> I wonder what my pair of 15" Tannoys
> is worth!
Next to nothing, Jerry, but I'd be happy to help you out if
they are cluttering up your garage. :-)
Kidding, of course. You'd probably get a pretty penny. If
you put a price on them, I would be interested (perhaps not
capable, but interested.) For directional mic
characterization a coaxial source is imperative.
Bob
--
"Things should be described as simply as possible, but no
simpler."
A. Einstein
Reply by Ben Bradley●June 29, 20052005-06-29
On 28 Jun 2005 09:15:00 -0700, "Andor" <an2or@mailcircuit.com> wrote:
>> > Instead of getting a program to work in systray, it probably would be
>> > much simpler, sound better, be more reliable, easier to install,
>> > operating system and computer hardware independent, extremely portable,
>> > upgrade/virus resistant and only marginally more expensive if you went
>> > to your local flee market (or e-bay) and got yourself a stereo graphic
>> > equalizer.
>>
>> Or decent speakers. You can sometimes simulate a silk purse using a
>> sow's ear, but the weave is never quite right.
>>
>> Translating, while you can equalize the frequency response of a
>> peaky/dippy speaker -- you can make it dead flat with enough care -- it
>> usually makes the phase response so bad that you'll wish you hadn't.
>
>Agreed - and we haven't even discussed the problem that the "frequency
>response" from the speakers to the ear is not well defined (it changes
>chaotically as one moves around or changes the position or aim of the
>speakers). Similarly, many dips and non-linearities in of the frequency
>response of a speaker are non-equalizable (usually due to reflections
>within the speaker cabinet).
There's even more to it, speakers (especially cheap ones) have
resonances that cannot be "eq'ed out" even with a good parametric. You
may be able to get the steady-state swept sine wave response flat, but
transients (which most sounds contain) will activate the resonances
just like strinking a bell (though the resonances may not have nearly
as high a Q as a bell).
(OOTC) DSP technology can be used to find these resonances, their
frequencies, Q, and phase in relation to the exciting signals, and
generate resonances with the same frequencies, Q, and oppsite polarity
such that they cancel out in the speaker. This may actually work, but
is cost-prohibitive to do with cheap speakers (at least it has been,
perhaps no longer), but I'm sure such things have been done with
high-end hi-fi speakers (which I imagine have fewer, well-controlled
resonances that are easier to cancel out). In the OP's case, this task
could take a substantial percentage of available CPU cycles.
>But for subjective frequency response
>adjustment (boosting the highs and lows is a favourite),
It even has a name, "smily-face" EQ.
>a graphic
>equalizer will do the job (a parametric would do it in a simpler
>manner, but they require more know-how to setup and usually are more
>expensive).
>
>
>> I haven't investigated what one can do with a single FIR that covers the
>> whole range. Since I don't know it won't work, that would be one of the
>> first ways I'd try. In any case, there's no need for variability.
For actual graphic equalizer code, a series of IIR's would closely
emulate an analog EQ.
Many audio recording programs can accept EQ plug-in's and run them
in realtime to do effects on the audio input, though I don't know how
to do 'signal routing' to have another program's audio output go
through the real-time effects. Any of them would be way overkill for
the task, but one would be N-Track Studio, <http://www.fasoft.com> and
this freeware also might do it:
<http://sourceforge.net/projects/audacity/>.
>Getting a good speaker would be the proper solution, as Jon already
>mentioned. However, I doubt one can get a good pair of speakers for $20
>(which seems to be about the price of a graphic stereo equalizer on
>ebay).
>
>Regards,
>Andor
"Jerry Avins" <jya@ieee.org> wrote in message
news:bIqdnQ0N_8cFW1_fRVn-3g@rcn.net...
> Jon wrote:
>>
>>>
>>>>Getting a good speaker would be the proper solution, as Jon already
>>>>mentioned. However, I doubt one can get a good pair of speakers for $20
>>>>(which seems to be about the price of a graphic stereo equalizer on
>>>>ebay).
