[SUGGESTION] Go neural

[MarcoRavich]

Hi there,
it's cool to hear that the project isn't adbandoned yet.

Bytheway, in order to be truly competitive on the scene, I believe it needs a serious modernization: since they are very fashionable nowadays, I would therefore suggest approaching to neural networks (perhaps by establishing some kind of collaborations, for example with @itu-lab - @Aeunal):

https://github.com/itu-lab/CorticalCodingCodec#readme

Hope that inspires !

[petterreinholdtsen]

This seem like a proposal to send upstream, not at the github git repo copy, ie https://gitlab.xiph.org/. I have no idea what "serious modernization" mean, and suggest you explain in more practical terms what you mean by this if you report it upstream.

Just to be clear, given that VP8, VP9 and AV1 exist, I doubt anyone is going to spend much time improving the compression and decompressions designs of theora. The release I plan for the coming weekend is to fix security and build issues, to ensure the library is still usable with the current set of compilers and libraries.