Sound waves and smells

[GregD]

I've just got round to wiring up the P3 OOP. I've never heard OOP before, so didn't quite know what to expect.

I selected Ella & Louis, a mono recording from the 1950s and pressed play. UGGHHH! It's really horrible to listen to. My brain got confused and thought Ella was inside my head at first, then I couldn't tell where she was at all. Really strange and confusing. I started to feel a bit sick too.

Well, I really can't understand how any passing visitors couldn't identify that SOMETHING was wrong in the room at Munich. It sounds totally unlike a normal hi-fi system, at least with a mono recording.

Let me now try a stereo recording...

1) Arthur Rubinstein - Beethoven Piano Sonata No.14 (RCA Living Stereo SACD)

Again impossible to locate the source of the sound apart from to say it is in front of me generally. I wasn't sure what to expect when using a stereo recording, the sound seems to be very unfocused and vague.

2) Miles Davis - Kind of Blue (Columbia SACD)

Instruments that are hard left/right are still obviously located there if lacking "solidity", but sounds that should be coming from the centre are very diffuse and impossible to locate. Miles trumpet is such a case, whereas Coltrane's tenor sax is on the left where it should be.

3) Heifitz/BSO - Beethoven Violin Concerto (RCA Living Stereo SACD)

Finally I thought I'd try a pure 2 track recording with no studio gimmickry to confuse the issue...

Hmmm, I can still tell the violins are on the left, yet Heifitz own violin seems very vague spacially. I can tell the orchestra is in front of me but it's all fuzzy and vague. No better than the Miles Davis recording really.

Conclusions

If I were at Munich, I think I would notice this effect - particularly with vocals (which are normally centrally placed). However, unfamiliar music might lead me to believe I was simply listening to some poor speakers that don't image well, possibly. But these are my first ever experiences of OOP sound and I would expect seasoned reviewers/technical journalists and industry professionals to notice it straight away - just as Pluto did.

[A.S.]

And how does that relate to the noxious smell in the room?

I don't know!

It's really simple. A (heavy) molecule that we call an odour presumably either hovers in the air, at rest (but under gravity) or it somehow attaches itself to an oxygen molecule. If so, we've seen with the hand-clap illustration that after the sound wave passes through any given volume of space, the air molecules return to their original place in the room, just where they were before the sound wave jiggled them about.

It must be true then that the reason a sound wave doesn't carry the smell is because there is no wind associated with the wave. For the smell to permeate the room there would have to be a real, progressive, irreversible motion of the molecules outwards away from the source. And we've seen that that's not the effect a sound wave has on the molecules suspended in your room.

Agree?

[GregD]

I agree Alan - the smell would be the same in each room.

Just a thought, the drive units move in AND out, does this make the air molecules move one way, then the other in synchrony with the dirve units as the wave passes them by (with no net movement overall, as shown in the hand-clap clip)?

[EricW]

It must be true then that the reason a sound wave doesn't carry the smell is because there is no wind associated with the wave. For the smell to permeate the room there would have to be a real, progressive, irreversible motion of the molecules outwards away from the source.
Agree?

That can't be entirely correct, surely? A bad smell will permeate a room even in the absence of noticeable air currents.

{Moderator's comment: obviously, a) there is no such room as an entirely draught free one and b) the point we are trying to make is that no matter how loud the speakers are playing they will not waft the odour through the room at all.}

[STHLS5]

....As I mentioned earlier, stereophony consists of mono material plus the additional information required to convey directional clues to the ear-brain of the listener. The former of these two (which we call the M component) is, by definition, fed identically to each loudspeaker. The latter (which we call the S component) is, by definition, fed to the right loudspeaker in reverse polarity compared to the left.

I am not sure if I understood you correctly. Not all stereophony effect is created by the net effect of differential phase reaching our left and right ears.

