[gst-devel] Re: Helix Player virtual team meeting

[Karl Lillevold]

Dear Rob,

I have to disagree in that there is little enthusiasm left for
RealVideo. If you would be so kind as to take a look in this forum:
http://forum.doom9.org/forumdisplay.php?s=&forumid=54 you will find a
lot of RealVideo enthusiasm, in fact I will dare to claim that RV9 is
currently the most popular New Codec, by far, among codec enthusiasts.

You will probably not believe our quality claims, so take a look in
this thread, one of the many independent codec comparisons that have
been posted in this forum:
http://forum.doom9.org/showthread.php?s=&threadid=64789
As you will see the current RV9-EHQ has the best quality among all the
codecs tested in terms of the objective measures. The codecs included
are RV9, WMV9, VP6, and two very popular MPEG-4 variants, XviD and
DivX.  Subjectively, of course, opinions differ greatly and each
codec has its own strength.  My point is that RV9 is highly
competitive, and is being improved every month, as you can see from
the news section in my 'sticky' RealVideo Information thread. Now
there is admittedly a lot of bad looking video on the Internet in
general. Much of this is older content, created with old codecs, and
using sub-standard encoding procedures and source material.

In this forum you will also find that one of the main reasons for this
popularity is that there are now many ways to play back RealVideo,
including a DirectShow wrapper on Windows, as well as advice on how to
make RealOne behave nicely. On Linux there is the Helix Player effort.
As a video codec engineer, as well as enthusiastic about video
playback on PCs, I need player choices. In fact, I use an alternative
player myself on a daily basis, since it communicates more directly
with the codec, and it is easier to isolate problems.

And before you send off a quick reply about the forums referenced
above, let me add that if it had not been for the useful feedback
there, many of the recent RealVideo improvements would not have
happened. I have attached a sample list below. There are participants
from all 4-5 leading codec research teams (RV9, WMV9 two MPEG-4
flavors, VP6), and independent tests of the same codecs, all of which
show that since RV9-EHQ became publicly available for testing, is
coming out on top in terms of compression efficiency.



# Sub-titles, RealText
# Native YV12 support in Helix Producer
# Motivation to work on RV9-EHQ and all the great feedback on its quality improvement.
# Duplicated frame dropping pre-filter (RV9 Animation DropDupe filter).
# Improved 2-pass VBR in RV9.
# Faster 1st pass when encoding RV9-EHQ.




Best regards,
 Karl Lillevold       mailto:karll at real.com
 Sr. Codec Engineer | RealNetworks Codec Group





Wednesday, December 10, 2003, 1:38:29 PM, Rob Adams wrote:
> On Wed, 2003-12-10 at 13:17, Thomas Vander Stichele wrote:
>> Looking at the actual situation, the real codecs are one
>> of the last to not have been reverse-engineered in some project.

> I should point out that window media and sorenson quicktime video
> haven't been reverse engineered by anyone -- rather wrappers have been
> written around the windows versions of the codecs to allow them to be
> played in mplayer or xine.  This works very well and so there's really
> no reason to bother reverse engineering the codec.

> I think that no effort has been applied to perform a similar effort on
> the real codecs because nobody uses realvideo anywhere unless they also
> have some other format.  This is mainly because the RealOne player can
> best be described as a trojan horse, slipping its foul tentacles into
> every aspect of the windows desktop, making it a decidedly unpleasant
> thing to deal with.  Combine that with the fact that RealVideo is
> substandard and you don't really have a lot of enthusiasm left for it.

> Now, if Real wants to try to resurrect itself with a new spirit of
> openness I applaud their efforts, but they've got a lot of ground to
> make up.

> -Rob