274

Every once in a while, my eighth-inch audio jack will slip loose and I'll seemingly lose only the voice part of a track -- leaving somewhat of a "karaoke" version. What I would guess about how audio plugs work suggests that I'd be making this up; however, I've asked and others tell me they've experienced this as well.

What causes this stripped vocals from audio when a 1/8" audio jack is partially unplugged?

clabacchio
  • 13,481
  • 4
  • 40
  • 80
editor
  • 2,481
  • 2
  • 13
  • 7

3 Answers3

435

When the plug starts to slip out of the jack, very often it's the ground contact (sleeve) that breaks its connection first, leaving the two "hot" leads (left and right, tip and ring) still connected.

With the ground open like this, both earpieces still get a signal, but now it's the "difference" signal between the left and right channels; any signal that is in-phase in both channels cancels out.

Recording engineers tend to place the lead vocal signal right in the middle of the stereo image, so that's just one example of an in-phase signal that disappears when you're listening to the difference signal.

Dave Tweed
  • 168,369
  • 17
  • 228
  • 393
  • 13
    Yes, for some reason, this answer really struck a chord (pardon the pun :-) with a lot of visitors. Totally unexpected ... – Dave Tweed Oct 03 '12 at 15:12
  • 28
    Its because you made hacker news: http://news.ycombinator.com/item?id=4607895 – bwawok Oct 03 '12 at 17:46
  • Here's an example http://blog.afandian.com/2012/04/almost-identifying-the-music-in-a-bbc-trailer/ – Joe Oct 03 '12 at 17:47
  • Why does disconnecting the ground mean that the earpieces now get a difference signal? – Chris Taylor Oct 03 '12 at 20:54
  • Wow, that was faster than I thought. This question also made the Stack Exchange "Hot questions" list: http://stackexchange.com/questions – Chris Laplante Oct 04 '12 at 00:41
  • 4
    Could this effect be used to hide a message in a stereo recording that isn't audible in either individual track, or even when both are listened to -- but does become audible when in this open-ground, cancelling-to-'difference' state? – gojomo Oct 04 '12 at 00:41
  • 3
    @gojomo: Not hide _completely_, but sure, you could make the difference signal really quiet and mix it with something loud and panned exactly to the center to drown it out when listened to normally. – Ilmari Karonen Oct 04 '12 at 10:15
  • 3
    @gojomo: Simple surround sound systems send anti-phase audio out the rear channel. Dolby Pro Logic(R) brand sound systems do some stuff that's a bit more complicated than that, but many surround-formatted videos will work reasonably well if one simply sends anti-phase audio to the rear channel, and that's what some cheap surround systems do. – supercat Feb 25 '13 at 23:25
  • 2
    @gojomo I know the comment is old but expanding on Ilmari's reply: The "hiding" would only be in that you have something that is quiet relative to the normal audio, and then rely on the normal audio becoming quieter than the hidden audio in a difference signal. You couldn't make it completely inaudible, as opposite phase to your left and right ear doesn't "cancel out" the sound - you'd still hear it but its perceived location would change. So hiding a signal this way depends entirely on the ability to make the hidden:normal ratio go to a perceptable level in a difference signal. – Jason C Jul 17 '14 at 15:23
  • (You also couldn't make the "hidden" signal become any more than twice the amplitude of its loudest channel when in "difference state"; most of this relies on the normal audio becoming disproportionately quieter.) – Jason C Jul 17 '14 at 15:25
  • if you're using a cable with 3.5mm plugs on both ends and an battery powered source. (eg: to connect powered speakers to an mp3 player) and the destination end half is unplugged to a you'll see the difference signal on the left channel and nothing on the right channel. – Jasen Слава Україні May 15 '16 at 03:16
  • 1
    @JasonC Interestingly, although it is technically possible to use phase differences to localise sounds, above around 1000Hz the head becomes bigger than half the wavelength, so the brain doesnt use phase difference there, only delays and amplitude differences. – BeB00 Jun 27 '16 at 00:50
  • @BWalker that's interesting but time delay above 1kHz has an equivalent as phase shift, but it may be many cycles, so the envelope delay is what is you mean or not? – Tony Stewart EE75 Oct 01 '16 at 13:18
96

The vocal track, particularly if it is just one singer, is usually found in the centre. This means it is mixed equally into left and right.

If you produce a difference signal, L-R or R-L, then this common mode material (anything mixed in equal proportions to left and right) will be attenuated.

Such a situation can happen in headphones if you break the ground connection.

Note that the headphone jack has only three conductors (tip, ring, sleeve). So the headphones share a common return path, or ground.

If this ground is not properly connected to the player, it still remains properly connected to both headphones through the jack. The headphones then form a series circuit: left amplifier output, left headphone, common ground, right headphone, right amplifier output.

What you're hearing in the headphones then is the voltage difference between the amps. Any component of the signal which is common mode (mixed into both channels equally) is suppressed. (If the amplifiers produced exactly the same signal, then the difference would be zero!)

So vocals in the center, and other things that are panned in the center such as (typically) bass guitar and kick drum, are faintly audible or not at all.

You hear a signal that lacks bass and in which the vocals are faint and distant.

But the reverb on the vocals may sound huge, because it is a stereo effect with a differing left and right signal! It may sound like the reverb mix is much more "wet" with respect to a tiny "dry" vocal signal.

Additional notes:

Why can the amplifiers work without a ground? Because each amplifier can regard the other as a ground, so to speak. A voltage amplifier has a low output impedance. One amplifier's output can serve as the ground or return path for another amplifier's output and vice versa. This is the basis for amplifier bridging. The main point is that the connection from one amplifier to the other is a complete circuit; lifting the headphone ground does not interrupt the circuit.

This type of connection between two amplifiers is exploited to bring about bridging. But bridging requires that one of the amplifiers receives an inverted signal, so that their difference is really addition! Bridging is a technique of using two weaker amplifiers to make a single more powerful amplifier. Bridging also allows an amplifier to be DC-coupled to the speaker, even if it is based on a single voltage supply (meaning that no coupling capacitor is required in series with the speaker to block DC). The technique is used in some small audio amplifier IC's that run off a single supply, but in the pro audio world, large stereo amplifiers sometimes support a bridging configuration. An important parameter of a stereo amp (to some users) is whether or not it can be easily "bridged mono" for more power, or driving of smaller impedance loads. So, what you've done with your jack is essentially bridged the left and right amps, except they have somewhat different signals, and one is not inverted with respect to the other!

Kaz
  • 19,838
  • 1
  • 39
  • 82
16

If you look carefully at the jackplug, you'll likely see three (or more) contacts. These are ground (shared), left and right.

I would guess that the vocals are on only one of the stereo tracks (ie. left or right but not both). When the plug comes partially out, you're getting mono.

Bruno Ferreira
  • 4,470
  • 2
  • 25
  • 36
Toby Jaffey
  • 28,796
  • 19
  • 96
  • 150