Hi,
The next thing that you heard with CC3D was another psychoacoustic
phenomenon that we kind of discovered last year about what sounds do
when they come closer versus moving farther away. And we found
that we
were able to simulate something that normally can?t be done with
traditional surround sound, which is proximity.
And again, that?s not just amplitude. So we?re taking advantage of
what we learned there to create this feeling that things are being
projected into space in the D axis, the depth axis.
From: J?rn Nettingsmeier
so this is 4d spacetime, right? x, y, z, and d :) now this funny
drone
noise, is that minkowski spinning in his grave?
As Dave Malham has already pointed out, d can be expressed in terms
of x,y,z, so is not an independent coordinate. This is like trying to
combine two coordinate systems describing the same position
(Cartesian and Polar), then saying we have six coordinates = 6D.
Date: Mon, 18 Jul 2011 09:17:41 +0100
From: Dave Malham <dave.mal...@york.ac.uk>
classical ambisonics doesn't really do that. on good recordings,
you will get a very nice sense of
distance, but that is due to distance cues which are more or less
independent of ambisonics (any
good recording method can do it).
what you definitely won't get (with any order less than
"ridiculously high") are sources closer
than the ring of speakers.
Whilst I agree that you can't generally get stationary audio
objects closer than the radius of the
speakers on low order systems (currently, only high order Ambisonic
systems, WFS or crosstalk
cancelled binaural systems can do that - oh, and the various
ultrasound based speakers), you can get
reasonably quickly moving objects to appear to pass close by,
especially if the acoustic of the
playback space is dead relative to the reproduced space, provided
you give enough cues (particularly
early reflection patterns and proximity effect) in the soundscape
to override the conflicting
playback space cues. Whilst this also occurs with any decent replay
methodology, it is easier with
Ambisonics because (I suspect) of the fact that there is always
more than one speaker producing
sound, so the local space cues conflict not just with the
soundscape cues, but also each other,
weakening the perceptual effects of the local cues.
It is true that 1st order ambisonics doesn't consider distance, with
all sources being reproduced at the distance of the speakers,
although Gerzon did consider distance panning. A Soundfield mic
recording contains distance information. If attempting spatial
synthesis, the ambisonic encoding equations do not include distance,
and this has to be added in various ways: amplitude variation
(inverse square or other law), hf air absorption, early reflections
and reverberation in a virtual space, source directivity, occluding
objects etc..
Sources inside the speaker distance cannot be be correctly
represented with 1st order ambisonics, as the x,y,z components all
diminish to zero at the listeners position, and this can be
compensated to some extent by increasing W to maintain a similar
loudness. As far as I can see, higher order components also tend
towards zero (apart from R, which tends towards a constant of -0.5).
Modelling near sources in HOA seems to depend mostly on the
'proximity effect': an increase of gain at low frequencies in the
directional components.
I'm not sure that this is really 'gimmickry' as J?rn suggests.
Date: Mon, 18 Jul 2011 19:27:26 +0200
From: J?rn Nettingsmeier <netti...@stackingdwarves.net>
distance cues are mostly gimmickry in my opinion. you can fake
distance
in a number of ways, but most are really dependent on the spectrum and
envelope of the program material. most aspects of distance encoding
are
also orthogonal to most surround techniques, which means they can be
added at will, today. they don't even necessitate a fancy new name.
Modelling distance, and controlling it on a per source basis, is
founded on sound physical principles and can be made 'convincing',
even with low order ambisonics. Agreed that it is 'bolted on', though
synthesis (being the converse of analysis) involves controlling a
large number of parameters to simulate what occurs naturally.
Even WFS, as described in the literature, suggests that sources be
recorded individually as dry and close as possible, and the 'scene'
then reconstructed on playback. So it too synthesises distance.
Ciao,
Dave Hunt
_______________________________________________
Sursound mailing list
Sursound@music.vt.edu
https://mail.music.vt.edu/mailman/listinfo/sursound
