Potential Encoding Option

Hi all,

I'm new to the list, and I haven't read the entirety of the archives 
yet.  But I think I have an idea of how this provisional codec is put 
together.  So I already have a few ideas; so I'll just fire one off:

This is a video codec right?  So individual frames can probably take 
a back seat to the general gist of the video.  There's little use in 
worrying about specific details of an object if it's only going to be 
there for a single frame.  With this in mind, it would be beneficial 
to kill some detail in the individual frames if that detail won't be 
appreciated by a human observer.

Thus, I propose a tack-on filter for the encoder to take advantage of 
this fact.
The decoder wouldn't have to be changed at all; but the encoder would 
probably be much slower.  It would probably be easy to make this 
filter a run-time option.

The idea is this, some algorithm looks for areas that are 
substantially different from the surrounding frames; then after the 
wavelet transformation we clear/quantize/alter the high-pass 
coefficients corresponding to that area so that they'll compress 
better.  After running through the entropy coder that area would be 
blurrier, but that wouldn't matter since the more important areas 
could be much clearer(with a constant bitrate).

The actual implementation could vary pretty wildly and, of course, 
this filter wouldn't change the bitstream format so only advanced 
applications would need to implement it.  On a more general note, 
it's interesting to note that, unlike some other compression 
standards, a good Tarkin encoder could drastically change the quality 
of the bitstream.

Am I on to something or am I just raving?

	-Alex

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to '[email protected]'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.