Frequently asked questions
Table of Contents
- How do I Purchase
Speak?
- Compared to other programs
- Polar response
- Why do I need to be concerned with polar response?
- What is your upgrade policy
?
- Modeling box stuffing
How do I Purchase
Speak
?
First download and install the demo. If you
want the program "unlocked" for full capability you simply send
GedLee a check or pay through Paypal and we will send you a key
code number via E-mail. The current cost is
$299.95, including the text.
Speak is the evolution of some twenty years of Loudspeaker modeling by its
author Dr. Earl R. Geddes. It is not a classic lumped parameter Thiele-Small
model as virtually all other simulation program are. This is very
important because it allows a flexibility that cannot be obtained by lumped
parameter applications. Speak uses T-matrix technology - well know in
numerical acoustics - to perform its analysis. The T-matrix approach
allows for a far greater degree of flexibility than can be obtained
otherwise. These techniques are well documented in the textbook
Audio Transducers which accompanies the program.
More and more designers are coming to the same conclusion that we came to
decades ago, that polar response is a critical factor in perceived sound
quality. Only SPEAK has polar response capabilities which are not
based on - inaccurate - piston assumptions. The array capabilities are also
unmatched, as well as the waveguide modeling capability.
Loudspeakers, or at least most of them, are not flat pistons. The
cone cavity has a very pronounced effect on the polar radiation pattern of
these devices. The waves radiating from the speaker cone tend to get
focused towards the center of the cone cavity creating a much larger amplitude
of sound in the middle of the aperture than at the edges. Resonance can
even occur across this aperture. This means that a typical cone
loudspeaker has a much wider radiation pattern than a flat piston model would
predict. Measurements bear this out. This effect is modeled in Speak
with the "cone depth" parameter. And even more
accurate polar response can be obtained with an actual cone model.
With horns/waveguides this issue is even more pronounced. The mouth of a horn
has a decidedly non-flat wave front and cannot be modeled as a piston.
Only Speak
correctly handles this complex situation.
The complete polar response is also known as the power response because the
integrated spatial response yields power. In the steady state, such as
the reverberant field it is the power response which one hears. Further
if a system is equalized using steady state methods then it is the power
response which has been equalized and not the axial response. Only when
a smooth and controlled power response, in conjunction with a smooth and
controlled axial response is obtained does one have a correct transient and reverberant
sound field.
Many loudspeaker simulations calculate the axial field and some even
"optimize" the crossover, etc. for a flat axial field, however this
is not a satisfactory approach since no attempt is made to correct for polar
response irregularities in the system design.
Speak
was written to specifically address this shortfall. Optimizing both the
axial field and the polar field simultaneously is far more complex than simply
optimizing the axial response. It is felt that it is better to be able to
calculate the polar response, as Speak
does, than to simply optimize the
axial response. Optimizing both is a task that has not yet been implemented
in any application.
Right now the policy is free upgrades, but that is going to change with
Version 3.0. The base program will upgrade for $100.00.
Will Speak "model" the effects of various
types of enclosure stuffing? As you know stuffing fill can have the effect
of increasing the "apparent" size of the enclosures volume with
Sealed drivers, and does other thing's with ported drivers.
The "effects" that you talked about
can in fact be modeled very effectively with Speak, albeit not in an automated
way. If a box is stuffed, loosely although
filled, the enclosure volume is found to increase by about 30% and the
"Internal Resistance" value goes to about 4 to 8 Kilo-ohms (acoustical). The box volume increase must
be done by hand and you must
also enter in the resistance value. The effects that you describe are then modeled very
accurately. An extensive discussion of these effects and how to model them
can be found in the article "In search of
parasitic parameters".
