If both formats are lossless, why isn't the conversion completely harmless?

Because losslessness only guarantees that the pixel data is not damaged during compression, not that every piece of information the format can carry will be preserved. If TIFF holds a feature PNG does not support, that feature is lost in the conversion — not because of compression, but because there is no equivalent in the target format. Multi-page support, the CMYK colour space and some embedded metadata fall into this category. Look at the pixels and nothing has been damaged; look at the file as a whole and there is a loss of capability.

Why does my PNG file sometimes come out larger than the TIFF?

This usually stems from the source TIFF already being compressed. TIFF can be stored uncompressed but it can also be compressed with lossless algorithms such as LZW or ZIP; in the latter case the file is already small and PNG retains no clear advantage. In addition, with photographic content there is limited ground for lossless compression to gain, because natural texture and continuous colour transitions cannot be encoded easily. In certain special cases, for example with particular colour layouts, the PNG output can even be larger than the source.

If I convert a 16-bit TIFF to PNG, does my colour depth drop?

PNG technically supports 16 bits of data per channel, so there is no obstacle as far as the format goes. In practice, however, a great many conversion tools and workflows reduce the output to 8 bits, because 8 bits is the standard for the web and general use. That reduction can lead to banding in transitions in images that will undergo heavy colour correction. If you need high colour depth, it is safer to check the file's bit depth after conversion and to do editing work on the original TIFF.

Can I keep my archive as PNG instead of TIFF?

It is technically possible because PNG is lossless and an open standard, but in corporate preservation practice it cannot fully take TIFF's place. Libraries, museums and archival institutions prefer TIFF because multi-page support, CMYK, rich metadata support and high colour depth come together in a single container. For personal use or a web-oriented archive, PNG is a perfectly reasonable choice. But for material connected to the print chain or requiring long-term corporate preservation, staying with TIFF is more appropriate.

TIFF to PNG Conversion: What Changes When You Go from Lossless to Lossless?

How much does the word "lossless" cover?

Both TIFF and PNG are lossless formats. Looked at with that in mind, the conversion appears completely harmless: you are moving from one lossless box to another lossless box, and the contents should stay the same.

Largely true — but not exactly. The term "lossless" only says that the compression algorithm discards no data. What kinds of information the format can carry is an entirely separate matter. Some of the things TIFF can carry have no equivalent in PNG, and they are quietly lost in the conversion.

In this article we explain exactly what is preserved, what is lost, and what that means for your work.

Preserved: the pixel data

Let us start with the most important thing. Moving from TIFF to PNG, the image's colour values are preserved exactly.

What happens with a JPG conversion does not happen here: no block artefacts form, no ringing appears around sharp edges, no fine texture is erased. Open and save the file as PNG a hundred times and no degradation accumulates. That eliminates generation loss, the most insidious problem of lossy formats.

Transparency is preserved too. If the TIFF carries an alpha channel, PNG fully supports it — both fully transparent areas and the partial transparency of soft edge transitions are transferred exactly. This is the point that makes PNG decisively superior to JPG in this scenario.

Lost: the format's capabilities

Now to the main issue. TIFF is a far more capable container than PNG, and those extra capabilities go away in the conversion.

Multi-page support

TIFF can hold more than one image inside a single file. All the pages of a scanned contract, the entirety of a fax printout or a sequence of microfilm frames can sit in a single .tif file.

PNG cannot do this. Every PNG file is a single image. When a multi-page TIFF is converted one of two things happens: either only the first page is taken, or each page comes out as a separate PNG.

This is the trap people working with scanned documents fall into most often. The conversion looks successful, the file opens, but there is only the cover page. Know how many pages the source file contains before converting.

CMYK colour space

TIFF supports CMYK, the colour model of the print world, along with embedded ICC colour profiles. This is the basic requirement of the print workflow.

PNG works only with RGB. When you convert a CMYK TIFF prepared for print into a PNG, a colour space conversion takes place. That conversion cannot preserve colours exactly, because the two models can produce different colour ranges. A noticeable shift on screen occurs particularly in saturated reds, oranges and blues.

This is not a fault of the converter but an unavoidable situation arising from the nature of the two colour models. The practical consequence: a file going to print cannot be a PNG. Always give the printer the original TIFF (or the PDF/X file the printer asks for).

Colour depth

TIFF can carry 16 bits of data per channel or more. Where a standard 8-bit image has 256 tones per channel, 16-bit has tens of thousands. That difference is not directly visible to the eye but emerges under heavy colour correction: in an 8-bit image, smooth transitions such as a sky begin to band while a 16-bit file stays smooth.

PNG supports 16 bits as a format, so there is no theoretical obstacle. In practice, though, a great many conversion routes reduce the output to 8 bits. That is why, if you are going to do heavy editing, it is safer to carry out the work on the original TIFF.

Metadata and layers

Thanks to its rich tag structure, TIFF can carry detailed metadata such as scanning conditions, device information and archival records; in some cases it also holds layer information. PNG's metadata support is more limited and it has no concept of layers. In corporate archives this metadata is valuable for verifying the provenance of a document.

And what about file size?

The expectation should run this way: content type is decisive.

With images containing flat colour areas, sharp edges and few colours — logos, drawings, screenshots, clean black-and-white text scans — PNG's lossless compression works very efficiently and the file shrinks considerably compared with an uncompressed TIFF.

With photographic content, the gain is limited. Natural texture and continuous colour transitions cannot be encoded easily by lossless algorithms. In addition, if the source TIFF is already compressed with LZW or ZIP, the difference between them narrows sharply.

If size is critical for you and the content is pure photography, converting to JPG rather than PNG may make more sense — you get a far smaller file in exchange for accepting lossiness.

When is PNG the right choice?

There is transparency in the image
The content is a sharp-edged graphic, logo, drawing or screenshot
The image will be edited repeatedly and generation loss is unwanted
The file will be displayed on the web and TIFF will not open
A clean text scan on which OCR will be run

When should you stay with TIFF?

The file is going to print
A multi-page document must stay whole
Corporate archiving is being done
16-bit colour depth must be preserved

In summary

Converting from TIFF to PNG is genuinely lossless at the pixel level — that is the most valuable aspect of this conversion. But things do get lost, and they stem not from compression but from PNG being a simpler container: multi-page support, CMYK, usually colour depth, and rich metadata. If those losses do not affect your use case, PNG is an excellent destination. If they do, the solution is not to skip the conversion — it is to do the conversion and keep the original too.