[LAD] Attenuation of sounds in 3D space

[Lorenzo]

fons at kokkinizita.net wrote:
> On Wed, Jul 28, 2010 at 10:47:03PM +0200, JohnLM wrote:
>
>    
>> I've been studying process of spatialization a bit, and ambisonics
>> figure in there quite a bit.
>> I was wondering if there is some "data loss" if I encode ambisonic
>> format and then decode it to whatever "direct channel-to-speaker"
>> format (stereo, 5.1, others) is needed, in stead of writing to
>> different channels directly.
>>
>> To make it a bit clearer: What difference would these methods cause?
>> 1. Inputs ->  B-format ->  5.1
>> or
>> 2. Inputs ->  5.1
>>
>> Simply first method allows me to concentrate making ambisonics only,
>> and then use already existing decoders to create final output.
>>      
> If you use 2nd order AMB and a good AMB->5.1 decoder the results
> should be very good.
That's what I was referring to in my previous email on this subject, 
although I think we did the decoding "by hand" the result (from a mere 
subjective point of view was remarkable). What - empirically - struck me 
is that the 'sweet spot' seemed to be particularly big, that is if I 
moved off the centre I could still hear things pretty localised, as if 
you move in along the axis of a stage...

Lorenzo
> The combimation AMB panner + AMB->5.1 decoder
> is in fact a 5.1 panner, but quite a sophisticated one. You could
> put one of theae in each channel, but since the second part is the
> same for all it's more effficient to do it after the mixing, in
> other words use an AMB mixing bus.
>
> Compared to conventional 5.1 pairwise panning the result will be
> more even, without emphasising the speaker locations as would be
> the case otherwise. In other words, the sound will be much less
> seem to originate in the speakers.
>
> Ciao,
>
>