Diffraction

Many performance advantages are the result of properly and efficiently coupling the high frequency transducer to the air load. Edge diffraction is virtually non existent as all frequencies that emanate "see" a more constant air load "impedance." This greatly improves the transient response of the speaker as there is no delayed energy reaching discontinuous enclosure surfaces and then "bouncing" back into the environment.

Other manufacturers' attempts at rounding the edges of--or making the driver mounting baffle narrower, making the baffle taper in a pyramid fashion or the use of absorbing foam near the tweeter only help to reduce this problem over a very narrow band of frequencies. These are modest attempts at best and do not address or correct the inherent cause of the problem. The reason these methods are inferior is relatively simple to explain. The simplified graphical representation below is provided to help you visualize this effect

.

The point source radiator above experiences no boundaries and propagates freely through the air in all directions. If an infinite plane bisects the point source as indicated by the red dotted line, then the source is said to be radiating into a hemispherical field on either side of the plane. Such a hemispherical field represents the air load presented to the average tweeter mounted on a traditional flat baffle. The significant exception to this fact is that the baffle plane is not truly infinite but is discontinuous and abruptly changes to a spherical field at the enclosure corners. This discontinuity is the source of diffraction.

As can be seen above, the expanding conic section of sound has traveled a considerable distance from the driver before reaching the enclosure edges. At this distance the wavefront has grown large and has had time to make the transition from a "high" to "low" acoustic impedance. This path is significantly further than the distance a wave travels before reaching the corners of a conventional flat baffle design. A wavefront having a lower acoustic impedance and larger total area is less sensitive to boundary transitions and the resultant diffraction effects.

The result of a larger total area is that a smaller total percentage of the radiating wave is encountering enclosure discontinuities. Any reflections that do occur contribute a lesser percentage of energy at any point further from the source. The resultant sum of all energy at any point in the far field is less affected and therefore less distorted from the original signal source.

Impedance matching is a common goal in many engineering disciplines. It is necessary in the fields of radio and television transmission to properly match antennas and other transmission lines to avoid reflected energy from wasting radiated power. Waveguides, in fact, are employed to transmit microwave signals and their theory of design and operation has been highly developed for many years now. Upon inspection, one would even find a waveguide in their microwave oven at home.

The effect of a lowered acoustic impedance is that when the expanding wave reaches the enclosure discontinuities there is less of a mismatch to that of the surrounding air. The abrupt impedance mismatch appears to the expanding wave as a sort of "invisible" barrier or wall, if you will, that causes part of the wave to reflect away from it. Any mismatch can result in a reflected wave emanating from the point of discontinuity and then recombining (albeit delayed in time due to the longer path it has traveled) with the un-reflected wave. The summed response of the two different waves can cause major fluctuations in the frequency response depending on the phase of one with respect to the other.

A properly designed waveguide will not suffer from this phenomenon. An extremely flat frequency response is a natural byproduct with frequency response variations less than +/- 1dB being quite common over most of the operating range. Complex DSP processing to correct for frequency response anomalies are not included in our designs for the simple fact that they are not required. Artificially forcing any system into a quasi-linear mode of operation by external means is inherently inferior to a system that is fundamentally linear to begin with. DSP processing has become relatively inexpensive and consequently very popular recently. It is an attractive method to a designer from a cost perspective in his/her attempt to linearize a fundamentally non-linear and inexpensive product.

As a side note, DSP based system designers will claim that diffraction effects are "corrected" as a side effect of frequency/phase correction. What they won't tell you is that the correction holds true for only one fixed point in space (usually on-axis) and falls apart at any other point (off-axis). This makes for a tiny little sweat spot - so be careful and try to hold your head real still.

Irrespective of these facts, a non-flat frequency response, by itself, is tolerated fairly well by the averaging effect of the human hearing mechanism. The greater problem arises from the fact that the delayed energy of diffraction tends to "smear" the impulse or transient response of the system, making fast rising signals such as snare drum strikes and plucked strings sound lifeless and artificial. Human hearing also localizes the point of origin (speaker location) more easily, thus destroying the sonic illusion of a three-dimensional space and a soundstage that extends beyond the boundary of the speakers. It is noteworthy that a poorly designed waveguide that has either a mouth area (the exiting point of the wave) that is too small or too abrupt will cause exactly the same phenomenon. The laws of acoustics simply must be accounted for in proper design.

