"Jim Lesurf" wrote in message
...
In article , Peter Scott

OK Wrong word. I meant the fact that you might not get as great an
increase in perceived power as you might expect.

OK. No, the ability to delay is unlikely to affect this very much - unless
the delays were so great as to disrupt the music.

Is phase (delay) not an issue in whether one speaker cancels or reinforces
another? As I understand it two speakers next to each other and in
phase should produce 3dB gain.

Peter Scott

In article , Peter Scott
wrote:

"Jim Lesurf" wrote in message
...
In article , Peter Scott

OK Wrong word. I meant the fact that you might not get as great an
increase in perceived power as you might expect.

OK. No, the ability to delay is unlikely to affect this very much -
unless the delays were so great as to disrupt the music.

Is phase (delay) not an issue in whether one speaker cancels or
reinforces another? As I understand it two speakers next to each other
and in phase should produce 3dB gain.

Alas, its not quite that simple. :-) [This response is a favourite one
for academics. ;- ]

The first point is to clarify what is meant by the phrase "next to each
other". I take this to mean "effectively equidistant from the listening
position".

Real speakers have a finite non-zero size, so the above also implies that
we can specify the point or plane from which their output seems to come
when heard from such a location.

In practice, as soon as we move to other locations the phase/time
relationship will change. Thus the result varies around the room. At high
frequencies quite small changes in location w.r.t. the speakers can affect
this. This arises due in part to the change in location altering the
relative lengths as you'd expect. However a less well-known effect is that
the nominal location of the effective source may also change. Real speakers
are unlikely to be point or plane sources in their actual radiation
behaviour and directional properties.

The results in practice will also probably be significantly affected by
reflected/reverberant sounds reaching the listening location via wall
reflections, etc.

If we ignore all the above complications and assume two speakers, each
radiating *in phase* at the same level then the result at a listening
location equidistant from both is a *6dB* rise. Not 3dB.

This is because we have to add the 'vectors' or pressure levels and the
sound power rises in free space as the square of this. Bit like double the
voltage implying four times the power.

However *averaged around the room* the level may rise by around 3dB. In
some places the rise will be more than 3dB, in others less than 3dB.
Indeed, there may well be places where the level *falls* when two speakers
are used as in some places the phase relationship may produce a partial (or
even near-total in principle!) cancellation!

This is due to the variations in phase relationship with listening
location. Can't be more specific without a lot of case-specific details.

When using a pair of speakers with a cross-over the results are much more
complex as the phase relationships vary with frequency, both due to the
crossover actions, and due to the inherent properties of the two speakers.
The result may well phase 'lead' the output from one speaker and 'lag' the
other so that the actual sum is much the same - at the normal/expected
listening locations - as if just one had been used with a flat response.
Again this depends on the details of the situation.

Hence what we get may be nothing like either 6dB or 3dB. Depends on the
details.

In practice, the speaker designer is probably trying to get a given
response, but has to worry about the user-choice of speaker and listening
locations and room acoustics all being outwith his control, and varying
from one user to another. Thus the designer has to make a 'guess' as to
what will suit enough people to make a speaker a commerical success!

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

"Jim Lesurf" wrote in message
...
In article , Peter Scott
wrote:

Hence what we get may be nothing like either 6dB or 3dB. Depends on the
details.

In practice, the speaker designer is probably trying to get a given
response, but has to worry about the user-choice of speaker and listening
locations and room acoustics all being outwith his control, and varying
from one user to another. Thus the designer has to make a 'guess' as to
what will suit enough people to make a speaker a commerical success!

Thanks for clarifying that. The more you think about it the more of a
compromise loudspeaker design is. Bit like designing a vehicle that might
run
on a road or perhaps on fields or up mountains, or maybe on water or.....
This must also be true for high-end speakers unless the designer exactly
specifies the room, speaker positions and surfaces. Going back to another
discussion perhaps there ought be a test criterion along the lines of
'performs well in a wide range of environments'? This might prove to
be a crucial, if not *the* crucial, factor for speakers.

In my question I was thinking of a simpler matter. According to Behringer
two speakers next to each other, so acting as a single point-source, would
show 3dB gain at low frequencies. I should have specified the frequency.
This region is on my mind because I'm pondering what to do about
the design for a sub-woofer. B says you need four drivers for 6dB.
Clearly this will not be the perceived gain throughout the listening
area but just that near the speakers.

Peter Scott

In article , Peter Scott
wrote:

"Jim Lesurf" wrote in message
...

[snip]


Thus the designer has to make a
'guess' as to what will suit enough people to make a speaker a
commerical success!

Thanks for clarifying that. The more you think about it the more of a
compromise loudspeaker design is. Bit like designing a vehicle that
might run on a road or perhaps on fields or up mountains, or maybe on
water or.....

Yes. FWIW When I worked at Armstrong mumble-mumble years ago I used to
get into conversations and work with a speaker designer[1] who developed
the Armstrong 602 speaker. This rapidly convinced me that designing
commercial speakers for domestic use is an absolute nightmare. Designing
amps is child's play in comparison. :-) A lot of the design decisions tend
to be based on experience and judgements of what will be most acceptable to
a given target audience of customers.

[1] Bill Perkiss. Came to Armstrong from Goodmans. Wonder what happened to
him after he left Armstrong?...

This must also be true for high-end speakers unless the designer exactly
specifies the room, speaker positions and surfaces.

