W3C home > Mailing lists > Public > www-multimedia@w3.org > October to December 1997

RE: MHEG

From: Marchisio Pietro <marchipi@XRR3.CSELT.STET.IT>
Date: Fri, 12 Dec 1997 15:29:25 +0100
To: www-multimedia@w3.org, "'Larry Goldberg'" <Larry_Goldberg@wgbh.org>
Cc: Geoff Freed <Geoff_Freed@wgbh.org>
Message-id: <E976E00A79ECD011B23000805FA6EA8E190145@xrr1.cselt.stet.it>
Dear Larry and other WWW-multimedia members,

I have been involved in the MHEG standardization (MHEG-1 and then
MHEG-5), and we at CSELT have implemented the following MHEG-5 tools
(for Windows95/NT):

*	an MHEG-5 RTE, for a "major" profile of MHEG-5
*	MediaTouch, which is a "native" authoring tool for MHEG-5
applications
*	ARMIDA, which is a complete "client-server" solution for
delivering high quality MPEG-2 sequences over ATM. It aims at being
"DAVIC compliant", and hence relies on an MHEG-5 RTE for conveying
multimedia applications to the final user. The latest version of ARMIDA
is also capable to get information via HTTP and to select TV channels,
as per DVB specifications.  
*	a set of MHEG-5 application examples, for the VoD, ITV and
tourist domains
*	a demo CD-ROM is also available (demo version of MediaTouch,
RTE, and application examples)

More details are available at "www.cselt.it/ufv". By the end of the year
we are also planning to make some MHEG-5 applications available on-line,
with the possibility of downloading the MHEG-5 RTE.

I also spend some words on the relationships between MHEG and SMIL,
focusing  on MHEG-1 which is quite more semantic carrying than MHEG-5.

The MHEG-1 support includes, but is not limited to:

*	object composition: more hierarchic levels are possible, i.e. a
"composite" object can contain both "composite" and "elementary" objects
*	layout: every object has a presentation space. There is a
hierarchical organization of "presentations spaces": if you scale a
composite, all descendants are affected
*	temporal dimension integrated in the presentation space, i.e. if
you change the "speed" of a composite, the speed of all the descendants
is affected too
*	audio dimension (volume control) integrated in the presentation
space, i.e. actual volume of an audio is normalized to the volume of its
ancestors
*	linking. A Link condition can be expressed as a combination of
"trigger" and "constraint" conditions. When a link fires, an "action
object" is executed. 
	Example of condition: ((button1 becomes selected) AND ((video1
IS NOT running) OR (currentTemporalPosition of Video2 > 100000))) 
*	action. Hierarchical organization of serial/parallel branches of
execution, with "elementary actions" at the leafs. An elementary action
can have a simultaneous effect on more target objects. Moreover a
"transition duration" on the effect of the action can also be specified
for certain elementary actions.
	Example of elementary action: setSize (<bitmap1, bitmap2,
bitmap3>, 100, 100, 5000)
	implies a gradual scaling of the size of the three bitmaps to
size (100, 100) in 5000 msec.
*	macro support for actions: the value of certain parameters of
the involved elementary actions are provided only when the macro is
executed
*	basic arithmetic operations and variable support

I think the following reasons fall in the list of those that hampered
the acceptance of MHEG-1 by the industry:

*	only the ASN.1 encoding format is supported, which is not
human-readable
*	it is too complex, not only to implement, but also to understand
*	no "minimum profile" was defined
*	when it was nearly ready (around 1995) DAVIC accepted it in
principle, but urged something simpler. And we have now MHEG-5 which can
be considered a "simplifying profile" of MHEG-1 even if the two syntaxes
are incompatible

It is however my feeling that there are a lot of conceptual
commonalities between MHEG-1 and SMIL, and that therefore the looking at
MHEG-1 will avoide to reinvent certain features starting from scratch.

Regards,
  Pietro Marchisio

> ----------
> From: 	Larry Goldberg[SMTP:Larry_Goldberg@wgbh.org]
> Sent: 	giovedý 11 dicembre 1997 17.32
> To: 	www-multimedia@w3.org
> Cc: 	Geoff Freed
> Subject: 	Re: MHEG
> 
>                       RE>>MHEG
> 12/11/97
> 
> thanks for the pointers.  now a developer's question:
> are there working examples or prototype implementations of SMIL, SAMI
> and MHEG or are all still theories and code at this point? any
> suggestions?
> 
> Larry Goldberg
> WGBH Educational Foundation
> 
Received on Friday, 12 December 1997 09:33:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.29 : Thursday, 2 December 2004 20:44:29 GMT