What changes in the file
- Pages become separate files
- A multi-page PDF exports as `name-1.png`, `name-2.png`, `name-3.png` and so on, into a folder named after the document, every page is numbered, including the first. A single-page PDF is written as one plain file, with no folder. The conversion reports how many pages it produced rather than leaving you to count.
- Text stops being text
- Vector glyphs are rasterised. The output is no longer searchable, selectable or reflowable, and it will not be picked up by a text index. This is the fundamental trade of the conversion.
- Vectors get a fixed resolution
- A PDF's line art is resolution-independent; a PNG is not. Once rasterised, zooming in shows pixels rather than sharper curves.
- Interactivity is dropped
- Links, form fields, annotations, layers and embedded attachments have no representation in a PNG and do not survive.
Why the 100-page cap exists
This is a real limit and it is stated on purpose rather than discovered at run time.
PDF rendering happens in memory: every requested page is rasterised before the writing starts. At the 2x viewport scale PixelFerry uses (roughly 144 dpi) a 500-page scanned document is gigabytes of bitmap, and the honest outcome of attempting it is that the application dies partway through and you get an incomplete folder with no explanation.
So the conversion stops at the first 100 pages and tells you it did. A truncated result you know about is recoverable; one you do not know about is a bug you find weeks later.
If you need more, split the PDF first. Preview’s sidebar will do it by dragging page ranges out into new documents.
PNG or JPG for pages?
It depends on what is on them.
PNG for anything that is mostly text, line art, tables or screenshots. The lossless compression keeps glyph edges sharp, and documents of that kind compress well because they are largely flat colour.
JPG for scanned photographs and image-heavy brochures, where PNG produces much larger files for no visible benefit. Scanned photographic pages are close to worst case for lossless compression.
If the document mixes both (a report with photographic plates) PNG is the safer default, because degraded body text is more obviously wrong than a slightly softer photograph.
Where the output goes
A multi-page PDF gets its own folder, named after the document; a single-page PDF is written as one plain file. The folder is deliberate, a batch of twenty PDFs at twenty pages each would otherwise put four hundred loose PNGs into one directory with names that collide the moment two documents both have a page 3.
Inside, pages are numbered in order. The numbers are not zero-padded, so Finder
sorts them naturally but a plain lexicographic listing (ls, a glob, most
readdir sorts) will put name-10.png before name-2.png. Worth knowing if
you are feeding them to a script.
Resolution
Set a width if the pages are headed somewhere specific. PixelFerry rasterises at a fixed 2x viewport scale, roughly 144 dpi, not user-configurable, which is still far wider than a wiki thumbnail or a slide needs, and it costs real disk space across a hundred pages.
Resizing during the conversion rather than after it means the large intermediate never touches your disk.
When people do this
- Producing page thumbnails of a document set for a CMS, a wiki, or an internal review tool.
- Extracting scanned pages as images so they can be cropped, straightened or run through OCR elsewhere.
- Getting a single page of a report into a slide deck or an email without attaching the whole document.
What macOS already does
Preview's export, or sips.Preview can export a PDF to PNG, and in the sidebar you can select specific pages first. `sips -s format png doc.pdf --out page.png` handles the first page from Terminal.
Where it stops being enough:You have several PDFs to do at once. Preview works document by document, so applying the same output width across a set of them is manual.
Limitations worth knowing first
- Long documents are capped at the first 100 pages. Rendering is held in memory, and an uncapped 500-page scan would take the app down, so the conversion says it stopped rather than silently truncating.
- PDF is input-only. PixelFerry does not assemble images back into a PDF.
- Password-protected and damaged PDFs will not render, and are reported as errors on their own row rather than failing the batch.
PixelFerry requires macOS 14 (Sonoma) or later on Apple silicon and Intel. Conversion runs entirely on your Mac. No file, and no part of one, is uploaded anywhere.