Reading question on SO such as this one, makes me wonder how Chrome can support PDF ≥ 1.7 (which can contains JPEG 2000 codestream), while at the same time claims that it does not support JPEG 2000:
What is the trick here ?
It's because PDFs in Chrome are rendered with PDFium, which does support JPEG2000 while HTML is rendered using Blink, which does not.
The issue has already been discussed here:
In particular comment #4 makes the following observation:
Chrome already has JPEG2000 support and includes openjpeg library for its PDF reader and has to deal with its fuzzing/security issues anyway so the actually supporting it like Safari is not opening a new attack surface and doesn't cost binary size also I'd guess.
Same goes here:
PDF is not part of the standard web platform, it is a plugin, so it is out of scope. However, I agree with you that it seems silly to support this in one context and not in another.
Full thread on blink-dev: