Tech Insider					   Technology and Trends


			   USENET Archives

From nicolas@bnp-eng.remcomp.com Mon, 21 Jun 1999 11:28:06 +0200
Date: Mon, 21 Jun 1999 11:28:06 +0200
From: Eric Nicolas nicolas@bnp-eng.remcomp.com
Subject: [LinuxDVD] I'm working on linux dvd software

Hi everybody,

I'm new on this list. I'm french (so please excuse if my English
gets bad sometimes :-). I'm really really fond of DVD. I own a
quite decent home theater equipment including a Sony DVD player,
as well as 150 or so DVDs (both Zone 1 and 2). I am also the 
creator and maintainor of the first French DVD information web
site http://www.dvdfr.com/.

Professionally, I'm a C++ developer for a big banking corporation.
At home I'm using Linux (RedHat 5.2, soon 6) and I have a 
bi-pentium II 233 mhz with Matrox Millenium 2 graphic card and
a DVD-ROM drive (Creative Labs, no decompression card).

I've watched the ML archives and saw that Paul Volko was wondering
who was working on linux dvd software. I am ! (Though I'm just
begining...)

More precisely, I want to write portable dvd software (at least
portability across Windows 98/nt/2000, Linux and if possible other
Unixes). I want to release the code in some sort of GPL license,
although some legal questions may arise concerning DVD, MPEG2 and 
Dolby Digital names/algorithms use.

I am not interested in hardware-helped solutions. I want a 100%
software dvd solution which will run on any PC with a DVD-ROM drive.
I am interested in both dvd playing and dvd authoring.

Here is my "business-plan" :-) :

- Write small and optimized software components for many dvd
  manipulations topics (maybe close to some DVD-API you were
  speaking about...). I've already identified :
  1 Back-End Video - Display a YUV420 frame onto the screen ;
  2 Back-End Audio - Send a stereo numeric signal to the soundcard ;
  3 MPEG2 video decoder
  4 Dolby Digital audio downmixer (X.X chanels -> stereo)
  5 DVD interactivity engine
  Most of this components can be tested and optimized almost alone,
  on test samples. My work has begun on topic 1 (back end video).
  My goal is to have it working for Linux XWindows and MsWindows
  by Mid-July, with full MMX optimisation.

- Integrate those components to build a dvd player that reads 
  non encrypted discs. This will demonstrate the feasibility or 
  not of 100% software playback under Linux. Some discs are not
  CSS encrypted (Zone 1 "Highlander" is not, most Vivid adult titles
  are not...).
  Have this player work exactly the same under Linux and MsWindows.

- Solve the problem of CSS. By this time the people working on the
  CSS negociation with the dvd-rom drive will probably have finished
  their job. If not, I will have to do some reverse-engeneering on a
  msWindows code such as PowerDVD diagnostic program.

- Start working on authoring code :
  1. UDF filesystem writer ;
  2. MPEG2 compression code, with quality control and bit-budget check ;
  3. DVD Menu/Interactivity manipulation program (lot of UI in this one.
     Java could help for portability)

Now, what I miss :
- Some other good guys to help me on some topics. I would appreciate
  those knowhow :
       good C/C++/ASM on linux or mswindows ;
       good reverse-engeneering ;
- DVD specifications (Book B mostly) (just like everyone else)
- MPEG2 specifications (just like everyone else)

I hope I will hear from you linux dvd interested guys !

-- Eric Nicolas 
-- Createur de http://www.dvdfr.com/

From derek@spider.com Mon, 21 Jun 1999 19:02:25 +0100
Date: Mon, 21 Jun 1999 19:02:25 +0100
From: Derek Fawcus derek@spider.com
Subject: [LinuxDVD] I'm working on linux dvd software

On Mon, Jun 21, 1999 at 11:28:06AM +0200, Eric Nicolas wrote:
> Hi everybody,
> 
> More precisely, I want to write portable dvd software (at least
> portability across Windows 98/nt/2000, Linux and if possible other
> Unixes). I want to release the code in some sort of GPL license,
> although some legal questions may arise concerning DVD, MPEG2 and 
> Dolby Digital names/algorithms use.
> 
> I am not interested in hardware-helped solutions. I want a 100%
> software dvd solution which will run on any PC with a DVD-ROM drive.
> I am interested in both dvd playing and dvd authoring.

  Well I'm interested in the above, but using the h/w assist where it
exists in various video cards.

> Here is my "business-plan" :-) :
> 
> - Write small and optimized software components for many dvd
>   manipulations topics (maybe close to some DVD-API you were
>   speaking about...). I've already identified :
>   1 Back-End Video - Display a YUV420 frame onto the screen ;

  This could be done as an enhancement to X.  Add a way to do XShmPutImage
with a YUV data format XImage.  That gives you the final state - YUV to
RGB colour space conversion.  Most video chip sets seem to support at
least this much in some form of hardware.  Saves you a table lookup or
a dither algorithm.

>   2 Back-End Audio - Send a stereo numeric signal to the soundcard ;
>   3 MPEG2 video decoder

  There are already a number of free (BSD licence) source bases available
for MPEG-1 / MPEG-2 decoding.

>   4 Dolby Digital audio downmixer (X.X chanels -> stereo)

  The specs for DD are available from Dolby's web site.

>   5 DVD interactivity engine
>   Most of this components can be tested and optimized almost alone,
>   on test samples. My work has begun on topic 1 (back end video).
>   My goal is to have it working for Linux XWindows and MsWindows
>   by Mid-July, with full MMX optimisation.

  I don't know that MMX will help that much,  but I believe the new
3D instructs (3Dnow! and intels version) will help for the IDCT bit.

> - Integrate those components to build a dvd player that reads 
>   non encrypted discs. This will demonstrate the feasibility or 
>   not of 100% software playback under Linux. Some discs are not
>   CSS encrypted (Zone 1 "Highlander" is not, most Vivid adult titles
>   are not...).

   Hmm,  well I'll have to see if I can get Highlander.

> - Some other good guys to help me on some topics. I would appreciate
>   those knowhow :
>        good C/C++/ASM on linux or mswindows ;
>        good reverse-engeneering ;
> - DVD specifications (Book B mostly) (just like everyone else)
> - MPEG2 specifications (just like everyone else)

  You can get draft versions of these from www.wotsit.org,  however the
MPEG-1 specs are a integral part and these aren't (as far as I know)
available.  At least one of the free s/w decoders seems to be a literal
translation of the MPEG-1 spec into C.

  I'd certainly be interested in this sort of thing.  At the moment I've
bought the MPEG-1 specs and intend to get a h/w assisted MPEG-1 video
decoder done.  I'll then enhance this for MPEG-2 once the first bit is
done.

DF
-- 
Derek Fawcus                                                    derek@spider.com
Spider Software Ltd.                                        +44 (0) 131 475 7034

From castongj@wam.umd.edu Mon, 21 Jun 1999 15:39:19 -0400 (EDT)
Date: Mon, 21 Jun 1999 15:39:19 -0400 (EDT)
From: Jason Castonguay castongj@wam.umd.edu
Subject: [LinuxDVD] I'm working on linux dvd software

On Mon, 21 Jun 1999, Eric Nicolas wrote:
...

> More precisely, I want to write portable dvd software (at least
> portability across Windows 98/nt/2000, Linux and if possible other
> Unixes). I want to release the code in some sort of GPL license,
> although some legal questions may arise concerning DVD, MPEG2 and 
> Dolby Digital names/algorithms use.
> 

Existing mpeg2 implimentations seem to be patented and require a fee for
comercial use.

...
> 3 MPEG2 video decoder

Try either http://mpeg.openprojects.net or get the SDL library at
http://www.devolution.com/~slouken/SDL and the mpeg add-on at
http://www.lokisoft.com/opensource/opensource.html which says it only does
mpeg1 decoding, but it does have some support for mpeg2.

also a couple (mpg123, splay) mp3 players support some forms of mpeg2
audio.

...

> - Solve the problem of CSS. By this time the people working on the
> CSS negociation with the dvd-rom drive will probably have finished
> their job. If not, I will have to do some reverse-engeneering on a
> msWindows code such as PowerDVD diagnostic program.

I've been meaning to ask someone about CSS. I've looked and looked for the
algorithm, but all I have is

1) that it is based on DES

2) it was a poor design to start with and they implimented it poorly

If anyone has more information then that, please send it to me. It has
been hinted time and time again that CSS is easy to break, but no one says
what it is.

> 
> - Start working on authoring code :
> 1. UDF filesystem writer ;
> 2. MPEG2 compression code, with quality control and bit-budget check ;
> 3. DVD Menu/Interactivity manipulation program (lot of UI in this one.
> Java could help for portability)
> 
for UDF try http://trylinux.com/projects/udf/
for MPEG2 try http://mpeg.openprojects.net
and for a UI, I would recomend modifying an existing UI for a mp3 player
or something. I would recomend against java, instead there are a couple
different librarys for both X and Windows. QT is the first to come to
mind, but i am pretty sure there are others (GTK ?).

It seems most everything is available for a opensource dvd player except
getting around CSS.

--
Jason Castonguay

From castongj@wam.umd.edu Mon, 21 Jun 1999 15:39:19 -0400 (EDT)
Date: Mon, 21 Jun 1999 15:39:19 -0400 (EDT)
From: Jason Castonguay castongj@wam.umd.edu
Subject: [LinuxDVD] I'm working on linux dvd software

On Mon, 21 Jun 1999, Eric Nicolas wrote:
...

> More precisely, I want to write portable dvd software (at least
> portability across Windows 98/nt/2000, Linux and if possible other
> Unixes). I want to release the code in some sort of GPL license,
> although some legal questions may arise concerning DVD, MPEG2 and 
> Dolby Digital names/algorithms use.
> 

Existing mpeg2 implimentations seem to be patented and require a fee for
comercial use.

...
>   3 MPEG2 video decoder

Try either http://mpeg.openprojects.net or get the SDL library at
http://www.devolution.com/~slouken/SDL and the mpeg add-on at
http://www.lokisoft.com/opensource/opensource.html which says it only does
mpeg1 decoding, but it does have some support for mpeg2.

also a couple (mpg123, splay) mp3 players support some forms of mpeg2
audio.

...
 
> - Solve the problem of CSS. By this time the people working on the
>   CSS negociation with the dvd-rom drive will probably have finished
>   their job. If not, I will have to do some reverse-engeneering on a
>   msWindows code such as PowerDVD diagnostic program.

I've been meaning to ask someone about CSS. I've looked and looked for the
algorithm, but all I have is

1) that it is based on DES

2) it was a poor design to start with and they implimented it poorly

If anyone has more information then that, please send it to me. It has
been hinted time and time again that CSS is easy to break, but no one says
what it is.

> 
> - Start working on authoring code :
>   1. UDF filesystem writer ;
>   2. MPEG2 compression code, with quality control and bit-budget check ;
>   3. DVD Menu/Interactivity manipulation program (lot of UI in this one.
>      Java could help for portability)
> 
for UDF try http://trylinux.com/projects/udf/
for MPEG2 try http://mpeg.openprojects.net
and for a UI, I would recomend modifying an existing UI for a mp3 player
or something. I would recomend against java, instead there are a couple
different librarys for both X and Windows. QT is the first to come to
mind, but i am pretty sure there are others (GTK ?).

It seems most everything is available for a opensource dvd player except
getting around CSS.

--
Jason Castonguay

From pvolcko@concentric.net Mon, 21 Jun 1999 17:47:01 -0400
Date: Mon, 21 Jun 1999 17:47:01 -0400
From: Paul Volcko pvolcko@concentric.net
Subject: [LinuxDVD] I'm working on linux dvd software

> > More precisely, I want to write portable dvd software (at least
> > portability across Windows 98/nt/2000, Linux and if possible other
> > Unixes). I want to release the code in some sort of GPL license,
> > although some legal questions may arise concerning DVD, MPEG2 and 
> > Dolby Digital names/algorithms use.
> 
> Existing mpeg2 implimentations seem to be patented and require a fee for
> comercial use.

I don't knwo about MPEG, but I know AC-3 has to be licensed for 
distribution of any kind (free or not), regardless of the specs being 
publicly available.
 
> > - Solve the problem of CSS. By this time the people working on the
> >   CSS negociation with the dvd-rom drive will probably have finished
> >   their job. If not, I will have to do some reverse-engeneering on a
> >   msWindows code such as PowerDVD diagnostic program.
> 
> I've been meaning to ask someone about CSS. I've looked and looked for the
> algorithm, but all I have is
> 
> 1) that it is based on DES
> 
> 2) it was a poor design to start with and they implimented it poorly
> 
> If anyone has more information then that, please send it to me. It has
> been hinted time and time again that CSS is easy to break, but no one says
> what it is.

You're not likely to find anyone who knows the algorithm or how it 
works to come out any say anything unless they has discovered it 
via RE or perhaps some kind of cryptanalysis.  It is *highly* 
gaurded within the industry.  CSS licensing is *very* strict in 
keeping the knowledge under wraps.  He or she who breaks the 
secrecy will probably be hunted down and subjected to never 
ending legal battles.