I have previously mentioned in

.

about the sound being slightly infront of the speakers. Previously, I have also mentioned that Elton John's voice in Candle in the Wind was way behind the speakers. I managed to correct them by moving the speakers closer to each other from 240cm to 164cm. So by moving the speakers I did not change the phase of the frequencies but the image moved closer towards me.

I also discoverd by chance while correcting the dip "a big W shaped response" from 65Hz to about 90Hz that by adding a thick Rockwool (4 x 4ft) on the floor in front of the speakers moved the sound from slightly "pinched" projection to the natural in between the speakers stage as described in the Tech Talk section.

In all the scenarios mentioned above, the image, sound stage and stereophonic effect changed by altering the distance between the left and right speakers and/or by reducing the reflected sound reaching our ears. So in my humble opinion phase difference alone does not determine the stereo effect, but it is a combination of intensity differential between left and right speakers and phase.

....So if we agree that the sound generated by the loudspeaker is not a gust of air, how does sound actually radiate from the speaker - or the birds high in the trees? Clue? What do barometers measure? (We hinted about this a few posts back...)>

Ahhhh... I am under pressure to get the right answer. (wink)

ST

[Pluto]

I am not sure if I understood you correctly. Not all stereophony effect is created by the net effect of differential phase reaching our left and right ears

Not necessarily differential phase - just a difference. Any and all stereo can be mathematically modelled in two ways. The more usual approach is termed AB (or sometimes XY), where A & B simply represent the signals presented to the left and right speakers respectively. This can be remodelled into the sum and difference domain where M = A + B and S = A - B. Each of these equations will be subject to a constant level reduction but we need not be concerned with this here.

This is fully reversible thus: A = M + S, B = M - S, an identical matrix to the one above. In theory you could convert from one domain to the other ad infinitum.

If you were given two signals, M & S, with the instruction to reproduce correct stereo on a pair of loudspeakers, you would have to send the M signal equally to each speaker and the S signal to the left speaker. You would also need to phase invert the S signal and send that to the right speaker.

The S component is the instant difference that results from A - B. Have a look at the applet on this page which will enable you to see the effect of adding two sine waves in different combinations of level and phase. Remember that subtraction is the same as adding, but with a 180° phase inversion!

Adding two identical (i.e. in phase) sine waves results in a 6dB increase. Phase shift one of the waves by 90° and you only get a 3dB increase. 120° is an interesting case as it results in a sine wave of the same level as the two being added, and 180° results in no output at all - total cancellation. So you can see that any phase shift at all between the two waves produces an output that is less than the best case (which we could call mono), that of the two waves being perfectly in phase.

You could use this applet to go on to explore the results of adding (or subtracting) waves of different frequencies and phase relationships but the important thing to realise for the purposes of this discussion is is that subtracting any two sine waves of non-identical frequency and/or phase will result in a signal, and it is this signal that we term "S". As long as an S signal exists, we have something that is not mono.

In real life these signals will all be vastly more complex than simple sine waves and the phase relationship between them will vary dynamically but the concept still holds good - the crucial point is that stereo signals can be illustrated in left & right (AB) terms or sum & difference (MS) terms. The former only considers levels, the latter tells us less about the real levels but more about the nature of the stereo. The greater the relative level of S to M, the wider the stereo will seem. No S at all, you get mono (A==B). All S, no M, results in mono which will be out of phase when reproduced on a stereo system. If you combined that signal electrically (to feed, say, a mono radio transmitter) you get total cancellation. Silence.

Any worthwhile stereo will contain a good balance of the M & S components. Too little S, you get something near mono. Too much, and it will tend to sound out of phase-ish until you reach the extreme case of all S, no M, at which point you have what we call "out of phase".

[A.S.]

The answer dear readers is here....

Just to conclude this line of thought, we have seen in the animation how the instant after the hand clap the pressure shock wave radiates away from the hands that generated it. That pressure wave reaches the ears and presses on the ear drum - we hear the sound of the clap.