>>>
>>>I believe that the best speakers on can buy for $20 a pair will sound
>>>better than ones with light magnets, thin cones, and the best of equalizers.
>>>
>>>A quick look around the web tells me I'm behind the times on price.
>>>Radio Shack no longer lists the replacement speakers I had in mind. I
>>>need to double my price estimate. I wonder what my pair of 15" Tannoys
>>>is worth!
>>
>>
>> Keep in mind that since this is a PC application, the OP probably
>> wants/needs _powered_ speakers since most PCs only have a line-level
>> output. Maybe that also affects your price estimate.
>
> I'm running a pair of KLHs off my system's unamplified output. They're plenty
> loud enough on most program material, but louder would be nice.
Yeah, you can get away with driving some small speakers with a headphone or line
output, but obviously volume will be limited, and highly dependent on your
particular sound hardware's output stage. In fact, a new HP PC I got at work
recently came with some passive speakers and an output on the PC (from the
built-in sound hardware) designed to drive them to a pretty decent volume. But
I replaced that set-up with some active Harmon/Kardons I had around. I like
having a physical volume knob I can grab and turn instead of fooling around with
software adjustments. (The HP also came with a keyboard that gave you special
volume control keys to deal with that problem, but the typing action on it was
junk so I didn't want to use it.)
Reply by Jerry Avins●June 29, 20052005-06-29
Andor wrote:
> Jerry Avins wrote:
>
>>I wonder what my pair of 15" Tannoys is worth!
>
>
> Coaxial?
Yes. Very directional above 15 KHz. The on-axis beam is only about 8 dB
down at 30 KHz, but it's dead a few degrees away. Like a flashlight.
FYI: The speaker's magnetic path has two gaps. The one at the front has
the cone's voice coil, and the one at the back, the tweeter dome's. The
pole piece is the start of the tweeter's exponential horn, and the cone
continues the flare. So the tweeter horn has a 13-inch mouth (obviously
unsegmented), which makes it very directional at the high end.
Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������
Reply by Jerry Avins●June 29, 20052005-06-29
Jon wrote:
>
> Jerry Avins wrote:
>
>>Andor wrote:
>> ...
>>
>>
>>>Getting a good speaker would be the proper solution, as Jon already
>>>mentioned. However, I doubt one can get a good pair of speakers for $20
>>>(which seems to be about the price of a graphic stereo equalizer on
>>>ebay).
>>
>>I believe that the best speakers on can buy for $20 a pair will sound
>>better than ones with light magnets, thin cones, and the best of equalizers.
>>
>>A quick look around the web tells me I'm behind the times on price.
>>Radio Shack no longer lists the replacement speakers I had in mind. I
>>need to double my price estimate. I wonder what my pair of 15" Tannoys
>>is worth!
>
>
> Keep in mind that since this is a PC application, the OP probably
> wants/needs _powered_ speakers since most PCs only have a line-level
> output. Maybe that also affects your price estimate.
I'm running a pair of KLHs off my system's unamplified output. They're
plenty loud enough on most program material, but louder would be nice.
Something like http://store.qkits.com/moreinfo.cfm/fk605 running off the
computer's underutilized 12V supply would add about $15. I have a pair
of powered Altec computer speakers he can have cheap. I have another
pair in cabinets only a little larger with heavy magnets and four-inch
cones than sound much better and are nearly as loud. Louder than the
KLHs in fact, but (obviously) less clean bass.
Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������
Reply by Andor●June 29, 20052005-06-29
Jerry Avins wrote:
> I wonder what my pair of 15" Tannoys is worth!
Coaxial?
Reply by Jon●June 28, 20052005-06-28
Jerry Avins wrote:
> Andor wrote:
> ...
>
> > Getting a good speaker would be the proper solution, as Jon already
> > mentioned. However, I doubt one can get a good pair of speakers for $20
> > (which seems to be about the price of a graphic stereo equalizer on
> > ebay).
>
> I believe that the best speakers on can buy for $20 a pair will sound
> better than ones with light magnets, thin cones, and the best of equalizers.
