Re: Re: hdot264 - Developer

Hi,

I'm a student, (hopefully) doing his last year of CS. As a final work I have
chosen to do some profiling on the encoder reference software
(http://bs.hhi.de/~suehring/tml/download/) and trying to optimize it.  I'm
currently reading the reference
(ftp://ftp.imtc-files.org/jvt-experts/2003_03_Pattaya/JVT-G050r1.zip) during
my summer job and plan on starting half september with the actual coding.

I haven't decided yet if I'm going to change the reference software or just
going to start from scratch.  The problem with starting from scratch will be
that looking for bugs will probably consume more time then actually
implementing it.  On the other hand, if I continu on the ref software I'm
going to be limited by the design choices made there.

I'm also still in doubt about how low-level I'm going to work.  I have a
decent notice of both C and assembler both have no idea on what to choose.
At the moment I'm thinking of implementing it in C (using Visual C/C++) and
then maybe adding inline assembler.  But I don't know how this affects
platform independency. I remember that linux gcc compiler mostly uses a
different assembler notation so this could be not such a good idea.  Maybe
someone knows a decent C + asm compiler which is available for different
platforms?

Anyway you probably gonna here from in the coming year.

Yves

----- Original Message -----
From: "Aitor Garay" <[email protected]>
To: "skal" <[email protected]>
Cc: <[email protected]>
Sent: Wednesday, July 09, 2003 2:41 PM
Subject: Re: [Hdot264-devel] Re: hdot264 - Developer


> Hi there!,
>
> > indeed, i'm playing with ref software in order to get a grasp of
> > H264 capabilities. And i must admit H264 is by far more complex
> > than i first thought.
>
>     What impressions did you get while playing with the reference
> software?  Do you think it's worth on using it as a base for developing
> a new codec, or it's better to start a new clean implementation?
>
> > And i doubt any real-time decoding can be
> > achieved before long, so that's a little depressing...
>
>     Encoding is even more depressing,   Some sources suggest that
> H.264/AVC encoding requires roughly 8 times more encoding processing
> power than MPEG-2.  Other sources say that it's only 4-5 more
> complex.  There is a big difference between both estimations!
>
> > Maybe
> > it's time for me to go on summer vacations :)
>
>     Enjoy.
>
>     /AITOR
>



-------------------------------------------------------
This SF.Net email sponsored by: Parasoft
Error proof Web apps, automate testing & more.
Download & eval WebKing and get a free book.
www.parasoft.com/bulletproofapps