point of origin of sound in a speaker

[Robert McLean]

This is a question that I am sure must be covered in the basic theory of speakers, but I havent seen it discussed anywhere, at least not in a form I recognize. Basically, does the sound come from the apex of the cone, ie voice coil, or the outside of cone, ie perimeter, or all the points in between ?

It is said that to get sound from two drivers to be in phase, lets say from a woofer and a tweeter, that the voice coils should be vertically aligned, that way the sound from each reaches the listener at the same time. Thus mounting each on a flat panel baffle is not accurate enough since the woofer is deeper than the tweeter, and sound from the woofer voice coil has farther to travel.

But that assumes that the sound is originating at the voice coil of each driver. It seems to me however that if the cone is moving as a piston, which I believe is what is strived for in a well designed driver, then the linear displacement of the cone is everywhere the same. So there is in fact more air dispaced at the outer edges of the cone then at the apex (voice coil end) of the cone, due to the larger circumference.

So why is it assumed that it is the voice coils that must be aligned ? If more sound comes from the outer edges should it not be the outer edges that should be aligned, which is what happens naturally on a flat baffle ?

Instead, is there some sort of center of gravity so to speak for the speaker sound, ie an effective point of origin somewhere between the voice coil and the outer edge ? and this is the point that should be aligned ?

Or is it the case that the sound is distributed, some coming from the voice coil end, some from the perimeter, and some from all points in between, and if so, would it not be the case that a narrow pulse of sound from the speaker will be recieved at a point a few meters away as spread out over a length of time equal to the time it takes for the sound to go from the voice coil to the perimeter of the speaker ? ( ie about .4mS for a 6" deep speaker by my calculation ) Ie sound from the edge gets there first, then .4 mS later the sound from the voice coil gets there ?

And would that mean that trying to phase align the drivers is futile since the sound is spread in out over time any way?

Or have I missed the point completely ?

[richie00boy]

The cone flexes to move the air, different parts of the cone will flex depending on the frequency. Generally the higher frequencies tend to come from towards the centre.

So from the centre right to the circumference is the short answer.

[MaVo]

wavelength is relevant. at 1000hz you have a 34cm long wave. at 100hz it is 3,4 meters long.

other thing is time alignment. woofers have inherent delay additional to their deeper cone, which can be quite big, depending on the woofer and enclosure. in a ms the sound travels 34cm, with a group delay of about 2-5 ms for closed boxes in the bass frequencies, the woofer will be virtually 1-2 meter behind the box.

are you speaking about aligning the drivers on the baffle of the speaker (which is more for aesthetic purposes), or in the depth dimension (which is time alignment)?

[thermal runaway]

Originally posted by richie00boy
The cone flexes to move the air, different parts of the cone will flex depending on the frequency. Generally the higher frequencies tend to come from towards the centre.

So from the centre right to the circumference is the short answer.

I would disagree with that , speaker designers strive to prevent cone breakup at all costs.

Speakers produce sound by moving air not cone distortion.

[MaVo]

check the enabl thread, there you will see that speaker designers love to drive cones into the breakup modes.

so much about generalisations

[Ron E]

The ear is insensitive to time alignment. We cannot hear waveshapes, only spectrum. If the phase or time alignment issue causes peaks or dips in response, you will hear it.

[BudP]

Ron E,

Are you saying that sound does not have a structure? That there are no internal time points, within a waveform? Or are you saying that it requires an ear/brain correlation, to extract meaning out of random frequencies, that have no meaningful structure, or information informed coherence and that are sprayed out in an indiscriminate fashion?

Bud

[MaVo]

Originally posted by Ron E
The ear is insensitive to time alignment. We cannot hear waveshapes, only spectrum. If the phase or time alignment issue causes peaks or dips in response, you will hear it.

music is all about time! delay a part of the waveform for a millisecond and the change is minimal. delay it for a second and you will see the problem. only with non transient signals like sine waves and continous noise, time is irrelevant. with transient material, time is crucial.

[roddyama]

Originally posted by Ron E
If the phase or time alignment issue causes peaks or dips in response, you will hear it.

This is exactly why time alignment is important, because it will cause peaks and dips in what is heard. However, just locating the so called "origin" of the sound wave is not the only factor to be considered. Assuming you can find the origin (not an easy task) you must also consider the group delays. And not just those caused by the cone breakup patterns and effects of defraction and dispersion but the crossovers and any other frequency dependent element in the sound reproduction path. The goal is not to obtain time alignment just to create the "perfect waveform, but to eliminate the peaks and dips refered to by Ron.

[purplepeople]

"We cannot hear waveshapes..."

I reject that wholeheartedly. In fact, I'd say that Bob Moog, Ray Kurzweil and Dave Smith would also disagree with you. The only reason we can have sound synthesis is because our ears can distinguish small changes in waveform.

:)ensen.