Does "true" stereo panning exist?

Question

I know that panning in a DAW is basically just turning the right/left channel up, making the sound pressure in one ear greater than the other. But as far as I know, in the real world, the perception of directionality doesn't come from a difference in volume as much as a tiny delay in the time of arrival from one ear to the other.

I have tried manually phase delaying one channel half a milisecond or so, and it did work to an extend (making the sound seem like it came from the left), but it made the sound thin and not very nice. I'm guessing there is more to it than that :-)

Is it possible to fake this delay in a convincing way, and if so are there any VST plugins for stereo panning that takes it into account, and perhaps other factors too that I don't know about? And what are the problems with doing so, since it doesn't seem to be very widespread? I can imagine that there's a reason no daws seem to have it as a default panning option, what am I missing?

Answer 1

First off, your understanding of the perception of directionality is a little flawed. Yes, "Interaural Time Difference" (ITD) affects our ability to localize, but so does "Interaural Level Difference" (ILD). Our brain makes use of both, because each lose effectiveness outside of certain frequency ranges. We need both mechanisms to localize across the frequency spectrum. High frequency sounds are prone to lose energy faster than lower frequency sounds. ILD helps localize high frequency sounds, and ITD helps localize low frequency sounds. It's not a hard cut crossover between the two; there's a lot of overlap. So we use both to cover a large chunk of the audible spectrum.

The problem you may be running into with using delays to pan is that each person has a unique HRTF (Head Related Transfer Function); due to the shape of the head, amount of hair, shape and position of the ears, etc.. Though I suppose that could also apply to the perception of level difference, so I don't know. Personally, I've found delays to be useful as a low dsp method of creating space (i.e. reverb). Any delay approaching 30ms will start to sound like an echo, but it can be very effective in the 5 to 20ms range. It's important to have a difference in volume between the two channels (or pairs of channels) as well. It's a useful method when you don't need a complex space, but need it to sound a little more "alive."

You might by interested in some of the things going on in "wave-field synthesis." They're using arrays of speakers to create "holographic" sound images; sounds that appear to be coming from a localized point within a three dimensional space. The theories and discussion all sound very interesting (though way over my head), and I'd love the chance to experience it. Probably way to complex for getting it into theatrical and home environments though.

Answer 2

I guess is not done often because mono compatibility concerns. Delay introduces all sort of phase issues. Mono is important in broadcast.

Answer 3

take the headphones off and everything goes back to the way it should be...

Answer 4

Perhaps it's worth trying the demo of Wave-arts Panorama 5. It uses the HRTF technique which Shaun describes plus what it calls "acoustic environment modeling". I've been meaning to give it a try myself. I think this technique is computationally too expensive to replace conventional panning on every track of your Daw, and probably less useful for music based applications. Seems like the effect is also much better with headphones.