[xine-devel] Multi channel sound positioning.

[James Courtier-D. <Ja...@su...>]

I have been thinking about getting xine to automatically adjust for the 
position of one's speakers.
So far I have worked out that 1 sample of 48khz audio is equivalent of 
about 7mm. (Using speed of sound in air=340m/s)
So, if one moves one speaker of a stereo speaker system 7mm further 
away, one could just delay the other channel by 1 sample, and the 
listener will hear no difference. Then, as the ears are about 120mm 
apart, how detailed is the human ear at perceiving time differences 
between sound reaching each ear? Should we resample to acheive 0.5 
sample resolution delay or not?
Can we use this information to add extra spacial sounds? So that the 
sound seems like it is coming from speakers much further apart than they 
currently are?

Then, what about power, If you double the distance from a speaker you 
1/4 the aplitude(volume) of the sound.
So, if you are sitting 1 meter away from your speakers, and you move one 
speaker 1 meter further away, the speaker further away will have to be 
twice as loud.
Does this relationship vary depending on the design of speaker 
(directional or not) ?

Some questions: -
1) Is the aplitude of the sound signal equal to the perceived loadness 
of the sound.
i.e. If one doubles the aplitude, does the sound actually sound twice as 
loud?
2) How does this map to a change in the mixer's volume slider?

Cheers
James