>
> A quick look around the web tells me I'm behind the times on price.
> Radio Shack no longer lists the replacement speakers I had in mind. I
> need to double my price estimate. I wonder what my pair of 15" Tannoys
> is worth!
Keep in mind that since this is a PC application, the OP probably
wants/needs _powered_ speakers since most PCs only have a line-level
output. Maybe that also affects your price estimate.
Reply by Jerry Avins●June 28, 20052005-06-28
Bob Cain wrote:
>
>
> Jerry Avins wrote:
>
>> Translating, while you can equalize the frequency response of a
>> peaky/dippy speaker -- you can make it dead flat with enough care --
>> it usually makes the phase response so bad that you'll wish you
>> hadn't. I haven't investigated what one can do with a single FIR that
>> covers the whole range. Since I don't know it won't work, that would
>> be one of the first ways I'd try. In any case, there's no need for
>> variability.
>
>
> Except for non-linearity, which cheap speakers usually feature in
> abundance, you can do a pretty darned good job of idealizing a speaker
> if you are willing to go to the effort of measuring its impulse response
> on axis in fairly anechoic conditions (on its back, out of doors far
> from reflections), inverting the IR (time or frequency domain with care)
> and using a convolver in place of an equalizer. This can pretty
> precisely fix crossover mangling, time alignment, internal reflection
> and other linear distortions.
>
> Rooms and dispersion characteristics are another whole matter and hardly
> worth discussing. Treatment and placement are all you can realistically
> do to effect improvement and that's almost purely an empirical process.
The rub comes when the steady-state response isn't what you would
predict from the impulse response. Weak magnets decouple the voice coil
from the driving voltage, and baskets that don't approximate rigid
bodies (Hi, Fred!) (not usually as big a problem; steel is cheap enough)
allow different resonances, some with high Q that don't show up with
transients. When Alnico was king, speakers were rated in ounces. (The
Tannoy, in pounds.) Weights are less with ceramic, but still important.
Cast frames, long-throw spiders and surrounds, cone design that avoids
breakup, edge-wound voice coils; those are all embellishments. You can
hear a voice coil and magnet without a cone. Without a magnet, there's
nothing. With a weak magnet, there's very little.
Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������
Reply by Bob Cain●June 28, 20052005-06-28
Jerry Avins wrote:
> Translating, while you can equalize the frequency response of a
> peaky/dippy speaker -- you can make it dead flat with enough care -- it
> usually makes the phase response so bad that you'll wish you hadn't. I
> haven't investigated what one can do with a single FIR that covers the
> whole range. Since I don't know it won't work, that would be one of the
> first ways I'd try. In any case, there's no need for variability.
Except for non-linearity, which cheap speakers usually
feature in abundance, you can do a pretty darned good job of
idealizing a speaker if you are willing to go to the effort
of measuring its impulse response on axis in fairly anechoic
conditions (on its back, out of doors far from reflections),
inverting the IR (time or frequency domain with care) and
using a convolver in place of an equalizer. This can pretty
precisely fix crossover mangling, time alignment, internal
reflection and other linear distortions.
Rooms and dispersion characteristics are another whole
matter and hardly worth discussing. Treatment and placement
are all you can realistically do to effect improvement and
that's almost purely an empirical process.
Bob
--
"Things should be described as simply as possible, but no
simpler."
A. Einstein
Reply by Jerry Avins●June 28, 20052005-06-28
Andor wrote:
...
> Getting a good speaker would be the proper solution, as Jon already
> mentioned. However, I doubt one can get a good pair of speakers for $20
> (which seems to be about the price of a graphic stereo equalizer on
> ebay).
I believe that the best speakers on can buy for $20 a pair will sound
better than ones with light magnets, thin cones, and the best of equalizers.
A quick look around the web tells me I'm behind the times on price.
Radio Shack no longer lists the replacement speakers I had in mind. I
need to double my price estimate. I wonder what my pair of 15" Tannoys
is worth!
Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������