possible file implementations of UA

5 views
Skip to first unread message

e deleflie

unread,
Sep 23, 2009, 4:04:39 AM9/23/09
to ambis...@googlegroups.com
All,

I'm currently exploring possible file implementations of UA to be used
on soundOfSpace.com

There are currently 2 possibilities. One is WavPack, the other is FLAC
based. I'm going to elaborate just on the WavPack issues in this
email.

A WavPack implementation of UA would involve a multichannel Wave file
wrapped and compressed by WavPack. One way to do it is for
multichannel wav file to contain no metadata. The WavPack wrapper
would. The WavPack wrapper would contain the mandatory wrapper to
express that this file is UA (i.e. UniversalAmbisonic = version
number), but may also contain the 'producer's recommended stereo
decode' (which involves 4 fields).

The questions I'd like to raise are:

- should the multichannel wav file also contain the same metadata
(duplication of info?)
- WavPack can also read the RIFF headers on a Wav file ... so it would
be possible to set the metadata on the multichannel wav file only and
have the WavPack API access them (as though they were written on the
wrapper itself) . The only issue here is that the WavPack command line
cant write these so extra command line programs would have to be
developped.
- should a new file extension be used so that applications can be associated.
- should files other than wav files be acceptible? ... such as aiff
files? (this would affect the RIFF based metadata version (unless AIFF
files can do RIFF as well (can anyone confirm this?))

Can anyone think of any other issues?

An advantage of going with WavPack is that the beta version has a
lossless mode as well, which apparently doesn't affect phase accuracy
... and radically, the lossy version can be returned to lossless by
the addition of an other file which contains correction information.
So WavPack could potentially be a processing _and_ a delivery format.

All comments welcome.

Etienne

Aaron Heller

unread,
Sep 23, 2009, 6:23:47 PM9/23/09
to ambis...@googlegroups.com
On Wed, Sep 23, 2009 at 1:04 AM, e deleflie <edel...@gmail.com> wrote:
>
> All,
>
> I'm currently exploring possible file implementations of UA to be used
> on soundOfSpace.com
[... snip ...]

>
> Can anyone think of any other issues?

The big issue with using WAV files as the underlying format is the 2GB limit.
Hence a 20 channel UA file at 48kHz sample rate with 16-bit words
would be limited to about 18 minutes.

IIRC, the IEM paper by Michael Chapman et al. suggests modifying
WavPack to handle CAF files, which do not have the 2GB limit.

> An advantage of going with WavPack is that the beta version has a
> lossless mode as well, which apparently doesn't affect phase accuracy

I think you meant "lossy", not "lossless."

> ... and radically, the lossy version can be returned to lossless by
> the addition of an other file which contains correction information.
> So WavPack could potentially be a processing _and_ a delivery format.

I do like WavPack's hybrid mode. I have done some informal listening
tests and haven't found any obvious problems.

e deleflie

unread,
Sep 23, 2009, 6:45:53 PM9/23/09
to ambis...@googlegroups.com
>> All,
>>
>> I'm currently exploring possible file implementations of UA to be used
>> on soundOfSpace.com
> [... snip ...]
>>
>> Can anyone think of any other issues?
>
> The big issue with using WAV files as the underlying format is the 2GB limit.
> Hence a 20 channel UA file at 48kHz sample rate with 16-bit words
> would be limited to about 18 minutes.
>
> IIRC, the IEM paper by Michael Chapman et al. suggests modifying
> WavPack to handle CAF files, which do not have the 2GB limit.

argghh good point ... that's crucial. And thanks to Michael I believe
that the upcoming version of WavPack does this.

There's also w64 files ... I've used those on Linux. And wikipedia
says there's also RF64 (max 18 channels).

what about the level of CAF compatibility on windows?

>> An advantage of going with WavPack is that the beta version has a
>> lossless mode as well, which apparently doesn't affect phase accuracy
>
> I think you meant "lossy", not "lossless."

yes, lossy, sorry.

There's a FLASH decompressor for WavPack ... which essentially means
that *if*, or hopefully *when* FLASH supports multi-channel audio then
we could theoretically have an ambisonic decoder embedded in web
pages.

So I'm wondering if CAF/W64 files can be read in FLASH....

Etienne

Oliver Thuns

unread,
Sep 24, 2009, 2:35:11 AM9/24/09
to ambis...@googlegroups.com
On Thu, Sep 24, 2009 at 12:45 AM, e deleflie <edel...@gmail.com> wrote:

> There's a FLASH decompressor for WavPack ... which essentially means
> that *if*, or hopefully *when* FLASH supports multi-channel audio then
> we could theoretically have an ambisonic decoder embedded in web
> pages.

This decoder consumes a lot of CPU cycles, so it's still experimental.
I'm sure there are ways to tune the performance, especially Flash's
Pixelbender could help.

> So I'm wondering if CAF/W64 files can be read in FLASH....

I don't know any lib for CAF or W64, but there is nothing in Flash 10
that prevents it.

as a side note: I hope that Google's nativeclient plugin will support
multi-channel audio output. Would be the perfect platform for
Ambisonics inside the browser.

http://code.google.com/p/nativeclient/

Reply all
Reply to author
Forward
0 new messages