is precisely the job it was designed to do.

What about transparency?

Here the picture is far more positive. Transparency is the concrete thing you genuinely gain in this conversion.

AVIF supports the alpha channel. The alpha channel stores an opacity value for each pixel in addition to its colour information. That value is not limited to the pair "visible" and "invisible"; it can carry every intermediate tone between them. A soft shadow, a logo with anti-aliased edges, a lightly transparent glass effect — all of these are expressed through intermediate alpha values.

PNG has the same capability. So during conversion the alpha channel is carried across exactly. Transparent areas stay transparent, and semi-transparent areas keep the same degree of transparency.

For comparison: had you converted the same image to JPG, all transparent regions would have been filled with a solid colour, since JPG has no alpha channel at all. A logo would come out trapped inside a white rectangle. This is the soundest justification for choosing to convert AVIF to PNG.

PNG's real benefit: generation loss stops

If quality does not improve, what is the point of PNG? The point lies in the future.

The insidious thing about working in a lossy format is that degradation accumulates. If you open an AVIF, crop it and save it as AVIF again, compression is applied afresh. If you open it the next day, add some text and save again, it is applied a third time. Each round leaves small damage, but the damage stacks until it becomes visible.

PNG breaks that chain. Because it is lossless, it makes no difference how many times you open and save the file; the data stays fixed.

That is why the professional workflow is clear: lossless working file, lossy delivery file. If you are going to work on an image, bring it into PNG; when you are done, output the final version in a lossy format suited to the web.

The cost in size

This trade has a price and it should not be underestimated.

Because PNG discards no information, it produces very large files for photographic content. Photographs contain almost no repeating patterns, so lossless compression has nothing to hold on to. The difference between the AVIF and PNG versions of a landscape photo can be several times over.

By contrast, PNG is quite efficient with graphics containing flat colour areas. A logo, an icon set or an interface screenshot stays at a perfectly reasonable size in PNG — because repeating patterns are plentiful in this kind of content.

Practical rule: graphic content loves PNG, photographic content does not.

Three questions to decide

Before converting, ask yourself:

Do I need transparency? If yes, PNG is mandatory and the discussion is over.
Am I going to edit the file and save it several times? If yes, PNG makes sense.
Is the content a photograph or a graphic? If it is a graphic, PNG is both correct and efficient. If it is a photograph and you answered no to the first two questions, PNG may be an unnecessarily large solution.

In summary

Converting AVIF to PNG is not a repair operation, it is a freezing operation. The image's current state — warts, artefacts and all — is placed into a lossless container and will not degrade from here on. Transparency is preserved completely, and in most scenarios that is the real justification for the conversion. In exchange, the file grows noticeably. Done with the aim of securing existing quality rather than the expectation of restoring it, this conversion is exactly the right call.