That said, there are some ofus who are working on this issue and 
finding out what kind of restirctions there are to get the license and 
whatnot.  Check the archives for this list to see some previous 
discussion on this.

> It seems most everything is available for a opensource dvd player except
> getting around CSS.

Not quite.  Also need to know how the chaining information is 
situated and what it means.  Also need to know the way that 
multiple streams of audio, video, and sub-picture data are encoded. 
The VOB files are not simply an mpeg-2 stream that gets dumped 
to a software or hardware decoder, at least in most cases.

From nicolas@bnp-eng.remcomp.com Tue, 22 Jun 1999 16:28:36 +0200
Date: Tue, 22 Jun 1999 16:28:36 +0200
From: Eric Nicolas nicolas@bnp-eng.remcomp.com
Subject: [LinuxDVD] I'm working on linux dvd software

Thanks to those of you who seems interested in my previous email.
I'll try to answer everybody and make a clear sum-up of what was said.

GENERAL
-------

> > - Start working on authoring code :
> >   1. UDF filesystem writer ;
> >   2. MPEG2 compression code, with quality control and bit-budget check ;
> >   3. DVD Menu/Interactivity manipulation program (lot of UI in this one.
> >      Java could help for portability)
> 
> This is a good idea that hasn't really been discussed before.  The
> proposed DVD API would easily be able to incorporate the
> necessary functionality to actually create all the data structures for
> the files and whatnot.

This would be a very interesting part of the project. As HD drives are
becoming bigger, and DVD writers becoming cheaper, the authoring of DVD
will become a mass market. Everybody with a camera will want to transfer
its happy holidays to DVD :-)

Jason Castonguay wrote:
> and for a UI, I would recomend modifying an existing UI for a mp3 player
> or something. I would recomend against java, instead there are a couple
> different librarys for both X and Windows. QT is the first to come to
> mind, but i am pretty sure there are others (GTK ?).

I was speaking of an authoring tool UI. That is collecting compressed/
uncompressed parts of the program, create chains, add interactivity,
compute big budgets, etc... Something like of UI like Macromedia I think.

For the UI of the dvd player itself, it is really straightforward : When
you know how to handle DVD interactivity, the control-panel for the DVD
player can be a "dvd-like" program :-))

By the way, I have a general question for you guys. Do you think a
CD-ROM disc with proper UDF filesystem and properly formatted VOB, etc...
files could be read by a standard DVD player ? This would allow a
very cheap short-dvd format (for video-clips, trailers, etc...).

THE BACK END VIDEO
------------------

Derek Fawcus wrote:
> > - Write small and optimized software components for many dvd
> >   manipulations topics (maybe close to some DVD-API you were
> >   speaking about...). I've already identified :
> >   1 Back-End Video - Display a YUV420 frame onto the screen ;
>   This could be done as an enhancement to X.  Add a way to do XShmPutImage
> with a YUV data format XImage.  That gives you the final state - YUV to
> RGB colour space conversion.  Most video chip sets seem to support at
> least this much in some form of hardware.  Saves you a table lookup or
> a dither algorithm.

As I said, I would like the code to be really portable. Writing an
X extension is a really tough work, really not portable (targetted to
XFree mostly). I would like the code to work for XWin, libsvga, glide,
mswindows, (+macos? +beos?) etc...
It is true that some card have some sort of YUV->RGB conversion builtin,
but documentation is really difficult to find most of the time.

Derek Fawcus wrote:
> >   My goal is to have it working for Linux XWindows and MsWindows
> >   by Mid-July, with full MMX optimisation.
>   I don't know that MMX will help that much,  but I believe the new
> 3D instructs (3Dnow! and intels version) will help for the IDCT bit.

There are quite a lot of code examples concerning mpeg decoding with
MMX on Intel site. For YUV->RGB conversion it is really convenient, and
MMX instructions allow converting 8 pixels in one clock cycle, which is
really impressive !

I didn't found any doc/examples about 3Dnow! instructions anywhere.

BACK END AUDIO
--------------

> >   2 Back-End Audio - Send a stereo numeric signal to the soundcard ;
> >   3 MPEG2 video decoder
>   There are already a number of free (BSD licence) source bases available
> for MPEG-1 / MPEG-2 decoding.

