Round-yer-Head Pan: Can this work?

[H S]

Varialbender's thread reminded me of a related idea, which didn't quite work.

It's a pan effect, where the L side volume goes from 0 to max and back to 0, and the R side does the same, but the LFO's are 180 degrees out of phase.  They could be tri or sin, whatever.  So far, standard pan.  But then you have each side switch polarity each time it hits 0.  So with every cycle, it goes left-to-right in-phase, and then right-to-left out-of-phase, like it's circling around your head.

I did a mock up of the effect on Sonar, but it wasn't really distinct.  (I was hoping for psychedelics and got a thin cup of coffee.)  I think too much can happen to mess up the phase relationships before the sound gets to your ears.  Also, the out-of-phase sound doesn't pan well.

Any way to make it work?

[MartyMart]

What about trying a small delay on one "half" ?
It's going to be quite fiddly to do ( and expensive ) but my understanding is that the
Roland RSS delay's/Reverbs use Phase "delay" to position stuff outside of normal stereo
This may also work on a "round your head" pan ..... make sense ?
I was watching an episode of CSI the other day on a normal "stereo" TV and some FX
seemed to be coming from my kitchen .... way off to the right in another room !
wierd .... but a great delayed/stereo placement trick

MM.

[H S]

What about trying a small delay on one "half" ?
It's going to be quite fiddly to do ( and expensive ) but my understanding is that the
Roland RSS delay's/Reverbs use Phase "delay" to position stuff outside of normal stereo
This may also work on a "round your head" pan ..... make sense ?
I was watching an episode of CSI the other day on a normal "stereo" TV and some FX
seemed to be coming from my kitchen .... way off to the right in another room !
wierd .... but a great delayed/stereo placement trick

MM.

So you might be able to make a unity gain "pan" by adding slight delay to the left, then the right . . . cool idea. 
7 inches between ears ~ 0.5 ms.  I guess this is part of what's going on with stereo Chorus and Flange.

[Paul Perry (Frostwave)]

Since we have only two ears (at most), there is a lot of complex shit going on inside the brain working out where sounds are coming from. Including: relative delay to each ear, phase differences to each ear (which are of course frequency dependant), and relative loudness in each ear. To say nothing of reverberation clues.
The reason ears are shaped the way they are, is not just to concentrate sound as an exponential horn does, but also to do complex transformations of phase relationships. (if it was just a matter of a horn, we'd all look like Shrek).
Getting a convincing 'round your head' experience is a complex task (especially if you are only using two speakers) & one that a HUGE amount of money has been spent on.

[5150]

I just got lazy and tried taking two copies of the same track (this is recording, here, not the faintest clue how to do this live), added a bit of treble and mid to one, and bass and reverb to the other, and put them through a *very* slow tremolo where one rose in volume as the other one fell... I then worked the pan left/right aspect so that the sound 'centered' when one was full-on and the other was full-off, and where the sides were where they were about 50-50.  Good enough for me (I might post a sound clip, if loading the project in Cubase stops crashing anytime soon... bloody RAM deficit).  Give that a try, if anyone thinks of anything more complex I'm up for trying it.

Jeff

[varialbender]

I made a graph in my thread that shows something similar to what you want. They'll pan like normal with different settings that give different phase cancellations at different places.

Check my thread out if you need more explanation. Hopefully we don't end up making something too similar, but I will tell you that this pan will follow a phaser, and you can choose to send full phased (vibrato) signal to the multiplier and mix dry in after the multiplier, so the amount of phase mix is panned around. The phase cancellation stuff should be a lot of fun too. I'm just starting breadboarding. I'll let you know how it sounds.

[calpolyengineer]

this reminds me of a sound clip my roommate showed me about a year ago. Here is a link, just know that you MUST use headphones to really hear it right.

http://digg.com/technology/Holophonic_Sound_-_Audio_So_Real_It_s_Scary

I did some research afterwards and found that this has been around since the 80s and is even used on the Pink FLoyd album Final Cut (even works after an mp3 conversion). This is some pretty amazing stuff, still sends chills down my spine.

-Joe

[Paul Perry (Frostwave)]

The most convincing 'home studio' method, is to use a binaural dummy head mic, and move the head around. When you play back (preferably on phones) it feels as though the sources are moving relative to the head.
Anyone who has made a bootleg live recording with earphone mics & accidentally moved their head during recording will know what I mean

[R.G.]

Something I've done on breadboards and always intended to do a PCB for is a four-way tremolo with fancy LFO modes.

You take four amplitude modulators and drive them with a multiphase LFO. A quadrature sine plus the two inversions of the sine gives you an appropriate set of waveforms. If you put a speaker for each output in each of the four corners of a room, it makes the sound appear to be rotating around the room around you. Not too practical for most gigs, though.

The other use I had for it was as an effect morpher, putting several effects chains in parallel, and fading in one as the other faded away, kind of a very sloooowww selector switch, either automatic or manual.

