21 Sep
2009
21 Sep
'09
6:39 p.m.
On Mon, Sep 21, 2009 at 06:48:31AM +0000, Michael Chapman wrote:
On Friday 18 September 2009 1:26 am, Paul Davis
wrote:
Well, that's one way to put it. As far as I'm concerned I would
consider it a disaster if this were accepted by the FLAC devs.
With all due respect to Etienne (whose motivation and energy is
beyond discussion), this proposal is seriously flawed even if it
has some merits.
It as absolutely unacceptable as a file format for production.
As a format for delivery to end users (listeners) it has some
merits, one of them being simplicity. But:
* It does not support 2nd order. Existing 2nd order material
can't be encoded without reducing it to first order. This is
an absolute no-go.
* It completely ignores the new mixed order schemes which are
probably the best options for distribution.
* Some of the technical choices (such as the special treatment
of the W channel) have no serious motivation at all. Whatever
is mentioned in the document is misleading to put it mildly.
i thought perhaps that our resident ambisonics
gurus/interested
parties should take a look at this.
God speed to Etienne and his work. He takes an interesting condition
precedent and builds on it. There is a more general solution that has attracted
wide
support. (see below).
True. And while this is not 100% what I'd want I do support it.
Note:
That paper refers to me twice, as the author of AmbDec and
Tetraproc. Neither of those reads or writes files, so the
comments made in those reference are somewhat out of place.
But anyway, AmbDec already supports SN3D (and N3D and FuMa),
and future releases of Tetraproc will do the same. I'm not
convinced that ACN is the best channel naming scheme, but at
least it is consistent and its use doesn't exclude more
informative forms such as those based on the l,m values.
Ciao,
--
FA
Io lo dico sempre: l'Italia รจ troppo stretta e lunga.