I haven't look at them, but I know they exist. Some exists too for mpeg2
video. But they generally are example or demo codes and need a full
rewrite in order to gain decent real-time speed.

Paul Volcko wrote:
> > - Integrate those components to build a dvd player that reads
> >   non encrypted discs. This will demonstrate the feasibility or
> >   not of 100% software playback under Linux. Some discs are not
> >   CSS encrypted (Zone 1 "Highlander" is not, most Vivid adult titles
> >   are not...).
> >   Have this player work exactly the same under Linux and MsWindows.
> 
> This will get you NTSC based titles.  PAL titles don't use AC-3 I
> don't think (someone please correct me if I'm wrong), I believe they
> use MPEG1 or MPEG2 audio encoding only (can still be
> multichannel, just not AC-3 encoded) or PCM (stereo or mono
> only).  I'm not sure on this and would like some verification if
> anyone knows about this.

Nope !
In the initial DVD specs, it is true that PAL titles were to be released
with MPEG2 multichanel audio, but manufacturers here (I'm in France : Zone 2,
Pal) realized that there were already a lot of home theater fans with DD
equipment (Laserdisc fans), whereas there were no MPEG2 audio decoders
available at all !

So MPEG2 audio is now abandonned in PAL zone2, and all titles are released
with DD sountracks. Of course there are some PCM tracks here and there, mostly
on music titles.

If you want to speak about that, there is still the problem of DTS soundtracks
too :-)

CSS
---

Jason Castonguay wrote:
> > - Solve the problem of CSS. By this time the people working on the
> >   CSS negociation with the dvd-rom drive will probably have finished
> >   their job. If not, I will have to do some reverse-engeneering on a
> >   msWindows code such as PowerDVD diagnostic program.
> 
> I've been meaning to ask someone about CSS. I've looked and looked for the
> algorithm, but all I have is
> 1) that it is based on DES
> 2) it was a poor design to start with and they implimented it poorly
> 
> If anyone has more information then that, please send it to me. It has
> been hinted time and time again that CSS is easy to break, but no one says
> what it is.

Paul Volcko wrote:
> You're not likely to find anyone who knows the algorithm or how it
> works to come out any say anything unless they has discovered it
> via RE or perhaps some kind of cryptanalysis.  It is *highly*
> gaurded within the industry.  CSS licensing is *very* strict in
> keeping the knowledge under wraps.  He or she who breaks the
> secrecy will probably be hunted down and subjected to never
> ending legal battles.
> 
> That said, there are some ofus who are working on this issue and
> finding out what kind of restirctions there are to get the license and
> whatnot.  Check the archives for this list to see some previous
> discussion on this.

I understood that you simply cannot RE the CSS algorithm (this is
forbidden). You only can use it, if you find how. I also understood
that it is based on a RSA challeng for the authentifications of the
player, then the data is DES encrypted during the stream transfert
(classic PGP envelope-type protection).

> > - Solve the problem of CSS. By this time the people working on the
> >   CSS negociation with the dvd-rom drive will probably have finished
> >   their job. If not, I will have to do some reverse-engeneering on a
> >   msWindows code such as PowerDVD diagnostic program.
> 
> Negotiation with the drive is one thing, actually descrambling data
> is another.  For a 100% software solution you need both.

I don't think so. This is a case where we *do need* HW help. The descrambling
is implemented in the dvd-rom drive. One the negociation with the drive is
ok, we can read the stream as usual, I think. For the challenge, we need
1. The hardware protocol (this is under progress with the dvd ioctl() patches)
2. The rsa keys, etc... (this is provided in the software distributions of
   PowerDVD for instance, so that could be RE).

DOCUMENTATION
-------------

> >   4 Dolby Digital audio downmixer (X.X chanels -> stereo)
>   The specs for DD are available from Dolby's web site.

True, I've got them.

> > - MPEG2 specifications (just like everyone else)
> 
>   You can get draft versions of these from www.wotsit.org.

Yes !
I found MPEG2 doc there. Many thanks for this url !

> > It seems most everything is available for a opensource dvd player except
> > getting around CSS.
> 
> Not quite.  Also need to know how the chaining information is
> situated and what it means.  Also need to know the way that
> multiple streams of audio, video, and sub-picture data are encoded.
> The VOB files are not simply an mpeg-2 stream that gets dumped
> to a software or hardware decoder, at least in most cases.

