I'm mostly just an interested reader on this group having an historical
interest in Ambisonics from days in the 70's when I was a member of the Oxford
Tape-Recording Society (which MAG and Peter Craven ran) and continue the
interest making amateur recordings in B format.
Although I have very limited experience of binaural I do remember the one and
only convincing binaural playback I've heard made in a demo at OUTRS using the
original quad speakers either side of the audience. The only poor localisation
was in front but height etc. very convincing. I don't remember the mic
configuration though. Could this good result be because the soundfield is
stable and head movements are used as would be the case for head-tracked
headphones? But then why front localisation still poor? Be interested to know
what people think.
Martin
From: Richard Lee <[email protected]>
To: "'[email protected]'" <[email protected]>
Sent: Tuesday, 14 June 2016, 12:35
Subject: Re: [Sursound] Using Ambisonic for a live streaming VR project
> The main mechanisms for disambiguating 'cones of confusion' (and this
includes front-back reversals) are: pinnae effects (Batteau) and
head-movements (Wallach) - so, without either of these mechanisms at play,
one would expect directional ambiguity.
You can test the relative importance of these for YOURSELF with the famous
Malham / Van-Gogh Experiment
http://www.ambisonia.com/Members/ricardo/PermAmbi.htm/#VanGogh.
I still have some Diamond encrusted caps with optional Golden Pinnae but
you need to pay in used bank notes. No Confederate money please.
Michael came up with his rE & rV theories ... not by considering how to
best replicate HRTFs bla bla .. but by asking ... "what information could
the Mk1 Human Head (+ torso + processing inside + bla bla) possibly have
available to determine localisation?"
If youi perform the above experiment, you'll find the Moving Head cues are
FAR more important than the Fixed Head cues (HRTFs bla bla).
Where the HRTFs have the most significance is in the vertical plane. It's
the different frequency response as a source moves off the horizontal plane
that allows the Mk1 HH to process 'height'. But even then, Moving Head
cues are far more unambiguous .. and don't require a priori knowledge of
the source.
If the HRTF cues break down completely (eg simulating a pair of coincident
back to back cardioids as the crudest possible binaural decode), simulating
the Moving Head cues (head tracking) lets the Mk1 HH decode all this
without any problem, fuss or discomfort.
> I would like a little more information on ?head movements?. I suspect
all head movements are being treated as equal, and I have a theory that
short rapid movements (like shaking the head) should be treated separately
from movements that include the shoulders, or even the whole body. Short
rapid movements of the eyeball have been studied and are well understood;
without these small movements the visual field collapses completely. Does
something similar happen for the aural field ?
One of the more surprising things that Michael worked out is that the
Moving Head localisation models gave the "same answers" regardless of
whether they assumed you turned your whole body to face the source (eg
Makita) .. or those that only allowed small involuntary head movements (eg
Clark, Dutton & Vanderlyn IIRC)
It's all there in his "General Metatheory .... " if you are prepared to
study it and follow up the references. See especially the 'stereo'
appendix.
http://www.aes.org/e-lib/browse.cfm?elib=6827
_______________________________________________
Sursound mailing list
[email protected]
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit
account or options, view archives and so on.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://mail.music.vt.edu/mailman/private/sursound/attachments/20160614/38fb93d9/attachment.html>
_______________________________________________
Sursound mailing list
[email protected]
https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit
account or options, view archives and so on.
