Re: Fw: Fw: sursound: The pinna and Ambisonics/cross post

[Peter Lennox <peter@xxxxxxxxxxxxxxxxxxxxxxxx>]

Ongoing:
Reply to the forward of some suggestion from this list
.....----- Original Message -----
From: "DG Malham" <dgm2@york.ac.uk>
To: "Peter Lennox" <peter@lennox01.freeserve.co.uk>
Cc: <sursound@darkwing.uoregon.edu>
Sent: 24 October 2000 17:09
Subject: Re: Fw: Fw: sursound: The pinna and Ambisonics/cross post


> There is nothing in this which in any way contradicts the point I was
> trying to make. The human ear, for continous sounds above a few kilohertz,
> has very poor localisation capabilities. Sounds containing both low and
> high frequencies and are continous are much easier to locate, lower (but
> not lowest) frequency continous sounds on their own are not as good mixed
> low and high, but much better than high frequency continous sounds on
> their own and discontinous sounds with a broad spectrum, such as music,
> are best of all. Whether the source is real or artificial is largely
> irrelevant - ever tried to locate the source of the high frequency whistle
> generated by escaping gas? Speakers do indeed contribute to poor
> localisation, especially from cabinet resonances - as was well known many
> years ago...that's one of why the BBC developed the highly damped, thin
> wall speaker cabinets for their own design of, highly regarded at the
> time, monitors.
>
> Dave
>
>

Right,
But in the case of ambisonics, weren't you saying that second-order
implementation is capable of generating a degree of resolution of
information concerning the angular separation of image sources which (for an
ideally placed listener) approaches the limits of discernment deployed by
humans? Because what comes out of this is that, for many 'naturally' (I
know, I know- it's a bit woolly) occuring hf-only sounds, (ie. non- steady
state ones), localisation -sorry,-lateralisation, is a great deal better
than 'within 30 degrees', no?
What I'm asking is, if the assumptions on which ambisonic decoding was
originally based extended to the notion that 'hf' is directionally perceived
via Inter-aural intensity differences, when in fact for most 'real world'
examples, direction perception relies much more heavily on the inter-aural
temporal differences available in the low-frequency modulation of the high
frequency-only sound in question, then is second-order really the upper
limit, or could one theoretically carry on going to higher orders,
continuing to reap significant benefits?
Actually, the choice of example of escaping gas is somewhat apposite ; I've
recently been doing just that! - what I found was that I had no difficulty
whatsoever in finding it, and that's with 47-year-old ears!. however, I AM
fully prepared to admit : a) that the hiss was probably modulated at a lower
frequency due to turbulent flow, and b) that I was not at all constrained to
one listening position. Furthermore, the leak in question was in a small,
semi-enclosed space, which would result in many early reflections,
comb-filterings and so on. So perhaps it was not a fair test of the example
you had in mind. But actually, it was representative of the sort of thing I
had in mind when I proposed that, push comes to shove, Hf-borne information
might well be more 'spatially informative' than Lf. So there.
But there was another thing here ; my contention that speakers sound more
like each other than they do sound like anything else.
 Actually, this bald statement breaks down into a host of others. For
instance, the sound of a mobile phone on tv is almost indistinguishable from
an actual phone (if they are the same tune); ....is this just me? - I find
the same with sirens, tills etc.
 Now, when was the last time you mistook the output from your tinny tv
speaker for anything other than a speaker? - the only time might be
something like voice, and ESPECIALLY when you are not in the same room. I
accept that, as tv sound systems get better, then some of the Lf stuff might
be confusing too, possibly sounding like a car door outside, or some such,
but that's not the same as mistaking a 'virtual' object for the equivalent
'real' one. And I just don't accept the hoary old argument that it is only
to do with prior knowledge of the world; ths argument is gainsaid somewhat
by the mistakes people DO make (for instance, when someone's phone goes off,
see how many people check their own even when it's a completely different
ring).
So, we could put this this way: "when does a speaker not sound like
itself" - a) when sounding like another, or b) when in a different room.
-I wonder if members of this list can come up with any other examples?

It does seem to me that this isn't something to do with frequency response,
dynamic range, colouration and so on; but more something to do with the way
we know of the corporeal nature of the 'things' in our environment; and that
this has something to do with informational regularities in the dispersion
patterns of such 'things', perhaps in conjunction with their consequent
interactions with local surfaces, which, over time, we are able to utilise.
I will indeed check out Eckard Blumschein's suggestion with reference to
speakers.
I do accept your point about cabinet resonances quite easily spoiling
'illusory' spatial information; a little contradictory information goes a
lot further than an awful lot of confirmatory stuff! - it's a neurological
something to do with humans detecting 'change' more readily than non-change,
I think. In other words, we tend to determine 'truth' by looking for
non-truth, I think; a bit like scientific method, really?

thanks for your time on this,
regards,
ppl
----- Original Message -----
From: "DG Malham" <dgm2@york.ac.uk>
To: "Peter Lennox" <peter@lennox01.freeserve.co.uk>
Cc: <sursound@darkwing.uoregon.edu>
Sent: 24 October 2000 17:09
Subject: Re: Fw: Fw: sursound: The pinna and Ambisonics/cross post