You're damn right !
And there are not only the VOB files, but also all the small files
that are around and store infos about interactivity, programs chaining,
etc... I really don't know where to start a RE on this topic. Start
with a minimal disc with only one chapter menu (there are some old DVDs
like that : "Cat People" with nastasja kinski for example) ?

Paul Volcko wrote:
> > - DVD specifications (Book B mostly) (just like everyone else)
> DVD specs will cost ya $5,000 ($500 each additional
> set) and an NDA, this also gets you at least part of licensing
> needed (I think) for product distribution.

Yes. $5000. I don't have this much money to spend on this project
yet :-) Maybe if we were, let's say... at least 10 peoples, $500
each would be ok, no ? We buy it, we copy it and send it to each
other by snail mail.

Who's interested ? :-))

Jason Castonguay wrote:
> >   1. UDF filesystem writer ;
> 
> for UDF try http://trylinux.com/projects/udf/

I've got the doc. We can watch what those people are doing quietly.
UDF is really a need if we want to do authoring anyway (the very last
part of the project, I think).

LEGAL CONSIDERATIONS
--------------------

Jason Castonguay wrote:
> Existing mpeg2 implimentations seem to be patented and require a fee for
> comercial use.

I don't think berkley mpeg_encode, mpeg_decode source code is patented. But
anyway, as I said before, all this stuff needs rewrite in order to be fast.
MPEG2 is an ISO standard, so everyone is free to implement it. no problem
on this point.

Paul Volcko wrote:
> I don't knwo about MPEG, but I know AC-3 has to be licensed for
> distribution of any kind (free or not), regardless of the specs being
> publicly available.

I think you need to be licensed by Dolby laboratories in order to use
the NAME and LOGO "DD","AC3"..., but you can distribute a software that
understand AC3 without saying it :-)
(this can be a way of starting : developing and distributing the software
without saying that it actually does AC3, and then later, if everything
goes fine, buy a license).

-- Eric Nicolas 
-- Createur de http://www.dvdfr.com/

From pvolcko@concentric.net Tue, 22 Jun 1999 11:55:23 -0400 (EDT)
Date: Tue, 22 Jun 1999 11:55:23 -0400 (EDT)
From: Paul Volcko pvolcko@concentric.net
Subject: [LinuxDVD] I'm working on linux dvd software

> This would be a very interesting part of the project. As HD drives are
> becoming bigger, and DVD writers becoming cheaper, the authoring of DVD
> will become a mass market. Everybody with a camera will want to transfer
> its happy holidays to DVD :-)

This is a good point.  In fact I approached by someone in the last few
days who wanted to make a DVD of his "baby tapes."  There is a definite
market for this I think.  And the thing about the home DVD thing is that
one could easily use mpeg-2 video and pcm audio encoding and not have to
deal with royalties (for family copies for instance).  Of course.  There
is not an affordable DVD Writer out there.  The hardware isn't there yet.
Once that comes out of the stratosphere, then I'd be willing to really put
some work into an authoring package.

> Paul Volcko wrote:
> > > - Integrate those components to build a dvd player that reads
> > >   non encrypted discs. This will demonstrate the feasibility or
> > >   not of 100% software playback under Linux. Some discs are not
> > >   CSS encrypted (Zone 1 "Highlander" is not, most Vivid adult titles
> > >   are not...).
> > >   Have this player work exactly the same under Linux and MsWindows.
> > 
> > This will get you NTSC based titles.  PAL titles don't use AC-3 I
> > don't think (someone please correct me if I'm wrong), I believe they
> > use MPEG1 or MPEG2 audio encoding only (can still be
> > multichannel, just not AC-3 encoded) or PCM (stereo or mono
> > only).  I'm not sure on this and would like some verification if
> > anyone knows about this.
> 
> Nope !
> In the initial DVD specs, it is true that PAL titles were to be released
> with MPEG2 multichanel audio, but manufacturers here (I'm in France : Zone 2,
> Pal) realized that there were already a lot of home theater fans with DD
> equipment (Laserdisc fans), whereas there were no MPEG2 audio decoders
> available at all !
> 
> So MPEG2 audio is now abandonned in PAL zone2, and all titles are released
> with DD sountracks. Of course there are some PCM tracks here and there, mostly
> on music titles.
> 
> If you want to speak about that, there is still the problem of DTS soundtracks
> too :-)

Ah.  Thank you for clearing that up.  It seemed to be a lame restriction.
Is this the case for all zones or just zone2?