But suppose we imagine this: If we held in our hands two halves of a container sealed together with rubber O rings. Inside the container is a perfect vacuum, a complete absence of air. Gripping the two halves we sharply pull them apart. The vacuum is now exposed to the air of the room. What happens is the reverse of when we squeeze the air of the room between our hands. Now the vacuum will create a pressure shock wave towards it as it tugs the air molecules across the room. The hand clap sends a 'positive' pressure wave into the room, the vacuum creates a 'negative' pressure wave towards it. If we listened carefully we'd hear a 'schloop' sound. But we wouldn't actually be able to tell by listening whether the direction of the wave was positive going (proceeding towards us) or negative going (air being pulled from us). The sound would be identical.

However, consider what happens if we replace the real human and his hand clap with a mono recording of the same hand clap which we are going to reproduce over a single loudspeaker corrected normally with red to +. The loudspeaker again sends out a positive going shock wave. And if we swap the connections so that red goes to the black socket? Then the reversed hand clap will initially suck the air from the vicinity of your ears right across the room.

And if we have two loudspeakers, one connected normally and one in reverse phase? What then? What do we experience sitting at the sweet spot listening to the hand-clap recording over the out-of-phase speaker pair? We experience a really weird physical sensation. One ear is bombarded with a positive-going shock wave from the normally-wired speaker which presses the listener's ear drum inwards. The other ear, simultaneously, nearer the reverse-connected speaker has the ear drum sucked outwards.

Is there any example in nature of this push-pull phenomena? Not that I can think of. What does than mean? It means that evolution has not prepared us for that experience. When our subconscious searches our inbuilt mental Experience Look-up Table there is no entry to describe the sensation of one ear being pressed and the other simultaneously being pulled. So it feels odd, unnatural and even a little frightening.

[Pluto]

So then, stereophony isn't simply differences in instrument loudness between left and right? That's what I always thought it was.

What a deep question this is!

It is of course true that a simple level difference between left and right will cause that source to appear anywhere between full left, then onto centre and continuing towards full right as the relative level is adjusted between left only, equal, right. This is precisely how an amplifier balance control works and a similar idea exists in each channel of a recording console, where the output of each individual microphone can be placed anywhere in the range between full left and full right: this control is called a "panpot" - panoramic potentiometer. However, simply "panning" mono mics into position provides, in and of itself, none of the more sophisticated directional clues which can give a recording real "life" and "sparkle" and elevate a run of the mill sound into one that excites and gives the listener a sense of "being there".

Two questions worthy of exploration:

1. Exactly how does a level difference between the two speakers of a stereo pair become interpreted by the ear-brain as a positional difference. Be warned - this is surprisingly subtle.

2. Having decided that a simple level difference between left & right will be interpreted by the ear-brain as positional information, how do we extend this idea to the perception of depth in a two channel stereo listening system?

Bear in mind that two speakers connected to two channels of signal can only convey differences of level and phase when presenting a sound, so how do we get from there to the perception of depth within the stereo "picture"?

[weaver]

In real life these signals will all be vastly more complex than simple sine waves and the phase relationship between them will vary dynamically but the concept still holds good - the crucial point is that stereo signals can be illustrated in left & right (AB) terms or sum & difference (MS) terms. The former only considers levels, the latter tells us less about the real levels but more about the nature of the stereo. The greater the relative level of S to M, the wider the stereo will seem. No S at all, you get mono (A==B). All S, no M, results in mono which will be out of phase when reproduced on a stereo system. If you combined that signal electrically (to feed, say, a mono radio transmitter) you get total cancellation. Silence.

There is what I assume must be a fundamental concept here that I am failing to get my head around and that is:

If we invert the phase of a complex signal (ie it is 180degrees out of phase) does that mean that all of the component wavelengths are 180 degrees out of phase with themselves?

The visualisation I have is based on the notion of a sound wave proceeding from a speaker down a room and then being reflected from the end wall back along itself; depending on the wavelength of the sound and the length of the room that sound will be either re-enforced or canceled or somewhere in between.

