[theora-codecs] Call for bugs and suggestions [One more try]

[Mauricio Piacentini]

Dan, I now have updated binaries compiled from cvs source code, but no place
yet to put them.  What happens is that the cvs tree is slightly broken for
the Mac, so a direct compile is not possible without some small tweakings.
There is also the issue of not being able to compile the codec directly with
CW 6 or 7: you have to compile the dxv libraries with CW5.3 first, and then
use the libraries to produce your final OSX binary in CW6 or 7 (this is all
in the readme, but difficult to follow for people that do not have the last
3 or 4 versions of CW installed, and different build environments setup). If
you try to build the libraries directly with CW 6 or 7 some assembler files
will not work. There is also an issue with the resource files when used with
the latest CW versions, where the resource produced contains some invalid
fields. It is necessary to manually edit some resources for the OSX codec,
so I doubt a lot of users can compile this code as it stands now.

Just to let you know, I am upgrading the Windows projects to use .Net, as
this will help Windows users get rid of the requirements to install VC 6,
some service packs, get the iaxmm.inc, and the processor pack.  For the Mac
I've been trying to make a build that is easy to compile with CW7, without
requiring the intermediary CW 5.3 build, but this has not worked so far. I
will post updated trees and binaries sometime later this week. Of course it
would be wonderful if it was possible to port everything to use free
compilers and tools, but this is not possible for the Mac and difficult on
the Windows at this stage (due to the assembler files.)

I investigated the issue with QT6 on the Mac and the issues reported on the
Mac (problem playing in scaled mode.) It turns out that this does not look
like something that can be addressed at the codec level. It appears that
there has been a change in QT6 (probably because of Quartz Extreme) in that
scaling is done in hardware on some machines, and there are slightly
different results depending on the video driver (ATI or Nvidia.). I have
some screenshots posted, you can compare scaling on the same machine, using
the same codec build, with QT5 and QT6. What is important to notice is that
exact scaling (2X larger, 4X larger) looks very good in QT6 and it is fast.
The problem is with non-exact scaling (320X240 to 460X320, for example). I
have not looked at the Windows-QT6 issue reported today yet.

Just a thought: one thing that might help a lot for the Mac side is to get
rid of QT4.x compatibility, and compile with QT5 or 6 headers directly. I
think this makes sense to make it easy to maintain the build, and older
users can still use the QT4.x versions. It is difficult (sometimes
impossible) to fix QT6 specific issues while trying to maintain a QT4 build
environment, using older Universal Interfaces and qt headers. It may even be
better to start maintaining only OSX builds from now on, and label the
existing one as the last version for OS9 and its current QT version (which
will not be updated further by Apple anyway.)

Regards,
Mauricio Piacentini
Tabuleiro

----- Original Message -----
From: "dan miller" <danbmil99 at yahoo.com>
To: <theora-codecs at xiph.org>
Sent: Wednesday, April 23, 2003 1:37 PM
Subject: Re: [theora-codecs] Call for bugs and suggestions [One more try]

<p>> I wouldn't assume the On2.com binaries are up to date.  Check out the code
from cvs and compile
> yourself to be certain you have the most up-to-date versions.
>
> I've suggested before that we should have binaries up at theora.org but I
don't think that's
> happened yet.

--- >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 'theora-codecs-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.