Forcing Acrobat to use XRef tables when saving/optimising PDFs

We have a workflow where I work of editing PDF files (adding bookmarks, changing default view settings, etc.) and then optimising the PDFs before deploying them for use.

Once deployed to the server, the PDFs are encrypted on-the-fly when downloaded, but the encryption program that does this is… well, quite old. It doesn’t understand XRef streams, for example, so if the PDF was created using XRef streams, the encryption just results in a corrupted file being downloaded. Works perfectly fine if the PDF was created using XRef tables, though.

When I export PDFs from InDesign, they use tables if I set compatibility to an early enough version (which makes sense since streams were introduced in… 1.4? 1.5?). This works fine. If I optimise such a file in Acrobat, it works perfectly fine, and the encryptor reads the file no problems.

Unfortunately, many of the PDFs I have are onees which I haven’t created myself, and they tend to have fonts embedded as OpenType, which requires a later version – and that’s when Acrobat then defaults to using streams instead of tables. In other words: if I try to optimise such a file to version 4.0, the optimisation fails because 4.0 does not support OpenType embedding; if I optimise to a later format, it uses streams and the encryptor won’t work.

Is there some way in Acrobat to optimise a file for an OpenType-compatible version of PDF, but still force it to use XRef tables instead of XRef streams to create the file?