The advantage of the 'amateur' is that they generally only have to develop
speakers for their own use, to suit themselves and their listening room.
The disadvantage is that they may lack to experience and knowledge and test
kit of a professional.

Going back to another discussion perhaps there ought be a test criterion
along the lines of 'performs well in a wide range of environments'? This
might prove to be a crucial, if not *the* crucial, factor for speakers.

The problem here is defining the relevant set of conditions and then
testing them all. This brings up once again that magazines and reviewers
generally simply don't have the time, money, skills, etc, to do this on a
routine basis.

In my question I was thinking of a simpler matter. According to
Behringer two speakers next to each other, so acting as a single
point-source, would show 3dB gain at low frequencies. I should have
specified the frequency. This region is on my mind because I'm pondering
what to do about the design for a sub-woofer. B says you need four
drivers for 6dB. Clearly this will not be the perceived gain throughout
the listening area but just that near the speakers.

I am not sure why they say the above. There are some other complications
which I didn't mention, though... :-)

One is that the pressure variations produced by one speaker may 'push on'
the other, thus altering its ability to move.

In general, conventional 'cone in a box' speakers have movements that are
mass-controlled. So the main force limiting their movement when driven is
that required to accellerate their mass. The air load is relatively small,
so they tend not to be affected much by the presence of a second unit.

However the same may not be the case at LF as the main force opposing
movement becomes the springiness of their support, and the air inside and
outside the cabinet.

Hence - for example - if two speakers are sharing the same box, at LF, and
the LF compliance (springiness) is mainly that of the air support, then
they might affect each other quite noticably. Even in different cabinets
they may do this.

The division between 'mass controlled' and 'compliance controlled' tends to
occur at the basic resonance frequency of the speaker system. For LF
systems this is typically well below 100Hz somewhere.

But I am not clear why this would mean the result would be a 3dB increase.
It only obviously implies that it may not be 6dB. I'd expect the results to
depend on the specific details of the situation.

Would need more information to know what B are specifically referring to
and why they say the change is 3dB. Or perhaps someone else can explain why
B say this?

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

"Jim Lesurf" wrote

If we ignore all the above complications and assume two speakers, each
radiating *in phase* at the same level then the result at a listening
location equidistant from both is a *6dB* rise. Not 3dB.

I thought that was only if they were reproducing the same signal.

"In phase" may imply that - how can they be in phase if they are reproducing
different signals - but I think some people take "in phase" to simply mean
wired correctly.

Tim

In article , Tim Martin
wrote:

"Jim Lesurf" wrote

If we ignore all the above complications and assume two speakers, each
radiating *in phase* at the same level then the result at a listening
location equidistant from both is a *6dB* rise. Not 3dB.

I thought that was only if they were reproducing the same signal.

The context of the discussion from which the above quote is snipped was
that the two speakers were being driven from the same source (signal)
albiet with a crossover that may affect the amplitude/phase relationship.
The "ignore all the above" indicated that the crossover, etc, didn't change
the phase relationship so the two units were being driven with the same
signal in time alignment.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

In polarity with a DC signal with the speaker occupying the same space,
the 6 dB differential is absolutely correct. With a complex phase
signal, which many commerically available loudspeakers being
non-coherent at best and which differ slightly between even sequential
serial numbers, the issue thus becomes significantly more complex.

This can be demonstrated with two stacked identical loudspeakers fed
identical, but with a relative polarity reversed, signal. This was one
of Don Davis's favorite tricks in his Syn-Aud-Con training sessions.

One of the easiest ways a non-technical person can evaluate a speaker
for obvious phase anomalies is to playback a full bandwidth swept
sinewave at a relatively low level. These are available on dozens of
commercial test CDs. If you hear "birdies", you have a phase problem
at that particular reproduction frequency. A birdie is immediately
obvious, sounding something akin to a Looney Tunes cartoon sound
effect. BEE-OOO-WOOP.

To the best of my experience (I've tested many), only some of the
planar electrostatic and plasma speakers have truly decent full
bandwidth phase responses over the majority of their bandwidth while
not exhibiting compensatory frequency anomalies. A few (a handful)
dynamic speakers from truly capable designers are also on the market.

This is where digital filtering (once it is fully understood) will
eventually take over the commercial signal processing market- the
theory being eventually we will be able to alter frequency without
requisite phase anomalies as happens in the analog world.

To listen to a high quality reproduction system with excellent phase
response is a truly great experience. It does transport you to a
different world, not unlike viewing a real painting masterpiece. A
picture in the book can only bring back memories of the original at
best.

That being said, the major thrust in the commercial market today is not
reproduction (including phase) accuracy- note how many commentators
marvel at higher bandwidth iPod iterations while exclaiming the virtues
of its near perfect performance. One wonders just how many Stax
headphones are actually connected to iPods in this world.

I'm sure we'll have to wait this trend out. Ten years ago 44.1 kHz
sampling on a CD was not good enough and we eventually ended up with
competing DVD Audio and SACD audio products. How an iPod sampling
(generally sourced a 44.1 CD and then heavily compressed) at 128kbps or
even 320k reaches perfection pushes the world of the illogic back to
flogiston theory. Mind you the DVD Audio bit rate is 9.6 MEGAbits per
second.

One of these days I hope I'll read a review of an iPod's technical
performance using high end test equipment. It would be great to see
bandwidth, polarity, phase, THD+Noise, S/N ratio, etc.

Yet I digress...