Re: Some hints for getting started WAS: Let's go for it then...

Hallo!,

    Interested in collaborating in the project too?  ;-)  We need to form
a task force of people willing to contribute, if not, the project will never
go on...

    /AITOR

----- Original Message ----- 
From: "Steffen Schulze" <[email protected]>
To: <[email protected]>
Sent: Friday, July 11, 2003 12:56 PM
Subject: [Hdot264-devel] Some hints for getting started WAS: Let's go for it then...


> Yves et al,
> 
> > My 'assignment' is mostly directed towards the encoding part so I'm
> planning
> > to spent most of my time on that part.  However, I don't think you can
> see
> > them both seperatly.
> 
> Yes, this is my opinion too. If you start the work on Encoder and
> Decoder together, you are more flexible to get a first milestone.
> 
> As Aitor already suggested, a first milestone can be an compliant
> intra coder. I think a first mile stone can be an Encoder (and the
> suitable Decoder) which is able to encode luminance information of MBs
> in Intra16x16 with DC-prediction only. For this simple task you will
> need already many of the core components of h.264: prediction, dct,
> quant, dequant, idct, reconstruction and entropy-coding (either
> UVLC/CAVLC or CABAC). For the actual codec you will need additional
> tools like bitstream management, memory management, reference frame
> management, parameter-IO (okay, perhaps not for this first version ;),
> and so on.
> 
> If you got this C-written very simple (and not neccessarily H.264
> compliant) Codec working, you can add more and more functionality to
> it. Note that those following task can be done independently.
> 
> - add chroma coding
> - add the remaining three Intra16x16 prediction modes
> - add Intra4x4
> - add the other entropy-coding method
> - optimize core components with assembler (but try to keep your
>   C-sources working - for compatibility)
> - ...
> 
> If you have this, then you can test this real intra coder with the
> reference software. ;)
> 
> An additional suggestion: While the development of the first codec
> version is in progress, you should start to create an test
> application for it. Searching bugs like wrong coeffs after entropy
> encoding/decoding can take a very long time if you only have a
> standard debugger... ;)
> 
> Ciao
> Steffen.
> 
> 
> 
> -------------------------------------------------------
> 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/bulletproofapps1
> _______________________________________________
> Hdot264-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/hdot264-devel
> 
> 



-------------------------------------------------------
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/bulletproofapps1