Is it correct to say that the reflected sound will be 'out of phase' with the original?

If so, then for a given length of room, the degree to which the sound is out of phase will depend on the wavelength of the sound and there will only be one wavelength that is actually 'inverted'.

Is there something basic I have missed?

[HUG-1]

Alan covered this with his explanation of push-pull. Think only about the wave generation at the speaker source.

[weaver]

Alan covered this with his explanation of push-pull. Think only about the wave generation at the speaker source.

In which case I'm afraid I must beg your indulgence for my being rather slow on the uptake.

... the vacuum will create a pressure shock wave towards it as it tugs the air molecules across the room. The hand clap sends a 'positive' pressure wave into the room, the vacuum creates a 'negative' pressure wave towards it. If we listened carefully we'd hear a 'schloop' sound. But we wouldn't actually be able to tell by listening whether the direction of the wave was positive going (proceeding towards us) or negative going (air being pulled from us). The sound would be identical.

However, consider what happens if we replace the real human and his hand clap with a mono recording of the same hand clap which we are going to reproduce over a single loudspeaker corrected normally with red to +. The loudspeaker again sends out a positive going shock wave. And if we swap the connections so that red goes to the black socket? Then the reversed hand clap will initially suck the air from the vicinity of your ears right across the room.

As was demonstrated earlier in the thread, a sound wave consists of a succession of compressions and rarefactions of the air molecules; a 'negative' pressure wave can no more 'suck' a smell across a room than a positive one can 'blow' it.

A speaker cone when at rest can either move outwards first - creating an initial compression, or inwards first - creating an initial rarefaction, what follows thereafter will always be the opposite will it not? - thereby creating the compression-rarefaction series that we hear as sound.

If we have two cones fed an identical sine wave with one moving in when the other moves out, then one will be creating compression when the other is creating rarefaction and they are 'out of phase'.

However, if one is fed a sine wave of half the wavelength of the other, then on every half cycle of the longer wavelength the cones will actually be moving in the same direction at the same time (the applet that Pluto links to in post #26 is what I'm looking at here) irrespective of whether the cones are in phase or 180 degrees out of phase.

[Pluto]

Weaver - can we please clarify what you are asking?

All stereo signals must contain a certain amount of out of phase material - they would not be stereo were this not so. The idea is to provide just enough to achieve a nice musical sense of width, but if you go too far down that road you arrive in nausea territory!

The use of static sine waves is purely as a learning aid and has little bearing on real world use of stereo. It certainly isn't worth complicating the subject by thinking about the relative phase of sound waves reflected off your walls. While such aspects might change the sound within your room by a small amount, the big picture will remain relatively unaffected. The human ear is rather good at "tuning out" unwanted reflections - perhaps because evolution has made us used to listening within an environment a few tens of cubic metres in size. Thinks - has our hearing evolved to perform optimally in the smallish spaces in which our ancestors lived?

[weaver]

It certainly isn't worth complicating the subject by thinking about the relative phase of sound waves reflected off your walls. While such aspects might change the sound within your room by a small amount, the big picture will remain relatively unaffected.

The 'reflected off walls' issue was just a way of trying to visualise (for myself) a situation where sound from a speaker would meet another version of itself that was slightly out of step.

If we have a single speaker playing, and assume a 'perfect' reflection coming back at it, is that situation comparable with a pair of speakers playing the same mono signal but out of phase?

Clearly, sound as it makes it way through the air does not exist as a set of discrete frequencies and amplitudes; they add and subtract from one another to give highly complex waveforms and I think that perhaps this is the root of my problem.

Would it be correct to say that, relative to a point source of sound that the angle of our heads (ie the distance of each ear to the source) gives rise to a difference in phase between the sound reaching each ear?

[A.S.]

...If we have a single speaker playing, and assume a 'perfect' reflection coming back at it, is that situation comparable with a pair of speakers playing the same mono signal but out of phase?...

