Why does stereo track sound funny if you invert one channel?

[Starkiller2]

Hello. I'm new to the forum and I'm curious about something.
Many people are aware that if you listen with headphones to a stereo track that has one of its channels inverted it sounds weird (I wasn't), but no one has said why exactly.
So I'd like to ask why does it sound so uncomfortable when you do this?
Thank you.

[steve]

Inverting a track reverses the "phase" of the audio (http://en.wikipedia.org/wiki/Phase_%28w ... hase_shift)
The relative phase of what the left ear hears and what the right ear hears is one of the mechanisms by which we gain positional information about the source of the sound.
Reversing the phase of one channel causes an unnatural shift in the phase which interferes with that sense of "position".

As a simple example,
1) Get an audio track in Audacity and make it mono (Tracks > Stereo Track to Mono), or just find a mono track.
2) Duplicate the track (Select the track then Ctrl+D)
3) Select half of one of the tracks and apply the "Invert" effect.
4) Using the track dropdown menu (http://manual.audacityteam.org/o/man/au ... _menu.html) of the upper track, select "Make Stereo Track".

You now have a 2 channel track in which half of the track has the same audio "in phase" in both channels, and half of the track with the same audio "out of phase" in each channel.

Now listen to the track with headphones.
Notice how the "in phase" section sounds as if the audio is in the middle of your head.
Notice how the "out of phase" section sounds as if the audio is outside of your head but you can't tell "where" it is coming from.

[Starkiller2]

Thank you, Steve.
The example you gave is what I did and brought me here. I've also seen you can achieve a very similar effect if instead of inverting one copy, you delay it by a few milliseconds, say no more than 6.

The relative phase of what the left ear hears and what the right ear hears is one of the mechanisms by which we gain positional information about the source of the sound

That refers to our brain being able to tell where a sound comes from based on how much later it arrived to one ear than to the other, right?

But when you invert the one of two identical tracks (e.g. your example) you're not delaying it. If you delayed it even a fraction of a second, they wouldn't cancel each other. All you're doing is making the positive negative and the negative positive, right? Isn't this what people call "polarity"? But then how can our ears/brain detect it? And how can it be useful to tell where a sound comes from when there are only two possibilities? (either the two tracks have the same polarities or one of them is inverted)

[flynwill]

For a continuous tone reversing it's polarity and delaying it a 1/2 cycle produces the same result. For a complex signal made of many frequencies inverting it is like creating a frequency dependent delay (higher frequencies are delayed less).

What your ears detect is the difference in phase. Remember that you ear's detectors are tiny hairs tuned to particular frequencies. Exactly how your brain manages to use the signals from these to produce an estimation of the direction a sound I have no idea, except that it must involve a lot of processing...

I expect that if you repeat the experiment carefully you will find that delay vs inverting of one channel do produce somewhat different effects.

[kozikowski]

This is a 39 second Left-Right sound test I shot a while ago. In the fourth segment I flip the phase of one side over and that's all I did. I shot it with one microphone and the whole show is two track mono.

http://www.kozco.com/tech/LRMonoPhase4.wav

This test is used to find sound systems that don't handle stereo sound correctly. This is a copy of the text that normally goes along with the sound file.


LRMonoPhase4.wav (Left, Right, In-Phase, Out-Of-Phase Test, 48k/16, WAV 7.3MB, 39 seconds)
Simple but for the end.
The last segment is intentionally damaged by flopping
the electrical phase of only one side.
I may sound like I'm in a deep hole
or possibly coming from behind you. On
mono sound systems I may vanish completely.
You're in trouble if the wrong segment does that.
Concert Hall Simulators may produce weird effects on
all the segments.

This clip isn't complicated. It's me, an RCA BK-5b ribbon
microphone, an FP-24 mixer, and a Mac PowerBook.


Koz

[kozikowski]

If you do delay instead of phase cancellation, mono sound systems may react very badly. Instead of just canceling out, you will be the vector sums of all the sounds. It might be cool. You'll get comb filtering effects.

Koz

[Starkiller2]

I'm really quite sad and disappointed I got absolutely no notification about your replies, guys. I thought they would be sent by default since this is a topic I Started Myself. What setting do I have to change in order to start getting email notifications every time someone replies to threads I created?

For a continuous tone reversing it's polarity and delaying it a 1/2 cycle produces the same result.

Do you mean that if I generate a tone on a mono track, duplicate the track and delay the duplicated track half a cycle they will cancel each other just like inverting one of them does? If so, how can I delay a track exactly half a cycle in audacity?

In the fourth segment I flip the phase of one side over and that's all I did.

Very interesting as well. How does our brain detect that the phase of one of the channels is inverted?

[Gale Andrews]

I'm really quite sad and disappointed I got absolutely no notification about your replies, guys. I thought they would be sent by default since this is a topic I Started Myself. What setting do I have to change in order to start getting email notifications every time someone replies to threads I created?

Reply notification is sent by default.

Either the notification went to your e-mail spam (add any address @audacityteam.org to your e-mail whitelist) or you unsubscribed the topic by accident.

If you look at the bottom of this topic it will say "Subscribe topic" or "Unsubscribe topic". If it says "Subscribe topic", click it to subscribe.

Gale