Ben, thank you for the history. I have survived most of it myself.

You seem to have thrown away the synergy of hyperlinking across multiple documents.
With Affinity, you will need to 'publish' every document and then work out what its link path will be.
Both Xara and Serif (noting your typo) handle internal links adequately but external ones require a collection area (c.f., the Internet).
With Obsidian (and MarkDown) the connectivity is down to the folder location (even across Vaults).

Using images and PDFs squander their information value. Most file formats wither. .txt (a.k.a., .md), UTF-8 et al, ASCII & extended and system flavours of EBCDIC all have simple (Regex) convertors, which I have had to use all of my career(s). Anything proprietary and you are out of the picture.

You were rather emphatic that your images were in an external file store.
In Obsidian, I can quickly link to any image with ![[...]]. If it is a mix of text, vector and graphic then I can link to its .XAR (or other file extension with [File](file:///I:\TG\2023.xar), edit that and export it as an image back to the required location.

Why stop at pictures?
With Boxy SVG, which I use, I can add in images, vectors, text and animate them.
The SVG format payload is quite light too and readily embedded into a MarkDown file.

Acorn