13

According to Ohm’s Law of Acoustics, the ear is phase insensitive due to its resonant structure. For example, changing the phase angle of the 3rd harmonic of a note drastically changes the shape of a waveform - but the ear cannot detect it.

If I play a note on my bass and feed it into stereo headphones as a mono signal, it centres in the middle of my head. If I invert the polarity of one channel, it changes the phase angle of each harmonic component by 180 deg. The puzzle is: I would expect the signals to cancel. Instead it appears all over my head like surround sound. Why is that? That’s why, if I practice on headphones, I invert one channel - I like it.

Update 1: If I add two electrical signals, one the inverse of the other, they cancel to give nothing. If I add two sound waves, one the inverse of the other, they should cancel. It’s this cancellation that causes beating between two close frequencies. This doesn’t happen with one such signal in each ear.

Update 2. Apologies for all the editing. Thanks to the discussion, I realise I’m confusing two issues. One is to do with the ear not detecting changes of the phase angle of harmonics. The other is to do with why the brain doesn’t cancel two sounds when one is a polarity inversion of the other. That’s the real puzzle.

P.s. I found this article about Ohm, Helmholz and Seebeck, which you need a science background to understand. My brief conclusion after a partial read, is that if the amazing signal processing abilities of the brain were realised at the time, there would have been more agreement between them. It’s the brain that creates sound, using information supplied by the ears. https://www.journals.uchicago.edu/doi/full/10.1086/710318

Brian F
  • 329
  • 1
  • 9
  • 1
    I don't know enough about this as 'science', to provide an answer [other than knowing that phase-shift is part of how you know where a sound comes from - Haas effect & all that] but this says Ohm had to give up on that idea - https://en.wikipedia.org/wiki/August_Seebeck – Tetsujin Jun 25 '21 at 10:55
  • One fact relevant to this question: **polarity isn’t phase**. You inverted the polarity, not the phase. You can’t “invert” phase. Polarity inversion is *similar* to 180° change of phase, but they are technically not the same. – Todd Wilcox Jun 25 '21 at 12:59
  • 1
    I had to look up polarity. Maybe someone else will find this helpful: _Polarity is a function of positive and negative wiring, while phase is a function of time._ Although, maybe that should be more generally "positive and negative _inversion_". Inversion has one result, phase time shift has infinite results. – Michael Curtis Jun 25 '21 at 13:55
  • 2
    @ToddWilcox "Polarity inversion is *similar* to 180° change of phase, but they are technically not the same." – Are you sure about that? Suppose that sound A is sin(x). Then suppose that you create sound B by inverting the polarity of sound A, meaning that sound B is -sin(x). Then you create sound C by doing a 180° change of phase on sound A, so that sound C is sin(x - 180°). Well, -sin(x) *is* sin(x - 180°), so sound B is the same thing as sound C. So it seems to me like polarity inversion *is* the same thing as a 180° change of phase. – Tanner Swett Jun 25 '21 at 14:11
  • @TannerSwett The thing about the math that you're doing is it's math about a sine *function* that has a domain of all real numbers. Sounds are not like that. They are transient. – Todd Wilcox Jun 25 '21 at 14:29
  • 3
    Regarding your last paragraph of the update - when you use headphones, you separate the brain's processing from the acoustics of the room (present when using speakers). You've discovered that the brain does not sum the two signals from each ear. It decodes them separately and then creates a sense of space (as best it can) based on correlation or decorrelation between the two signals. If you want to test *phase* sensitivity in human hearing to verify Ohm, the only scientific way is to alter the phases of two signals presented to **one ear only**. – Todd Wilcox Jun 25 '21 at 14:43
  • 1
    @ToddWilcox I think the same goes for finite-time sounds, though. You can take any real-life sound and approximate it arbitrarily well as a sum of lots of sine waves of various frequencies. Then, if you phase-shift all those sine waves by 180°, the result is the polarity inversion of the original sound. That said, I admit that if you have a sound which just consists of a single pulse (an increase in pressure followed by a decrease in pressure), it's not really clear that the idea of "phase-shift that pulse by 180°" really makes sense. – Tanner Swett Jun 25 '21 at 15:04
  • I could also imagine that it might be possible to think of an example of a sound where phase-shifting it by 180° would be something different than inverting the polarity. – Tanner Swett Jun 25 '21 at 15:04
  • @TannerSwett That "single pulse" thing you mentioned is happening all the time with sounds created by musical instruments and voices. There's no steady state in real world music. A thought or real world experiment you might conduct is to consider a monophonic audio recording of music on a track in a DAW. Duplicate that track to two other tracks. Invert the polarity of one duplicate and mute the other and play back - there should be no sound at all. Now mute the inverted track and unmute the other dupe and try to phase rotate it to totally cancel out the original. It won't happen. – Todd Wilcox Jun 25 '21 at 17:11
  • If you add two sound waves, one the inverse of the other, they _do_ cancel. But that's a physical effect when the two waves go through the same medium, the peak of one wave 'fills in' the trough of the other, see any [wave interference](https://youtu.be/Vma1gALd-Xw) physics experiment. The ears are different sets of sensors and capture independent waves in your example, the brain is just doing processing on the data. Just like there is no reason inverse electric signals should cancel if they go through different cables. That's up to the processing on the receiver end. – j-g-faustus Jun 25 '21 at 17:13
  • @ToddWilcox I guess it depends on exactly what you mean by "phase shift." Maybe there are multiple reasonable definitions. Suppose I have a recording of a guitar string plucked after tuning it to 400 Hz. If I use a Fourier transform or something to phase shift every sine wave in that signal by 180°, the result will be the same as just inverting the signal. On the other hand, if I instead shift the recording forwards or backwards by 1.25 ms (half the period of the note), then, as you point out, that will produce something very different. – Tanner Swett Jun 25 '21 at 19:06
  • 1
    @TannerSwett if you shift the signal by 1.25 ms, you're shifting the fundamental by 180 degrees, second harmonic by 360 degrees, third harmonic by 540 degrees and so on. – ojs Jun 25 '21 at 19:54