[Mark Hammer]

There are two potential objectives here.  One is to mimic, precisely, the effect of having a source move around the listener in a continuous manner.  This is close, but not identical, to th effect of having the listener rotate and listen to a stationary source.  The reason why I say one is not identical to the other is that stationary sources have the same reflection pattern in the same space, and the reflection cues the listener adopts have to change dynamically, whereas the stationary listener has a fixed spatial environment and reflection cues, and must determine the position of the moving sound source dependant on known reflection parameters.  The one is computationally more demanding than the other.  Mimicking this in digital form is difficult but clearly feasible.  All you need to do ( ) is map the direct/reflected properties of the sound arriving at a binaural head mic, and turn that into a set of algorithms.  Mimicking it in the analog domain is VERY, very hard.

The second objective is a more modest one, and that is simply achieving a form of panning that goes beyond the simple side-to-side and introduces apparent movement off that axis, perhaps mimicking some forward or rearward motion.  That may be much more feasible to create in the analog domain, especially with the sorts of budgets, circuit complexity, and board sizes most of us are accustomed to working with.

So, as with anything, one starts by asking oneself "What is it that changes when sounds move around the listener?".  Certainly, relative amplitude changes, but there are also changes in inter-ear arrival times as well as the spectral content of what arrives.  Keep in mind that reflected sound is almost always of different spectral content than direct sound, since most reflective surfaces are imperfect, no matter how close they are.

I have a couple of gadgets at home that do in the analog domain what software applets like the SRS Wow system does in your Windows Media Player.  These were popular amongst audiophiles in the early 80's and went by the name of image enhancers or if you sprung the bucks for Carver equipment, a "sonic hologram".  The general manner in which they worked was as follows:
1) Derive unique left/right channel information (i.e., stuff that is ONLY on one side) by subtracting the right from the left channel signal, and the left from the right-channel signal.
2) Feed the derived information through some lowpass filtering and adjust level.
3) Crossfeed the derived content to the opposite channel with a few milliseconds delay.
4) Adjust mix level of crossfeed information.

The result of this was to create a "sound shadow".  In the "real" world, when a sound is off-axis, the closer ear receives more of the direct sound, and the opposite ear receives more of the "shadow", which is the sound reflected off surfaces near that opposite ear.  That sound will, of course, arrive a short time later (depending on distance from reflecting surfaces) at a somewhat lower amplitude than the direct sound, and with some of the high end shaved off.  Traditional studio mic-ing and mixing techniques would simply pan the content somewhere between 50/50 and 100/0 mixes to achieve stereo positioning, but clearly this doesn't mimic what happens in the real world of sound shadows.  Binaural recordings (using a dummy head with mic cartridges where the ear canals are) were able to accomplish this, but since not all recordings are or were binaural, you were SOL if your favourite artist did not decide to do a binaural recording.  Happily, the electronic synthesis of sound shadows by these image enhancers was able to achieve some very striking improvement of the stereo sound field, even for recordings that were made the old-fashioned way, decades earlier.  As a listener, once you pushed the magic button, you could easily close your eyes and imagine that the drums were "here", and the conga player a little to the left, with the singer standing closer to the front.  Pushing the button again to disable the effect would make the "stereo" simply vanish.  It was like all the musicians kept playing but squashed in close together, like you were taking a group photo with a lousy camera from close up.  The biggest problems with it (and reasons why it never gained widespread acceptance) were that the intensity of the effect depended fundamentally on the original mix (some tracks sounded awful) but more particularly that surface noise on tape and vinyl was almost always unique to one channel (i.e., a click or pop was never "shared" by both channels equally).  That second aspect meant that finding unique channel info and cross-feeding simply increased the overall amplitude of impulse noise and hiss because now it was on BOTH channels instead of one.  Bluntly put, if you didn't keep scrupulous care of your records and tapes, this effect could make them irritatingly more noisy as it improved the stereo spread.

What can we learn from this digression into audio-tech history?  For one thing, apparent movement outside the strict left-to-right axis is likely going to require some adjustment of time/phase, something *other* than hard-panning, and some form of tone adjustment.  So, for instance, as a sound that WAS formerly pointed directly at your left ear moves behind you, what happens?  First, there is a slight rolling off of the high end to the left ear.  Second, the relative loudness (how much louder the one is than the other) of what reaches the left and right ear changes.  Third, the difference in delay time between the two ears changes, and perhaps the relative phase difference of different parts of the spectrum changes as well.  Note that, if we assume a fixed rather than infinite listening space, there will always be some reflected sound reaching the opposite ear.  So, a better emulation of sound-around will NEVER result in a 100/0 mix (hard pan), and only sporadically achieve a perfect 50/50 balance with equal tonal parameters on each side.

There.  That's a start.  I'm guessing here that the "basic" version will sync a) a panning circuit, b) a phase-shift circuit, and c) a stereo voltage-controlled lowpass filter circuit to the same master LFO.  If you're up to that, we can do this together.  But you have to admit, it's pretty daunting as an analog effect.  You've seen how tricky it is to mimic a rotating speaker.  In that case, the listener is stationary, and while the sound source moves, it is not changing its location relative to the room, merely how it exploits the reflective properties of the same place in the room.  That's actually a simpler task, and even THAT is pretty tricky.  This is a little harder still.