[sdiy] Panning (again)

[Brian Willoughby]

The brain uses different cues at different frequencies to determine panoramic position, and height is very specific to the individual.

Low frequencies have long wavelengths, and the brain uses the delay between the ears to determine the angle of the source. Since low frequencies tend to bend around obstacles with little attenuation, loudness is pretty much useless for locating low frequencies.

High frequencies have short wavelengths, and the brain can't even tell which period correlates between left and right, so time delay is useless. Fortunately, high frequencies are reliably attenuated by an obstacle - like your head - so the brain uses amplitude (loudness) differences to locate the sound source.

We're actually directionally deaf for middle frequencies with medium wavelengths. Although stereo salesmen were fond of telling you that you only need one subwoofer since we're directionally deaf at low frequencies, that's actually not a limitation of the human brain (see above). What is true is that pan-pot based "multi-mono" mixing produces a sound field that has no directional information in the low frequencies. A binaural recording - or a modern mix that uses inter channel delays - does not have this limitation (so you'll need multiple subwoofers).

Determination of the height of a sound source is much less precise than left-to-right position. Here, our brains learn the folds of our ears, with the resulting comb filtering that occurs at different angles, and interpret height from that. So, the comments about certain pan laws going "over the top of your head" when wearing cans is perhaps way more complex than simply the amplitude or attenuation. Frequency response changes - or lack thereof - can seriously affect directional placement.

I suppose it's a bit of a miss to try to correct Graham Hinton here, rather than on MuffWiggler. Oh well. Just be careful not to put too much effort into a fancy pan circuit without grabbing a book or two on the psychoacoustics of the human hearing system.

Brian Willoughby


On Dec 22, 2020, at 09:03, Neil Johnson <neil.johnson71 at gmail.com> wrote:
> Someone a while back posted a link to a post by Graham Hinton on
> MuffWiggler with some relevant information:
> https://www.muffwiggler.com/forum/viewtopic.php?p=1674967#p1674967
> 
> I'll copy the pertinent text here for those who don't frequent MW:
> 
> ---
> 
> Stereo panning is an illusion based on a loudness balance whereas the
> brain decides the panoramic position of a sound from the delay between
> arriving at the ears. That is why panning on headphones goes over the
> top of your head rather than around the front--the differential delay
> of panpots is always zero. To get a law that feels right when turning
> the knob and doesn't rush across the middle you need the BBC law. The
> balance is very delicate and a 0.5dB error can produce a 25% position
> shift. I know this because I designed the Solid State Logic
> Programmable (read VC) Panpots for the BBC in the 80s. I also designed
> their Programmable Surround Joysticks for a Lucasfilm console and the
> laws for that are equally non-obvious.