with the BBC's reputation in complete tatters
If that's re: Hutton, I really don't know a single soul who thinks the Hutton report an accurate report of events. The man had to be either gullible, senile, or open to bungs in a brown paper envelope in a remote lay by.
I think they use OBMC... at least it implies that in the readme
"Motion compensation uses overlapped blocks to reduce block artefacts that
would upset the transform coding stage"
My own experience is that OBMC is better than deblocking (well it ought to be!) but isn't worth the additional complexity over a good deblocker.
That's true for still images, but video still requires motion compensation - which is generally block based and causes block artifact with both wavelets and DCTs. In fact it can be worse with wavelets if you don't know what you're doing...
I did a psychovisual evaluation of still image compression with wavelets vs DCTs for my PhD a few years back... most people preferred the 'rain on the lens' degredation of wavelets, but a significant number still preferred the old DCT for specific images.
Slashdot Mirror
with the BBC's reputation in complete tatters If that's re: Hutton, I really don't know a single soul who thinks the Hutton report an accurate report of events. The man had to be either gullible, senile, or open to bungs in a brown paper envelope in a remote lay by.
I think they use OBMC... at least it implies that in the readme "Motion compensation uses overlapped blocks to reduce block artefacts that would upset the transform coding stage" My own experience is that OBMC is better than deblocking (well it ought to be!) but isn't worth the additional complexity over a good deblocker.
That's true for still images, but video still requires motion compensation - which is generally block based and causes block artifact with both wavelets and DCTs. In fact it can be worse with wavelets if you don't know what you're doing... I did a psychovisual evaluation of still image compression with wavelets vs DCTs for my PhD a few years back... most people preferred the 'rain on the lens' degredation of wavelets, but a significant number still preferred the old DCT for specific images.