This is a difficult question to answer, as elaborated below. What we 
have said is that Dirac is comparable to H264 in quality and, in 
principle, is simpler to implement (therefore lower processing 
requirements). Comparable in quality means that it has similar 
quality at similar bit rates and, therefore, requires similar disk space.

Do you mean how do the current implementations of codecs compare 
TODAY? Or do you mean, how would mature and equally well developed 
codecs compare in the future? I have found that many people who ask 
for comparisons are not clear which question they are asking.

It is, of course, reasonable to ask how the current implementations 
work today. However the results are not very meaningful. Dirac is 
less mature than H264. Consequently it is improving faster than H264. 
Even in the past few weeks there have been improvements in coding 
quality, particularly at low bit rates. Over the past few years there 
have been really dramatic falls in bit rate (for example the bit rate 
has halved over the past year). So, if you do a comparison today it 
will be out of date in a few weeks time.

The Dirac team publicly demonstrated Dirac at the International 
Broadcasting Convention (IBC) in Amsterdam in September. We were 
showing HDTV at 6Mbit/s. Our demos where more than 10 minutes long 
and IBC had about 50,000 visitors from the broadcasting industry. 
There was nowhere to hide in terms of picture quality! But our 
picture quality certainly stood up to scrutiny and in my (admittedly 
biased) opinion the picture quality was better than the vast majority 
of H264 pictures on show at the same bit rate.

So, at the moment, Dirac compression quality is comparable to the 
best H264 quality and is improving. For some content Dirac is better 
than H264. For some material it will be worse. But Dirac is improving 
more rapidly than H264.

Similarly it is not very meaningful to ask about processing 
requirements. The Schro decoder has reasonable performance in terms 
of processing requirements. But we know, even following work over the 
past few weeks, that there are quite a few ways we can optimise 
processing. At the moment H264 probably has lower processing 
requirements. But the situation is fluid and, as for compression 
efficiency, Dirac is likely to improve faster than H264. Also 
consider whether processing in the CPU versus the GPU is an issue. At 
the moment Scho runs in the CPU. In the future (like some H264 
implementations) Dirac will run in the GPU, which will dramatically 
improve performance.

So at the moment I would say Dirac compression (using the Dirac 
research coder) is comparable to H264. The Schro encoder is not quite 
there yet but we are working on it an I expect it to reach similar 
levels of performance by the end of the year. In terms of processing 
requirements H264 is probably ahead at the moment. But Dirac is 
probably improving more quickly than H264 so the situation is moving 
in favour of Dirac.

What of the more fundamental question about comparison of fully 
mature implementations of Dirac and H264? The Dirac specification may 
well allow Dirac to provide both better compression performance and 
lower processing requirements than H264 in the future.

Tim Borer

At 21:10 17/10/2008, Stas Oskin wrote:
>Hi.
>
>Are there any comparison details for how Dirac compares to MP4 (the 
>codec) and H.264 in terms of:
>
>a) Processing requirements
>b) Disk space requirements?
>
>Regards.
>-------------------------------------------------------------------------
>This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>Build the coolest Linux based applications with Moblin SDK & win great prizes
>Grand prize is a trip for two to an Open Source event anywhere in the world
>http://moblin-contest.org/redirect.php?banner_id=100&url=/
>_______________________________________________
>Schrodinger-devel mailing list
>[email protected]
>https://lists.sourceforge.net/lists/listinfo/schrodinger-devel


-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Schrodinger-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/schrodinger-devel

Reply via email to