Luis (et al):

I'm a big fan of ambisonics, and code all my works in it for the reasons you 
mention.

However, I've used bformenc instead of spat3d. Assoming I'm using free field 
spatialization (no room acoustics or doppler), do you see any advantage to 
spat3d?

Art Hunkins

----- Original Message ----- 
From: "luis jure" 
To: 
Sent: Tuesday, May 27, 2008 9:23 PM
Subject: [Csnd] Re: Re: Re: Re: Re: Re: spatialisation for headphones


El Tue, 27 May 2008 09:27:40 -0400
"Art Hunkins"  escribió:

> In your opinion, which opcodes, in and of themselves, give the truest
> localization in stereo or quad?

i haven't been able yet to set up a system beyond a simple two-channels
stereo, can't comment on quad.

locsig gives good panning in stereo, and it can also generate quad
output. but it is strictly intensity panning, no doppler. good only for
fixed or slowly moving sources. the best from what i tested is spat3d,
because it uses a physical model of the movement, which gives a
realistic doppler effect, at least for my ears. another plus of spat3d
is that the ouput is in ambisonics format, and IMO that is the way to
go. once the signal is coded in ambisonics, you can decode it into any
speaker disposition you want. with so many multi-channel formats and
systems available today i think we should consider the different
possible set-ups we could have at our disposal to play our pieces.
that's why i think opcodes using ambisonics, like spat3d, are very
convenient. (disclaimer: i know _very_ little about the theory behind
ambisonics)

> Is the answer different for headphones and speakers?

certainly, but i'm not very interested in headphones so i never
experimented with that. perhaps the hrtf related opcodes?

best,

lj


Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe 
csound"= 

El Tue, 27 May 2008 22:35:20 -0400
"Art Hunkins"  escribió:


> However, I've used bformenc instead of spat3d. Assoming I'm using
> free field spatialization (no room acoustics or doppler), do you see
> any advantage to spat3d?


mmm... i'm not sure i understand the kord0 and kord1 arguments in
bformenc. i don't know much about ambisonics, and the manual page is
rather terse. i tried a simple example and i got only 0s. then i
compiled the example in the manual, and i got 0s also. is the opcode
broken? could anyone provide a working example, or check if the
example in the manual is working?  

There's also a model out there somewhere which offers the most  
interesting (IMO) headphone option.
It permits the encoding of signals into a virtual ambisonic space and  
then reflects it back into HRTF encoding for listening.

It was a Masters thesis at York (UK), as I remember. Anyone remember it?

David.

On 28/05/2008, at 1:29 PM, luis jure wrote:

> El Tue, 27 May 2008 22:35:20 -0400
> "Art Hunkins"  escribió:
>
>
>> However, I've used bformenc instead of spat3d. Assoming I'm using
>> free field spatialization (no room acoustics or doppler), do you see
>> any advantage to spat3d?
>
>
> mmm... i'm not sure i understand the kord0 and kord1 arguments in
> bformenc. i don't know much about ambisonics, and the manual page is
> rather terse. i tried a simple example and i got only 0s. then i
> compiled the example in the manual, and i got 0s also. is the opcode
> broken? could anyone provide a working example, or check if the
> example in the manual is working?

Yes, this paper describes the method well.
I have written an article implementing this method as a UDO. It will
hopefully appear in the next edition of the journal.


Brad Fuller-5 wrote:
> 
> 
> There is a paper entitled:
> 
> "A 3D AMBISONIC BASED BINAURAL SOUND REPRODUCTION SYSTEM"
> by
> "MARKUS NOISTERNIG, ALOIS SONTACCHI, THOMAS MUSIL, AND ROBERT HÖLDRICH"
> Institute of Electronic Music and Acoustics,
> University of Music and Dramatic Arts, Graz, Austria
> 

-- 
View this message in context: http://www.nabble.com/spatialisation-for-headphones-tp17468535p17618633.html
Sent from the Csound - General mailing list archive at Nabble.com.

The kord parameters refer to the order of the ambisonic encode...
Higher order = more accurate spatialisation.
So, the example in the menu fades from second order (better spatialisation)
to zero order.


luis jure wrote:
> 
> 
> mmm... i'm not sure i understand the kord0 and kord1 arguments in
> bformenc. 
> 

-- 
View this message in context: http://www.nabble.com/spatialisation-for-headphones-tp17468535p17618822.html
Sent from the Csound - General mailing list archive at Nabble.com.

Hi!
   Sorry, Ive been out of it for so long. Thanks for all the responses. I also 
would like to see the hrtf-based model in a csound opcode/udo. Very much so!
   Btw.: I got my results a bit better, by using another udo stereoBformat.udo. 
A bit more hilbert in there, than I used. OK: 100% more hilbert. :-)
   Kindest regards
        Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Hi again!
   One thing about the spatialisation I'm rendering at the moment. Is it usual, 
that it takes VERY long?
   Here's an excerpt from my orc:
*** cut ***
         a1, a2 soundin "t2.wav"
         a1 = a1 *9
         a2 = a2 *9
         denorm a1, a2
         aW1, aX1, aY1, aZ1 spat3d a1, 4.5, 3.5, 1.7, 0.2, 1, 4, 2, 8
         aW2, aX2, aY2, aZ2 spat3d a2, -4.5, 3.5, 1.7, 0.2, 1, 4, 2, 8
         aw = aW1 + aW2
         ax = aX1 + aX2
         ay = aY1 + aY2
         az = aZ1 + aZ2
         aL, aR stereoBformat aw, ax, ay, az
         outs aL, aR
*** cut ***
   The stereoBformat udo can be found from csounds.com and the room has 6 walls 
and the first entries in the table defining the rooms looks like this:
6, 96, -1, 0.06, .123, -1, \
   Sorry I always keep forgetting what they are, here's what I remember:
1. Number of first reflections or something, needs to be 6 if I have all walls 
, floor and ceiling.
2. late reflections (I'm not even sure spat3d reads this)
5. randomisation (random seed, I believe)
   Can anyone give me a hint if this should take roughly 45min-1hour for about 
4min of orginal audio (4min * 2, for I take stereo track).
   Kindest regards
         Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de