Logically it can't be can it. A pair of speakers = two speakers at two different points in space in the listening room. However you look at it, such a pair cannot ever be considered the same sound generator as one soiltary speaker at one point in space can it?

Clearly, sound ... add and subtract from one another to give highly complex waveforms and I think that perhaps this is the root of my problem.

Are you sure about that? Superposition? Do the sound waves "know" that they are "out of phase"? Unlikely. Their phase status must be relative to the observer, be that (a pair of) human ears or a measuring microphone.

Best surely to think of a pressure wave? As I showed in my trumbone example http://www.harbeth.co.uk/usergroup/s...665#post14665?

Would it be correct to say that, relative to a point source of sound that the angle of our heads (ie the distance of each ear to the source) gives rise to a difference in phase between the sound reaching each ear?

Evolution has given us two ears on the sides of our head (not on the top like a horse) for resons to do with localising sound direction. One technique used by the brain is relative loudness left/right - the other is relative phase, another word for arrival time.

[Pluto]

If we have a single speaker playing, and assume a 'perfect' reflection coming back at it, is that situation comparable with a pair of speakers playing the same mono signal but out of phase?

No. At least, not in the real world. While there are many experiments in physics that involve light beams interfering with one another, AFAIK they deal with monochromatic light (i.e. a single fixed frequency). I suspect that the idea of sound waves doing the same across a broad range of frequencies isn't practical. It would require an extremely large reflective surface if nothing else. It might be noteworthy that digital noise cancelling techniques operate by synthesizing the effect I believe you are thinking about - adjusting relative phase at different frequencies to achieve a cancellation.

Clearly, sound as it makes it way through the air does not exist as a set of discrete frequencies and amplitudes; they add and subtract from one another to give highly complex waveforms and I think that perhaps this is the root of my problem.

Very true. Sine waves are all very well for learning concepts, but real life ain't like that!

Would it be correct to say that, relative to a point source of sound that the angle of our heads (ie the distance of each ear to the source) gives rise to a difference in phase between the sound reaching each ear?

Indeed. The principal mechanism by which we locate the position of sound sources at lower frequencies (which for these purposes means frequencies at which the half-wavelength is larger than the size of a typical head - roughly 1kHz) is essentially one of phase difference.

[Pluto]

A pair of speakers equidistant from the ears, radiating out of phase could never be compared to a single speaker plus its reflected sound...even if 100% coherent reflection at all frequencies was possible...which it isn't in the real world. Because....

...the velocity of sound varies slightlywith frequency so by the time the wave front has made the trip from a single point source, to you, to the infinite wall on which it is reflected and back again to you, the phase relationships between frequencies will be completely changed.

The out-of-phase nastiness originating from a pair of incorrectly wired speakers is 100% incoherent - a situation that simply does not exist in nature, hence it's disturbing quality.

[weaver]

Thanks to the posts above the situation is starting to become clearer.

I would like to just explore the time/phase relationship a little though.

Can we assume a 'perfect' pair of speakers being fed 'perfect' test tones and reproducing them perfectly such that what we hear is just pure single tones.

If we play a mono 100 Hz signal through both speakers and listen at the prefect sweetspot in our perfect room we should hear a tone from midway between the speakers shouldn't we.

If we then introduce a delay to the tone to one of the speakers of 1/200th of a second (ie half the period of the tone) would the speakers relative to the listener now be 180 degrees out of phase?

Assuming the answer to that is yes, we move on

Keeping the delay of 1/200th sec we move to a tone of 50Hz, the time delay is now one quarter the period of our tone so are the speakers now out of phase by 90 degrees?

Then again we move to a tone of 200 Hz but still keep the 1/200th delay; the delay is now equal to the period so are we now 360 degrees out of phase and in this particular (artificial) instance does that mean that they are effectively back in phase?