4 Answers4

20

You don't notice the phase of a particular note that arrives in one ear. But the brain can detect phase differences between your ears. That's the main mehod your brain uses to work out which direction a sound is coming from. If a sound is coming from the left, it will arrive earlier in the left ear than the right, and there will be a phase difference.

Inverting the polarity of one stereo channel gives the brain no way to work out where the sound is coming from, which is why it can sound odd. To me, it sounds strangely hollow.

Todd Wilcox
  • 53,618
  • 6
  • 108
  • 185
Simon B
  • 1,607
  • 10
  • 13
  • 1
    It comes from 'everywhere except where it could' so your brain turns to mush trying to work it out. It's physically impossible in real life, so you cannot work it out. – Tetsujin Jun 25 '21 at 12:40
  • Todd and Simon, isn’t it time difference or delay that the ear detects and uses for location. I’m now reluctant to use phase delay because of different interpretations :-) I’m an electrical engineer. – Brian F Jun 25 '21 at 14:25
  • 1
    @BrianF As an EE, I think you have a leg up on audio engineering and acoustics, but there are some important differences. First, I suggest letting go of the steady state when it comes to audio. Even though the time scales can be very long relative to the wavelengths, considering audio (in electrical or acoustic form) as transient is helpful in distinguishing idealized theory from practical understanding. Also, our ears are not scientific devices, so at least a bit of research into auditory anatomy and psychoacoustics might help a lot. – Todd Wilcox Jun 25 '21 at 14:39
  • Todd, you’re right about studying the ear, but I have looked at it a little. I’m fascinated how the brain substitutes a missing fundamental from information on the harmonics. Ohms Fourier Analysis model of the ear seems logical to me, as it fits the way the ear is constructed with resonant cilia. The brain does the rest. Being a bass player, I know how transients from plucking a string make the difference between a double bass and an electric bass. But whatever the waveform, inverting the polarity of a mono signal in one ear does not result in cancellation. An interesting puzzle. – Brian F Jun 25 '21 at 15:07
  • 3
    Do you have sources/references for this? – Mateen Ulhaq Jun 26 '21 at 05:17
  • Mateen, if it's my original post you are referring to, I discovered the effect of a polarity inversion in one ear when I was wiring up a homemade headphone amp. A mono signal in the centre of your head can get a bit boring. Interestingly, in the ‘60s, creating fake surround sound effects from 2 channels could be achieved by wiring the rear speakers in series, and across the L and R amplifier live outputs. So one rear speaker had L-R, and the other R-L. – Brian F Jun 26 '21 at 11:07
  • 1
    When you say "it sounds strangely hollow," do you mean hollow in a metaphorical sense, meaning it sounds empty and has no substance? Or do you mean that it sounds like something that you might hear while standing in a hollow tunnel and the sounds are echoing all around the walls and you can't tell where they are coming from? – Ben Miller - Remember Monica Jun 26 '21 at 13:24
  • @BenMiller-RememberMonica more the former. It's hard to find good words to describe an experience that couldn't be replicated in normal life. – Simon B Jun 26 '21 at 18:54
15

I used to be a sensory neuroscientist, and I teach systems physiology. The ear is phase sensitive at low frequency, which is why inter aural time delay is the dominant mode of horizontal auditory localization at low frequency. Hair cell dynamics are too slow for phase sensitivity at high frequency, where interaural intensity differences, caused by the head's acoustic shadow, dominate localization.

