[dr1394]

H.264 has been getting a lot of press lately. Some

feel that it's the preferable codec to replace MPEG-2

in the future. See these links:


H.264 codec jeopardizes MPEG-4's ascendancy
http://www.eetimes.com/story/OEG20020920S0049


DVD quality over the Internet?
http://zdnet.com.com/2100-1104-961707.html


H.264 tutorial
http://www.vcodex.fsnet.co.uk/h264.html


I also heard a funny story (don't know if it's really

true) about a Microsoft employee who almost got fired

after writing and circulating an in-house paper that

compared H.264 to WM9. His analysis showed that H.264

is approximately 2dB better than WM9 with regards to

coded signal to original signal MSE (mean squared error)

ratio.


Ron

 

[trbarry]

It sounds like the technology is very nice.


But the wide spread implementation of MPEG-4 has been held up for years in the MPEG-LA licensing negotiations. Now with H26L becoming H264/AVC/MPEG-4,part 10 the MPEG-LA is controlling the process again with a current call for AVC patents.


And I worry that the process of making it into a flavor of MPEG-4 will be muddied by attempts to make it depend upon other intellectual property by the current MPEG-4 IP holders.


This could take a while.


- Tom

 

[JStanton]

What are the potential uses for this?


- Is there any chance of the DVDForum approving this for HD-DVD?

- Could the DBS providers switch to this to alleviate their bandwidth constraints?

- When Sinclair gets their wish and the FCC reviews ATSC, could this be the new encoding method for the new transmission standard (and can we get broadcast DTS too)?


(Only kidding on the last one, actually curious about the other two.)


- Jim

 

[jamoka]

I would hate to be the one going up against Microsoft with


1. A ~1 year WM9 running start including free distribution of encoder and player


2. Only a 2% quality advantage, and


3. Significantly steeper (more expensive) hardware requirements


Marketing, politics, and wheelin' dealin' will probably decide the ultimate winner. Here, Microsoft is in another league compared to the bungling MPEG4 effort.


Joe

 

[jsgoldbe]

Quote:

Originally posted by jamoka I would hate to be the one going up against Microsoft with [...] 2. Only a 2% quality advantage, and

Actually it was posted as a 2 dB advantage, which is about a 60% improvement.


HTH

 

[Dean Roddey]

Hmmm... the plot thickens, and it was already opaque to begin with I'm all for the technically superior solution, but as mentioned there will probably be large style politics involved when it means that someone is going to own the next generation codec technology. Personally, I think that MS' technology would have to be the one that's 60% better in order to justify allowing them to extend their reach in such an important way.

 

[jamoka]

Whoa. 60% better is another story altogether. Next thing you know, we'll be getting HD movies on a CDR.


Joe

 

[dr1394]

Some nice things about H.264:


1) Still does well with interlaced formats. Aside from

the patent issues, poor performance on interlaced video

is one of the major roadblocks to the acceptance of

"regular" MPEG-4 for broadcast.


2) Will have specific profiles for SD and HD resolutions.

Avoids the confusion of the MPEG-4 "toolkit" concept.


3) The old-school MPEG-2 crowd (which is a lot of fairly

important people) that scoffed at MPEG-4, love H.264 since

it solves an actual problem (bandwidth limited MPEG-2

applications).


4) It's not from Microsoft


Disadvantages:


1) A fairly complex decoder and a crazy complex encoder.


2) Inclusion into MPEG-4 chapter 10. As Tom posted, it just

gets in the way.


3) Still a ways off to a finished specification and real

implementations.


Ron

 

[cymro]

FYI, a 2db SNR advantage translates to a Bit Error Rate (BER) improvement of something in the range of 1000:1.


Communication engineers kill for even 0.5 db SNR improvements.