One of the most challenging aspects of a instrument designs, is the design of transducers and in the case of a stereo equipment, the design on speakers and their enclosures.
While the engineering of the electronic part of stereo equipment certainly has presented challenges in the past, the current state of the technology including associated test equipment and especially the development of digital signal processing, makes the development of electronic stereo equipment a craft rather than an art.
Development and manufacturing of speakers require a larger investment in test equipment such as calibrated microphones, laser interferometers (to analyze vibration modes of speaker diaphragms) and even investment in sound-dead rooms.
But before anything can be build a basic design philosophy needs to be applied and in my view it should be as follows:
Stereo equipment is to re-produce the creation of the artist as truthful as possible.
So what about the acoustic characteristics of the listening room, in which the stereo equipment is used, should that be compensated for? In my opinion most often not, but in some cases it might be necessary to compensate for the worst reflections and then only through use of electronics (filtering) and not through speaker design. That room acoustics affects the performance of stereo equipment can easily be verified by listening to a car stereo while opening and closing the windows. Not compensating for room acoustics can also be justified by accepting the stereo equipment as a means to bring the concert hall into the listening room (with all its flaws). Again should the listened decide to contribute to the music as an artist it should be done through the more flexible electronics.
The traditional way of documenting the performance of stereo equipment is by showing the frequency response and in some cases the phase response. Rarely is the acoustic distortion, especially at higher volume settings, reported.
But what really is missing is an impulse response, which can reveal electronic (some filters) or acoustic (wave-guides, resonance boxes, etc.) compensations for dips in the frequency response. The most pronounced examples of prolonged impulse responses can be heard on equipment with "mega bass", boom boxes and especially cars acting as mega boom boxes. The most difficult instrument to re-produce (low frequency short duration) is probably the bass drum in a rock band. Some drummers even attenuate and shorten the sound by putting a pillow in the bass drum.
With all this said I have made a set of stereo speakers, focusing on a short impulse response, by making a small rigid airtight cabinet measuring only 9¼" x 5" x 3¾" and using two 3½" dual car speakers (Blaupunkt PCx352) connected in series. The materials used are solid ¼" birch (any other hardwood can be used) with a pegboard glued on the inside with the rough side inward to stiffen the construction, prevent warping and minimize internal acoustic reflection between the walls. Furthermore an off angle pegboard separation between the speakers is glued in to break up characteristic distances, thus minimizing internal reverberation.
They actually sound very good and I have had an opportunity to measure their electrical impedance and listen to the frequency response using a (50 ohm) function generator loaded with a 5 ohm carbon resistor (and the speaker) with and without back-plate. I believe I was successful in crating a very short impulse response judging by the reproduction of bass drums and other music with high instantaneous dynamic range, as well as the fact that very little vibration of the cabinet can be felt even at painfully high volume levels, which by the way did not distort the sound. The only draw-back is the expected roll off at lower frequencies (<150Hz), so a sub-woofer will improve on this.
![]() |
Impedance without back-plate (HP 4194a). The resonance frequency is here at about 150Hz |
![]() |
Impedance with back-plate (HP 4194a). The resonance frequency is as expected higher at about 250 Hz. What is somewhat surprising is that magnitude of the impedance is between 4 and 7 ohm above 350 Hz, since the nominal impedance of each of the two serially connected speakers is 4 ohms. |

Frequency response of the speaker. The frequency response was acquired in a non scientific manner by listening to the speaker and manually plotting dips and peaks of the response, calibrating the ear for amplitude variations by changing the driving voltage. Above 10 kHz the response was flat until 15 kHz, beyond which my ears could not follow.
![]() |
Speakers in their natural environment where they take up very little space. |