http://www.content-technology.com/standards/?p=162
Key issues to be addressed are, a compact and bit-efficient
representation of multi-channel audio programs with the ability to
scale up to transparent quality and the flexible rendering of an audio
program to an arbitrary number of loudspeakers for arbitrary
configurations. 3D Audio support via headphones is a key capability in
order to deliver an immersive experience for users of mobile devices.
The "mobile device part" sounds very much like me, but anyway... :-D
This technology will form the basis for MPEG-H 3D Audio, the Audio
part (Part 3) of the MPEG-H (ISO/IEC 23008) suite of technologies.
MPEG has launched a new Ad hoc Group (AhG) activity towards responding
to industry needs on the adoption of MPEG Audio technology. This AhG
will assess industry needs in the marketplace, respond to inquiries
and make available information such as conformance streams,
application notes and white papers, that might support adoption of
MPEG Audio technology in the various application domains. The
information will be disseminated by various means, including websites
and user forums and participation in the AhG is open to the public.
So, supposedly there is an AhG "responding to inquiries".
I, for my part, will try to get into some constructive contact with this
< new > group.
Best,
Stefan Schreiber
P.S.: They possibly should check some of our recent discussion.
Bo-Erik Sandholm wrote:
I was at the Stockholm MPEG meeting listening to the 3D-audio session.
If I remember correctly the guy holding the "keynote" speech actually talked
about several choices of 3D audio encodings in the future standard to choose from for
different
Target devices and distribution types.
I got the impression that there would be a head parameter field having a flag telling how many sound channels to use and according to what algorithms the
sound channels following where to be decoded.
When I listened to him I kept thinking that ambisonics B-format in varying orders, maybe in combination with a few dedicated channels for LFE effects and Dialogue would fulfill the
stated "vision" for 3D surround sound at least for applications with a modest bandwidth need.
How BAD would 3D audio really be with a Ambisonics B+ system with 4 channels
for 3D soundscape and 2 additional discrete channels for Dialogue and LFE?
This as a low entry point and higher or mixed orders could be used for more
advanced systems.
One of the speakers presented the sound design process for movie sound, The impression of that process was that almost 100% of current movie sound spatialisation is
completely artificial and very multichannel with a lot of mono sound samples to be placed and controlled in 3D, and looks a lot how the sound scape artists currently work.
In that context I felt that not all people in the industry have understood the
great idea of separating the working tools and formats for creation of a
soundscape and the
Distribution formats and decoding to arbitrary speaker localization.
From the 3D-Auro presentation, I got the impression that they need 24 bits in the audio
5.1 files to "hide" their 3D information, so their format might not be ideal
for distribution to
Bandwidth starved devices (if there will be any such devices in the future :-).
So 3D-auro is mostly the old 5.1 with a added higer level, and how well will
that work for realistic soundscapes :-).
The 3D-auro idea of 5 slim high speaker cabinets with speaker elements for
ground level and height in addition to LFE was smart, but Ambisonics could use
the same speakers and in
Addition compensate if the speakers are not placed according to the standard.
The Japanese 22.2 format is a fixed speaker setup and it is a lot of discrete
channels to distribute, my wife will NOT accept a 22.2 system in our living
room :-).
I really think this is the real (last ?) window of opportunity for more
widespread use of ambisonics.
Maybe our target for ambisonics is not to be a solution for all
implementations, but to find a niche in the device landscape with different
distribution formats.
I agree with Stefan but in addition I would like to push the advantage of a B+
with only 6 channels.
But as ambisonics have been in use for large outdoor conserts it should
probably be good enough for a cinema:-).
And it would be an extremely large advantage I think to be able to present a
LIVE demo of a 3D sound system at the next session, even if only a small one.
I was a bit disappointed that the 3D-Auro session was only in stereo :-), the
being the only talk with any sound as far as I remember.
Bo-Erik Sandholm
Ambisonic hobby recorder.
Stockholm Sweden
-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Stefan Schreiber
Sent: den 24 januari 2013 04:33
To: Surround Sound discussion group
Subject: Re: [Sursound] an exploratory mail
Richard Dobson wrote:
On 23/01/2013 01:39, Stefan Schreiber wrote:
..
Why are you actually not reading what I was posting? One of the
requirements is "arbitrary" speaker layouts. Full stop. (There will
be some fixed layouts, I guess. But still.)
...
of multi-channel audio programs and the ability to flexibly render
an audio program to an arbitrary number of loudspeakers with
arbitrary configurations.
Possibly people were working on the basis of your initial comment:
"However, my impression is that the MPEG's intention is more to settle
on something relatively simple, like 22.2, Auro-3D speaker layout etc.
"
Which would narrow the range of layouts considerably.
I can well understand the attraction that dealing with specific
companies would have for the MPEG committees. They have clearly
identified and authoritative individuals to deal with who represent
the company - whether a CTO or a CEO. Who will stand up to be the CTO
or CEO of "Ambisonics", with the support of the community? It would
ostensibly need to be someone (or a small group) not encumbered by
possible conflicts of interests with commercial organisations they
work for. But also someone who can discuss and accommodate the
special needs of cinema while making the broader argument too. So they
would still need one way or another to speak with the authentic voice
of "the industry".
Ambisonics may in all sorts of ways be both the superior and the most
appropriate technology, but even now it has barely escaped the
laboratory and the concert hall. We can be sure Auro-3D etc will be
lobbying intensively, not least on the strength of existing industry
adoption (e.g. Auro-3D's list of cinemas using their system). So
ambisonics has quite a bit of ground to make up, in effect not only to
make its case, but also to make the case against the existing and
already more established choices.
Richard Dobson
I would suggest that the "fast track" Ambisonics based proposal which I have made - relying on
already defined elements, or elements which could be defined and implemented in an easy and fast way, like a
"modified B+ scheme with higher order and more direct channels" - is way superiour to Auro-3D, even
if using "only" 3rd order.
In fact, the "3D performance" of any reasonable .AMB/HOA based approach (referring to the
potential "vertical resolution" of 2nd+ order height
elements) seems to beat any proposed commercial 3D audio solution for cinemas which I
have seen recently. (Let us also consider that the "3D performance" of
classical WFS doesn't seem to be that good, too. :-) )
Therefore, Ambisonics fits best into the official requirements of the MPEG's
CfP: It is a very natural approach for 3D audio, and arbitrary speaker layouts
are no problem.
If I am right (and it seems like...), it would be a no-brainer to chose some Ambisonics based solution as the fexible part of the new MPEG-H part 3 standard. (Of course they still might want to define 11.1 or 22.2 as typical "fixed" solutions. If we speak about solutions for 3D audio, I would not expect that you can install dozens of speakers everywhere.
However, you should keep the chance to implement some "excellent"
solutions for concert halls/big cinemas etc. But you would have to cover smaller/cheaper installations, home installations, mobile audio via
headphones...)
In spite of a former commentary of Dave Malham I would expect that it would be <
very > difficult to agree on any standard for audio objects, as there isn't any
common approach available. (They would have to chose some clearly propietary
solution, and this doesn't look like a reasonable standard definition.)
Therefore, Ambisonics seems to win again. In spite of different opinions on
some issues, the technique is based on a mathematical theory, and there is a
considerable implementation tradition.
If anything, there is far more experience available to implement sound field
based solutions than to implement to completely new object based solutions.
In this sense, I think we don't have to be too shy. Ambisonics is used in many
installations. In fact, Dolby Atmos and (DTS) Headphone:X or MDA are the
newcomers Ambisonics is used for open air concerts.
FLAC made it without CEOs or CTO's, so what? (It got a typical de-facto standard for
lossless audio compression, without < any > company support.)
But also someone who can discuss and accommodate the special needs of
cinema while making the broader argument too. So they would still need
one way or another to speak with the authentic voice of "the industry".
Yes. Audio "for 3D movies" was the initial motivation, and they are certainly
interested in 3D audio for UHD TV...
I gave some feedback to think about a flexible and broad solution for everything. They should
present a 3D audio codec in the wide sense, because this is what they really seem to need. If the
required and "to be defined" 3D audio codec is strong enough for cinema use and not too
"heavy", it should fit for UHD, games and mobile audio, too.
Best regards
Stefan
_______________________________________________
Sursound mailing list
[email protected]
https://mail.music.vt.edu/mailman/listinfo/sursound
_______________________________________________
Sursound mailing list
[email protected]
https://mail.music.vt.edu/mailman/listinfo/sursound
