Yeah I think for the chairman
introduction in Atlanta yeah great

it's great to be here
very excited honestly.

Was not at the first party conference but
you know the vapid and obvious things are.

Very important things also for
me and my colleague and

the company of a king for we are working
for the I heard she that is the R. and D.

Institute of the public broadcasters
in Germany Austria and Switzerland.

And yeah I will talk about the.

Framework that's a transcript framework
for optic based rendering in browsers.

Parts of this very careful been developed
in the you from the project Orpheus.

Approach last our own to two and a half
years moralists at the beginning and

the main goal of the project is to develop
implement and well a day enter and

up to based media chain we have
ten partners in the project and.

You might know some of them and there are
some very valid on company companies and

institutes from all over Europe
they're dealing with the best audio or

spatial audio and such things.

So even after a talk this morning of their
great keynote from Frank this morning

some of you might ask what
the hell is based on what you.

Base maybe a simple answer.

Based audio is not the message for you
as it is often equated to it is rather.

Audio.

Plus measure they took.

I mean Frank already told about
this keep this rather short but

just to get a better understanding
of what optic based audio is

we can compare it with
the situation today.

Audio is produced and distributed.

In the current broadcast and network and
this is the so-called channel based

approach so let's say we have to for
instance or is this.

Capturing them with microphones and
recording signals these signals and

then go to a kind of post production
there so whatever you queueing filtering

effects are applied and then there's
a mix down produced and this mix tone is.

Produced for a very specific target for me
in this case it's for a stereo target for

Mitt and
then the assumption is that the user or

the audience has speakers
set up that is suitable or

not just a little that is exactly
the same as is intended to so

in this case this uses
hysteresis up this very well but

if the user has noticed there
is sort of but in this case or

in this example if I've had one
set up it is no more working.

No more of wrecking is good.

So it's a channel based approach is a
means to make a production for one format.

And adaptation afterwards
is only hardly possible.

Sort of look at the two based workflow
began Yeah I think during the capturing

those not very much change because any
signals going to the post production step

but then we don't produce a mixdown.

Rather combining the older single spaced
additional meter data such as position.

Semantic information whatsoever and
this is then the so-called optic be.

This audio scene is then
transmitted to the audience and

there is a so-called render that produces
an audio signal that is suitable for

yeah all the target
devices that are connected

to their interests of this can be five
to one stereo headphones whatsoever and

the render takes you into counts
different situations such as.

Misplaced speaker of entrance or
if the user is walking the streets so.

Based audit was formatted gnostic would
provide basically accessibility so

one could very easily.

Offer in multiple languages or
over speak only one production.

Personally station in this
very big point here and

interactivity OK just awfully You got that

now let's come to the Baucus frame very
the reason I am here on the right side.

You can see a diagram and

overview of the classes that
are implemented in the frame vary and

how they are connected to each other and
with each other.

It's basically a childhood friend very
it's published under the MIT license

on get up it's written in
ECMA Script five standard and

uses very very many parts and.

Yeah also using some third party lips.

For the basic concepts and you can see
here on the right third of the screen

short code snippets how to integrate
it in your H.T.M.L. page and

so you didn't need to very much
you only need a scene file and

in the scene file there's all the relevant
info written and can be read from.

The framework office basically three
options to load or play audio signals.

Can be a single audio text that Rather.

Shorrosh and

there are fires these are then connected
with an audiobook a soulless note.

And they're loaded with a similar request
can be multiple single objects Groot.

Same and these can be one or
more long objects with the same duration.

These can have any duration and can be
represented in the file or a stream.

This is then connected with a media
element sorry snow which So an audio and

video.

Fife element.

Currently implemented of trick descriptors
are again position interactive and active.

On the slide just to give you
an impression how this scene file looks

like and what information is written there
don't want to go into details just to

show you how this looks like
it's following the spot if.

Format.

But this is only for the.

Some for the basic concepts.

Time changes.

Scheduling synchronization and
timing therefore every use the rep clock

thankfully it's a it's a great library
that helps us a lot for the timing here.

And there's also a U.I. manager class
in the frame very good that provides

some very basic functionality for
a two dimensional user interface and

yeah I thought I already set.

A lot of fights and streams.

For to be used as all your bets for
instance the so

called media controller class should
be used this new controller class.

Uses media buffer so it's not so basically
you can connect an audio to video

from a five element two to two
if to the frame very good but

the problem here be experienced.

I'm guessing that.

Is the channel order of coded
multichannel streams or flights.

And that's why be implemented a class it's
called Channel order test on the right

side you can see again to code snippets
how to use it class hopes you two

to detect automatic Uli the order of
the decoded model general checks.

For transit you take and
therefore some test files

are provided the different
channel numbers and

in each channel there is
an increased thunderstorm frequency.

Just to yeah how to detect
the order of the channel.

Decoder afterwards OK very short demo.

We.

Hear you OK So this is as I said
a very very basic use interface

just to demonstrate the object
based on the approach itself.

Just starting.

Yeah.

That's what it is it sure

was.

Interesting that it's never.

That good if your current theory I

think that if you've got
your own to it that.

They get the.

Solo stuff.

Playing around and you have to give
the user a print of what up to.

Basically.

OK great it's not working
anyway we don't have time.

Coming back to the presentation.

Just just a short summary.

Of the book to its framework has multiple
options to log in play only signals that

office automatic detection of Decoded
multichannel tricorder printers scenes for

stereo and Binaural various on moralists
all modern browsers that supported.

And my plans for the future US
are quite something I will implement.

In a very potent standard the I two
R.B.'s twenty seventy six who for those

of you that and that familiar with these
standards thing that's the so-called or

your definition model that will be become
very important in the future hope so

and the representation will be no more
despotic like but rather cheese chase and

representation and you know once it is
available I will implement the spatial

plan and
out the spool offer much more features for

me to palatial and positioning things
you know into the streaming but

capabilities should be a bit
extended OK that's it thanks very

much.

I'm not sure if it got the

question.

OK.

I see OK got it yeah but.

So the question was.

Whether.

We ever experienced or think people
experience problems with the amount of

state or because of the potentially
many channels or many objects or

audio signals that have
to be downloaded right.

Yes that's a very very common question
that pops up from time to time and.

Not like.

How do you know that that we say we
need this number it's rather that we're

thinking especially for this for
this browser case it should be.

Possible to make something
dynamic Ali that is scalable.

In that it just takes into
account the computing.

Couple that use off of the browser off or
that device and they're.

Not.

Not such experiences yet are numbers that
I can say we need at least I don't know.

Six objects or twelve very I don't
know in this all depends a little bit

on the use case and on there the piece
of content you are transmitting.

I hope this answers your question a little

bit.

So you ask for it personally

sation.

So it's not just a start and
stop it's also like

positioning of the of the process of
thought object of course it's like

you can make sure to check to research
movements of objects of course and

you can make them basically you could
it's not very much implemented yet

in the Schoenberg honestly there are many
things will come in the future but

basically you this author could just say

the user shall have this degrees
of freedom for this object or for

those objects so this is everything and
possible and this will hopefully then come

with the yeah I am implementation.

Thank you.