[skandium]

This is my field as well, although I come from the neural network angle.

Learned video codecs definitely do look promising: Microsoft's DCVC-FM (https://github.com/microsoft/DCVC) beats H.267 in BD-rate. Another benefit of the learned approach is being able to run on soon commodity NPUs, without special hardware accommodation requirements.

In the CLIC challenge, hybrid codecs (traditional + learned components) are so far the best, so that has been a letdown for pure end to end learned codecs, agree. But something like H.267 is currently not cheap to run either.

[zbobet2012]

Winning in bd rate though isn't hard. You need to win in bd rate and have a hardware implementable, power efficient, cheap decoder.

Agreed hybrid presents real opportunity.

[AzzyHN]

Did you mean H.266? Or is there some secret H.267 that hasn't been agreed upon yet