Concerning impedance, a properly designed waveguide has the characteristic of raising the air load impedance to a given diaphragm in a way that can be considered a "conjugate match" with regard to what the diaphragm prefers to see when transferring its energy of motion into the air. It is often referred to in engineering terms as an "acoustic transformer," not unlike a transformer used in electrical circuits. If done correctly, the wave will transition to the air in a smooth and gentle progression without experiencing abrupt changes and the resulting reflections.

A flat mounting baffle, on the other hand, can be thought of as (and in fact IS) a waveguide having an effective dispersion pattern of 180° both vertically and horizontally. In other words, it produces a full hemispherical radiation pattern. Such a pattern is analogous to dividing in half the radiation field as represented by the Point Source graphic above.

It is the nature of virtually all piston type drivers (regardless of diaphragm geometry or size) to prefer an acoustic load or impedance that remains constant at all frequencies. As a point of interest, all power sources are the most efficient when their internal impedance matches exactly that of the load they are driving. A speaker driver is nothing more than an electric motor (power source) designed to move air.

A flat baffle does not have this constant impedance characteristic, i.e., the acoustic impedance decreases as the frequency decreases. In fact, this is the very nature of air itself in free space and is represented by the complete field of the Point Source. To be accurate, the impedance of a hemispherical radiation field is twice that of a spherical field. Unfortunately this increase is linear but the characteristic impedance decrease of air is not. In fact, to correct for this, a non-linear transfer function is required in the mathematical terms defining virtually all horns of various types and efficiencies.

Finally, a third rather complex source of diffraction requires our attention. The predominantly linear air load impedance represented by the hemispherical field of a flat baffle represents an impedance mismatch at distances very near the tweeter diaphragm. Enclosure edges are not at issue here. At certain wavelengths, usually at higher frequencies, there is an abrupt transition that takes place at the high impedance field near the diaphragm surface with respect to the lower impedance air load. "Pressure Zone" boundary conditions due even to the presence of the tweeter faceplate and the tweeter diaphragm diameter are sufficient to cause undesirable "virtual" secondary source radiation. Explanation of such boundary effects is beyond the scope of this text. The interested reader is advised to refer to engineering documents available through the Audio Engineering Society.

Such diffraction "type" events tend to occur at periodic intervals and are similar to harmonic orderings in music. These periodic transitions are dependent on the diaphragm size, mass and geometry (seeing that the air load cannot change). Periodic amplitude response variations at higher frequencies are a telltale sign this type of diffraction is taking place.

Making the diaphragm smaller has the potential of eliminating this problem except that the size reduction required does not provide enough surface area to produce reasonable sound pressure levels (volume). Heroic efforts have been made at altering geometries (parabolic and inverted domes, etc.) but their success is limited at best. Enclosure geometry has virtually no effect here save the recent introduction of those silly looking "eyeball" tweeters mounted on top of boxes. These types have the advantage of eliminating the boundary conditions that are the source of the problem but restrict operating bandwidth and force designers to use higher than optimal crossover points. The ideal solution is once again the waveguide. Impedance levels and boundary conditions are both optimized facilitating extended bandwidth and edge diffraction elimination.

To sum up then, a flat baffle can be considered as little more than a poorly designed waveguide with a purely linear transfer function. It does not correct for the impedance non-linearity of the air medium AND it introduces boundary condition errors. It seems therefore to be of relatively little value. It only finds a place in audio systems do to the necessary evil of having to have someplace to mount the drivers. Logic dictates that a truly progressive and efficient design would require the baffle to be a functional and synergistic part of the overall concept. If a component is necessary for one reason then it makes sense to extract as many functions from it as possible. As in most areas of life though, It is cheaper and easier to not be logical or creative -- it's just a question of who pays the price and when.

Thankfully nature has provided an elegant solution to these problems. Unfortunately, for the sake of manufacturing simplicity and reduced cost, it IS NOT the predominant "flat baffle" design and their variations being offered by others.

SUPERIOR PERFORMANCE MANDATES ABANDONING CONVENTIONAL SOLUTIONS
No other design, regardless of price, can compete with the advantages gained from waveguide technology as the laws of physics cannot be engineered out by the use of exotic and over priced materials being combined with fundamentally inferior methods.

Horn Theory | Waveguide Theory | Bass Driver Tech | Enclosure Tech | Crossover Tech