[MPlayer-dev-eng] aes & bluray

[Nico Sabbi]

Il venerdì 11 settembre 2009 11:01:07 Reimar Döffinger ha scritto:
> On Fri, Sep 11, 2009 at 10:39:48AM +0200, Nico Sabbi wrote:
> > Il venerdì 11 settembre 2009 10:20:42 Reimar Döffinger ha scritto:
> > > On Fri, Sep 11, 2009 at 04:51:56AM +0100, Jonathan Nell wrote:
> > > > 2009/9/10 Nico Sabbi <nicola.sabbi at poste.it>:
> > > > > I didn't read the specs, but unless decryption can't be
> > > > > implemented otherwise I'd prefer to keep it in the stream
> > > > > layer, just like for CSS
> > >
> > > And why? I really think it is a mistake to do it there for CSS
> > > already. The encryption part of CSS does not really depends on
> > > a physical medium at all and could be used for any MPEG stream
> > > - if it wasn't always implemented "at the wrong place".
> >
> > because different media could use different encryptions in the
> > same mux format, think for example of dvd and hd-dvd: both use
> > mpeg-ps as a container but with very different encryptions.
> > Even though there's nothing stream-specific in the encryption
> > method, as a matter of fact the media type (thus the stream)
> > involves an encryption algorithm.
>
> I admit I expected that they could be easily distinguished e.g. by
> the value of PES_scrambling_control, but I have not found out what
> the value for DVDs is (it is 01 for HD-DVD).
> And while I don't know if it is enough to decide the type of
> encryption, it is PES_scrambling_control that determines if it
> needs to be decrypted, not the source media.
>

yes, but "how to decrypt it" is the problem.
You would have to add decryption members in the demuxers, making the 
code even worse than it is now, while keeping the decrypters in the 
stream readers the code looks much cleaner.