> > You're not likely to find anyone who knows the algorithm or how it
> > works to come out any say anything unless they has discovered it
> > via RE or perhaps some kind of cryptanalysis.  It is *highly*
> > gaurded within the industry.  CSS licensing is *very* strict in
> > keeping the knowledge under wraps.  He or she who breaks the
> > secrecy will probably be hunted down and subjected to never
> > ending legal battles.
> > 
> > That said, there are some ofus who are working on this issue and
> > finding out what kind of restirctions there are to get the license and
> > whatnot.  Check the archives for this list to see some previous
> > discussion on this.
> 
> I understood that you simply cannot RE the CSS algorithm (this is
> forbidden). You only can use it, if you find how. I also understood
> that it is based on a RSA challeng for the authentifications of the
> player, then the data is DES encrypted during the stream transfert
> (classic PGP envelope-type protection).

Form my understanding of it the authentication with the drive is simply to
establish that the keys needed to do the decryption by the decoder card or
software CSS decryption module.  The drive would not be abel to output raw
unencrypted data, I don't think.  If it does, then that would seem to be a
major shortcoming of the protection.  It would seem that those protecting
their data would want it to be encoded until the last possible moment (ie
the decoder card or CSS decryption in software before stream decoding).
Also, what is the point of putting anything on a decoder card related to
CSS if it isn't doing actual decryption.  I'm not familiar enough with the
restrictions surrounding CSS implementation to know any of this for sure.
Hopefully this will change in the not too distant future.

> > Not quite.  Also need to know how the chaining information is
> > situated and what it means.  Also need to know the way that
> > multiple streams of audio, video, and sub-picture data are encoded.
> > The VOB files are not simply an mpeg-2 stream that gets dumped
> > to a software or hardware decoder, at least in most cases.
> 
> You're damn right !
> And there are not only the VOB files, but also all the small files
> that are around and store infos about interactivity, programs chaining,
> etc... I really don't know where to start a RE on this topic. Start
> with a minimal disc with only one chapter menu (there are some old DVDs
> like that : "Cat People" with nastasja kinski for example) ?

Good luck.  I can't help in anything along the lines of REing the DVD
specific stuff as I have access to the DVD Specs.  You might want to just
start doing some fund raising instead, though, so you cna get yourself the
specs. :)  

> Paul Volcko wrote:
> > > - DVD specifications (Book B mostly) (just like everyone else)
> > DVD specs will cost ya $5,000 ($500 each additional
> > set) and an NDA, this also gets you at least part of licensing
> > needed (I think) for product distribution.
> 
> Yes. $5000. I don't have this much money to spend on this project
> yet :-) Maybe if we were, let's say... at least 10 peoples, $500
> each would be ok, no ? We buy it, we copy it and send it to each
> other by snail mail.

you get these under NDA.  No copying.  It is plastered on the bottom of
each page.  Sorry.  

> Jason Castonguay wrote:
> > Existing mpeg2 implimentations seem to be patented and require a fee for
> > comercial use.
> 
> I don't think berkley mpeg_encode, mpeg_decode source code is patented. But
> anyway, as I said before, all this stuff needs rewrite in order to be fast.
> MPEG2 is an ISO standard, so everyone is free to implement it. no problem
> on this point.
> 
> Paul Volcko wrote:
> > I don't knwo about MPEG, but I know AC-3 has to be licensed for
> > distribution of any kind (free or not), regardless of the specs being
> > publicly available.
> 
> I think you need to be licensed by Dolby laboratories in order to use
> the NAME and LOGO "DD","AC3"..., but you can distribute a software that
> understand AC3 without saying it :-)
> (this can be a way of starting : developing and distributing the software
> without saying that it actually does AC3, and then later, if everything
> goes fine, buy a license).

I can check into this some more.  It would be great if it didn't need to
be licensed until AC-3 support was actually "officially" said to be in it.
I think that it might be just a bit too semantic to fly in court.  If you
are developing with the intent of implementing AC-3 and release
code/binaries that do infact do the decoding, even if you don't claim it
does it, you have knowingly put a product out there that does it.  I would
think that Dolby could come after you for that.  But like I said, I'll
check into it.

			   USENET Archives


The materials and information included in this website may only be used
for purposes such as criticism, review, private study, scholarship, or 
research.


Electronic mail:			      WorldWideWeb:
   tech-insider@outlook.com		         http://tech-insider.org/