Scott Seidman
  • 261
  • 1
  • 5
  • 1
    I agree, the ear is phase delay sensitive, meaning the timing delay between wave fronts. Ohm was really considering the effect of shifting the phase angle of a harmonic sine wave. I think these different interpretations have caused some confusion. I still have some confusion, but putting it down to the fact that no-one really knows how the brain analyses sound waves, or do they? – Brian F Jun 26 '21 at 11:25
8

The ears, and our neuro processing, are amazing gadgets. Yes, the phase difference between left and right contributes to directional location, but there's more than that going on. If that were the only difference, we couldn't tell , say, 30 degrees left - front from 30 degrees left- rear.
Each ear - the external part, has all sorts of ridges and stuff that lead to an incoming wavefront generating phase-shifted echoes. The phases of these echos vary depending on where in the hemisphere the sound originated. The brain then "analyzes" the whole shebang to locate the source. This is in addition to the left vs. right processing. Try these two experiments (which have been done in controlled lab settings:

  1. completely block one ear. You'll still get a pretty decent idea of the direction a sound is coming from.
  2. now "fill in" the other external ear with putty so it's a blank smooth surface. Now you lose almost all directional information.
Carl Witthoft
  • 14,656
  • 2
  • 18
  • 41
  • Very interesting. I’m concluding that the brain is an amazing gadget, and that “sound” is generated by the brain and not the ears. There’s some amazing signal processing going on there. Another puzzle is why I’m gradually losing my sense of pitch, along with my hearing - a real bugger. I blame the ear for that, and might have to switch to percussion. – Brian F Jun 25 '21 at 16:52
  • This article is interesting. If the cilia change length with hearing loss, that could be why my sense of pitch is slightly different between my 2 ears. https://www.sciencedaily.com/releases/2012/10/121002113414.htm – Brian F Jun 25 '21 at 17:22
  • 3
    @BrianF - I once had an ear infection for a week [during which I had to do live shows, bass & vocals] where one ear was quite clearly the best part of a semitone out from the other. I knew which was "correct" but by heck it was a tough couple of gigs, trying to ignore the "wrong" one. Intellectually, I knew which was which, but my internal 'brain tuner' believed both, equally. – Tetsujin Jun 25 '21 at 17:28
  • Is putting putty in your ear really advisable? – user253751 Jun 25 '21 at 21:30
  • The putty thing will impact vertical localization of sound, but not horizontal. Horizontal localization is poor with one ear, always. – Scott Seidman Jun 26 '21 at 01:43
  • @ScottSeidman Thanks for the additional info. Any references? – Carl Witthoft Jun 28 '21 at 13:39
  • I'll try to dig one up. – Scott Seidman Jun 28 '21 at 23:56
  • Hofman, P., Van Riswick, J. & Van Opstal, A. Relearning sound localization with new ears. Nat Neurosci 1, 417–421 (1998). https://doi.org/10.1038/1633 – Scott Seidman Jun 29 '21 at 00:00
  • van Opstal is an old colleague of mine, in a previous life. – Scott Seidman Jun 29 '21 at 00:02
-3

Audiophiles sometimes claim that absolute polarity can be detected - that a bass drum note sounds different whether the initial transient is a pressure wave 'blow' or 'suck'. But audiophiles claim all kinds of peculiar things. Prepare to be amazed at the products on sale here!

https://www.russandrews.com/

However, as you say, RELATIVE polarity can certainly be detected. It may not matter with which polarity the two sides of a headset or a pair of speakers are connected, but it needs to be the SAME polarity.

As to cancellation - yes, if two speakers, in a perfect room, fed opposite polarity audio to a one-eared listener, you'd get perfect cancellation. In practice, the result is reduced bass. The pragmatic test for (stereo) speaker polarity is to place the two speakers side by side and play a mono signal. Swap the wires on one speaker. Whichever way gives most bass is right. (Well, maybe not right if you believe in perception of absolute priority, but they'll both be the SAME polarity.)

Isolated separate feeds to each ear are another matter completely. The brain makes some sense out of it, but not always in a predictable way. You can create some interesting 'unreal' effects.

Laurence
  • 84,790
  • 5
  • 59
  • 178
  • 4
    What's the purpose of the link to a shopping website in the middle of this answer? – Tanner Swett Jun 25 '21 at 14:13
  • To illustrate the gullibility of some audiophiles. – Laurence Jun 25 '21 at 15:34
  • but it doesn't… it's just a shopfront. – Tetsujin Jun 25 '21 at 15:53
  • Don't you ever look under the surface? Look at some of the products. They're all snake oil. – Laurence Jun 25 '21 at 17:01
  • 4
    A couple of actual examples might have been amusing. A storefront is just three clicks past being bothered. – Tetsujin Jun 25 '21 at 17:07
  • You like instant gratification then? :-) – Laurence Jun 25 '21 at 18:51
  • 3
    I guess I didn't know what I was supposed to click on after clicking that link. I was expecting some kind of essay or blog or something and was a little confused to see a shopfront. But I'm having a thick-headed day today, so maybe it's my fault. :D I probably would have gotten the point immediately if you'd written "Look at some of the products on this website and prepare to be amazed!" – Tanner Swett Jun 25 '21 at 19:14
  • Just explore! Don't glance-and-reject. EVERYTHING in life should be an invitation to explore! You're not THAT short of time! – Laurence Jun 27 '21 at 23:08