Lastly, we remove our time delay and instead we reverse the polarity (black for red) on the back of one of our speakers and the replay each of our tones - 50 Hz, 100 Hz, 200 Hz separately, what is the effect on the phase of each of these tones (relative to the perfect listener) now?

[Pluto]

Thanks to the posts above the situation is starting to become clearer.

I would like to just explore the time/phase relationship a little though.

Can we assume a 'perfect' pair of speakers being fed 'perfect' test tones and reproducing them perfectly such that what we hear is just pure single tones.

You must add something rather abstract to your list of assumptions; an infinite room to listen in. Tones do not sit well in real rooms but OK, let's continue.

If we play a mono 100 Hz signal through both speakers and listen at the prefect sweetspot in our perfect room we should hear a tone from midway between the speakers shouldn't we.

For the purposes of this discussion only, OK. In reality, pure tones are often quite hard to locate but I've never listened in a perfect room of infinite size.

If we then introduce a delay to the tone to one of the speakers of 1/200th of a second (ie half the period of the tone) would the speakers relative to the listener now be 180 degrees out of phase?

Yes. At this frequency one speaker would be moving forwards, the other backwards in a symmetrical, complementary manner.

Assuming the answer to that is yes, we move on

Keeping the delay of 1/200th sec we move to a tone of 50Hz, the time delay is now one quarter the period of our tone so are the speakers now out of phase by 90 degrees?

Yes.

Then again we move to a tone of 200 Hz but still keep the 1/200th delay; the delay is now equal to the period so are we now 360 degrees out of phase and in this particular (artificial) instance does that mean that they are effectively back in phase?

Yes.

Lastly, we remove our time delay and instead we reverse the polarity (black for red) on the back of one of our speakers and the replay each of our tones - 50 Hz, 100 Hz, 200 Hz separately, what is the effect on the phase of each of these tones (relative to the perfect listener) now?

They will all be out of phase - as will everything.

Finite time delays cause a complex patterning of boosting and cancellation across the spectrum. I suggest you explore the concept of comb filtering, particularly in the context of real room reflections and their effect on the response perceived by the listener.

A time delay may be thought of as a variable phase shift - a shift of 360° at a given frequency may be thought of as "no change" but this is untrue for a variety of reasons* - think of it as two people in a walking race in step with each other, but one of them exactly two, four, six or eight paces ahead. They may well be "in phase" but the one ahead still wins!

* Mainly that the phase relationship between two random signals based on a time delay will vary with frequency. This is why, for stereo to work properly, the timing relationship between the two channels must be maintained to within a very small magnitude.

[weaver]

There is what I assume must be a fundamental concept here that I am failing to get my head around and that is:

If we invert the phase of a complex signal (ie it is 180degrees out of phase) does that mean that all of the component wavelengths are 180 degrees out of phase with themselves?
...

Is there something basic I have missed?

Firstly, thanks again Pluto.

Next - the basic concept I had missed?

When people talk about a pair of speakers being 'out of phase' there is in fact no 'phase shift' involved.

By reversing the polarity, the signal to one speaker has been inverted or mirrored.

Conversely, shifting the phase of a music signal by 180 degrees such that it is in anti phase wrt the original does a very different thing.

Trying to reconcile these two situations on the assumption that they were the same was my mistake.

It seems that the phrase 'out of phase' can mean two entirely different things depending on whether it is applied to a 'reverse polarity' situation or a 'phase shift' situation.

If we look at simple sine waves it can appear (it did to me) that we are looking at the same thing when in fact we are not.

[Pluto]

...By reversing the polarity, the signal to one speaker has been inverted or mirrored....Conversely, shifting the phase of a music signal by 180 degrees such that it is in anti phase wrt the original does a very different thing.

I'm still not sure you have it... The statement "shifting the phase of a music signal by 180°" really has to be qualified by asking "at what frequency" because anything other than an electrical inversion is likely to have a phase shifting effect that varies with frequency.