Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Because it's a neural technique, not one based on pixels or frames.

https://blog.metaphysic.ai/what-is-neural-compression/

Instead of artifacts in pixels, you'll see artifacts in larger features.

https://arxiv.org/abs/2412.11379

Look at figure 5 and beyond.





Like a visual version of psychoacoustic compression. Neat. Thanks for sharing.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: