> Are you suggesting that you can discern something is out by 1 sample?
John,
No, that's why I said 1 per degree. I'm writing a "blog" on stereo and have
been running tests. I sort of persuaded myself I could hear a 2 sample shift
with a sharp transient test sound but that was wishful thinking, but five
samples are clearly audible at about 5 deg. The angle per time difference
stops for me at about 40 deg but still stays left or right with much larger
differences until you can hear two distinct sounds.
For a full stereo sound image, you also need volume differences and I
describe the placement of the sound elements (single birds etc) in terms of
sonels (sound pixels). A very good stereo image can resolve nine sonel
directions. (left, centre, right, half left, half right, and four more
sonels in between all those.)
In case anybody asks, volume stereo overrides time difference stereo, but
both together give the cleanest stereo image. Most studio mixes rely on
pan-potted volume stereo.
Just to complete the stereo "picture" the third component in a stereo image
is reverberation, which is sounds with little or no phase relationships
between L and R and rear L and rear R. Interestingly, some sounds which
reach the ears phase inverted sound as if they come from behind, especially
with headphones.
David Brinicombe
|