Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/3/84; site teddy.UUCP
Path: utzoo!decvax!decwrl!greipa!pesnta!amd!amdcad!lll-crg!seismo!harvard!talcott!panda!teddy!rdp
From: rdp@teddy.UUCP
Newsgroups: net.audio
Subject: Re: subwoofers (bass from small speakers)
Message-ID: <1517@teddy.UUCP>
Date: Mon, 28-Oct-85 18:05:49 EST
Article-I.D.: teddy.1517
Posted: Mon Oct 28 18:05:49 1985
Date-Received: Wed, 30-Oct-85 12:38:39 EST
References: <808@psivax.UUCP> <303@watmum.UUCP>
Reply-To: rdp@teddy.UUCP (Richard D. Pierce)
Distribution: na
Organization: GenRad, Inc., Concord, Mass.
Lines: 118

In article <303@watmum.UUCP> gvcormack@watmum.UUCP (Gordon V. Cormack) writes:
>> Ahem...the audio-pro B2-50 subwoofer uses (2) 5-1/4" drivers and I
>> believe it is fairly flat down to 20Hz.
>
>>  ...  other testimonial evidence
>
>I have a couple of comments.  First, speaker manufactures LIE about
>their low-end frequency response (and accuracy, for that matter).
>I don't know about audio-pro, but look for independent test results
>to verify any manufactures claims.  Polk, for example, claims the
>monitor 10b is 22Hz - 22Khz +- 2db.  Tests in SOUND revealed it
>was pretty flat down to 40, but was down >10 db at 30.  That still
>makes it a pretty decent speaker, but not as fantastic as claimed.
>

First of all, the statement should be qualified to read: "SOME speaker
manufacturers LIE about their ." Some don't.

Secondly, I am sitting here looking at a copy "High Fidelity's 1986
Audio and Video Buying Guide". Without first refering to the caveats
section usually present in such journals, one first notices that EVERY
speaker measurement shows nearly the same set of gross anomolies below
500 Hz. Well, some of the reviews go on to explain about 315 Hz dips
caused by room reflections and so forth. The point is that many of these
reviews' measurements are as flawed and misleading as some manufacturers'
data. It seems that, despite their assurances to the contrary, their
measurement technique causes common biases in the results, often swamping
meaningful differences that might be had otherwise. Also, the insistance
on 1/3 octave measurement techniques as the only response measurement 
results in to little response information being made available. Some
pretty gross (and quite audible) narrow-band effects can be completely
missed using this technique. A classic example being the Bose 901. On
1/3 octave measurements, the 901's, while not measuring terribly well,
do not exhibit the very gross anomolies that are apparent using other
techniques, for instance, a 6 db peak 1/20 octave wide where ALL the cones
have decided to go into their 1st mode rim resonance.

>Second, it isn't hard to make small speakers have excellent low-frequency
>response AT LOW OVERALL SOUND LEVELS.  But in order to get
>a decent sound pressure level out of them, small speakers would have
>to have a cone excursion of several inches.  This is nearly impossible
>to accomplish, and were it accomplished I doubt that the suspension
>would be very linear.  

In general I agree, but the number may be be a bit exaggerated. For example,
2 5 1/4" woofers have about 1/2 the radiating area of a single 12" woofer
(2 x .012 sq m vs. .05 sq m). Therefore, for a given sound pressure level
at a given frequency (requiring a certain volume-velocity), the tandem
pair of 5" woofers would have only to excurse twice the distance of the 12"
woofer. I know of no 12" woofers having a real inch of linear excursion, much
less that much voice coil/magnet. So, reversing the argument for the moment,
if all other things were equal, then a similarly based dual 5 1/4" system might
be capable of attaining levels only 3 to 6 db less than the twelve inch based
system.

Which is not to say that such systems are practical. I once (as an excersize)
built a transmission line system around the KEF B-110 woofer/midrange. It's
anechoic response was remarkably linear (within 2 db or so) down to 29 Hz,
below which it dropped like the proverbial stone (12db/oct due to compliance
controlled operation, like normal systems, and another 6 db/octave due to
line cancellation, plus another 18 db/oct for offending some engineering
god's sensibilities!). The speaker, at low levels with quite disks sounded
great! However, the presence of any warp, very low frequency disturbances,
tone-arm resonance, whatever, all conspired to cause this poor little woofer
to unmercifully beat its brains out banging against the back of the magnet.
The inclusion of a 15 Hz sharp high-pass filter improved things greatly, and
effectively raised its power handling by as much as 10 dB!.


> Even if a linear ultra-long-throw suspension
>were developed, there would still be doppler distortion of the higher
>frequencies (maybe this wouldn't be a problem in a subwoofer).

OH GOD! DOPPLER DISTORTION!!! (I think Paul Klipsch is probably old enough
to have been present at Herr Doppler's original experiment).

				:-)

>So, when you talk "frequency response", make sure you find out at
>what Sound Pressure Level.

And what measurement technique, and room conditions, and so on and so on.

>
>Here are some test pieces to use when checking out the bass response
>of a speaker system (see if you can play them at a realistic level). 
>
> Time Warp (Telarc cd-80106)  1st and 2nd cut.
> 1812 Overture (Telarc cd-80041)  12:00-15:20 in particular
> Bach Toccata (Archiv 410 999-2)
> Pink Floyd - The Wall (CBS C2K-36183) "Don't Leave me now" --
>               has some very low synthesizer tones, but they are not
>               all that loud. 


OK, What is a realistic level for synthesizer tones? What sort of SPL
does whatever organ playing whatever random Bach Toccata produce? How
loud was the orchestra playing in the 1812 overture? 

That sort of recommendation has it's own set of flaws. One experiment I
performed years ago was measuring actual performance levels of the BSO
and other orchestras, then asking people to play recordings of these
performances at "realistic" levels on stereo systems. Invariably, people
played them significantly louder in their homes than what was measured
their favorite seat in the hall. Also, the longer it had been since they had
heard the original concert, the louder their impression of "realistic" was,
reaching a plateau averaging as much as 10 db above the actual levels measured.
Another interesting effect: If the concert was simulcast, the presence of the
visual cues caused people to lower what the expected as "realistic" levels!

What is a far better recommendation is to listen to favorite recordings at
LEVELS THAT YOU FIND SATISFACTORY AND NORMAL. Your judge of what constitutes
realistics may differ greatly from this persons, my own, or anyone elses. You
will not be listening at the levels I like, you will be listening at levels 
that YOU like.


Dick Pierce