When indigenous people were first shown a photograph they could not at first recognise a picture. Over time the mind unconsciously learns to interpret information (no matter how abstract) and associate it with experience. This illusion is so great that only by being attentive we can hear a speaker is not and cannot be a violin or saxophone.
Speakers and musical instruments do share similar properties including the laws governing their behaviour. It is a miracle a speaker system can be made to sound real. But the fact is most don't. Loudspeakers have principally remained unchanged for approx 50 years. There have been few improvements in since their invention. One example is epoxy compounds for bonding voice coils to improve power handling.
The greatest advances have been in marketing where illusions of hearing differences are mostly generated by descriptions. As with musical instruments it is possible to make a speaker system come alive. All information on this site is directed to the understanding of bringing sound systems alive by application of four way active technology.
Speaker Specifications
Specifications are a guide similar to a road map. No matter how detailed it cannot show you what you will actually see until you get there. Another example is a list of ingredients for a recipe, again it cannot tell you how it will actually taste until you make it. Specifications that are similar for the same size musical instruments or speakers do not tell you which sounds best. You actually have to listen. Large speakers and musical instruments suit low frequencies and vice versa. A speaker and most musical instruments can only function efficiently within 3 octaves (octave is ratio 1:2) or 1 decade, ratio 1:10.
1. Frequency response measurement. All electronics today has frequency response capability that extends well beyond the audio range. The measured sound spectrum is 20Hz to 20KHz. A speaker and most musical instruments have a limited frequency response and only accurate within 3 octaves (octave is ratio 1:2) or 1 decade (decade is ratio 1:10). 3.2 octaves is 1 decade. The standard way of measuring a speaker or speaker system is with a microphone, at one point at 1 meter on axis and a tone sweep across the frequency spectrum and plotted. dB/mW This measurement provides vital information for detailed research but limited for how it sounds.
2. polar Response Dispersion. Measurement of sound dispersion around the sound system, plotted at different frequencies. Small speaker systems (passive) are the majority and have wide dispersion at low frequencies but narrow beaming dispersion at high frequencies. An ideal sound system will have even dispersion at all frequencies. To achieve this the diameters of speakers must change each 2-3 octaves, this requires 4 different size speakers to cover the sound spectrum. 90deg is ideal.
Theoretically a single speaker would have to change diameter from (1in - 24ft) or (20mm - 8m) to maintain similar level and polar dispersion over the frequency spectrum.
3. Spectral Energy Response. Includes Energy response, Frequency response, polar response and Efficiency for the whole acoustic energy delivered from a speaker system giving a closer understanding of how a system behaves.
(A) The on axis frequency response appears flat. The high frequency energy is small but highly directional. The low frequency energy is larger but spread around the box. Off axis, there is little to no high frequency and the sound will be muggy. (B) On axis the frequency response is trebly. Both high and low frequency energy is equal but the high frequency energy is highly directional and the low frequencies are spread around the box. (C) On axis frequency response is flat. Both high and low frequencies have a similar polar response and their energies are equal. Off axis the sound will remain balanced and the level will evenly decrease.
To achieve (C) an ideal system would have 4 speakers each covering 2-3 octaves.
(12-15/18in for Sub-bass) . (8-10in for Lower voice) . (4-5in for Upper voice) . (1in for the Harmonics)
4. Efficiency and power. Cone Speakers are approx 2% efficient. Power and efficiency is dependant on magnet and voice coil size. (BL) B is total magnetic flux in the gap, and L is length of voice coil wire. Also mass and area of cone, suspension compliance, damping and frequency.
The fine art of musical instrument and speaker making, is combining efficiency and power, responding with clarity and evenness, complying within the 3 octave rule, 1 decade. A musical instrument and speaker may be efficient but uncontrolled which means the sound is colored the notes uneven without clarity. Many cheap musical instruments and speakers behave this way. An instrument or speaker may be in-efficient having a flat response the notes even but lacking dynamic expression and responsiveness, requiring to be played harder to be heard.
A 3dB efficiency difference between speakers may be aurally small but 3dB is double or half the power. A 10dB efficiency difference is heard as double or half as loud. 10dB is ten times or one tenth the power. It is easier and less expensive for manufacturers to make low efficient speakers have a flat response. With small domestic sound systems this may not be a problem as modern high powered amps are inexpensive. With large professional sound systems speaker efficiency is very important.
5. Inter-modulation. Linearity and Intelligibility. Muddle-ness and inter-cluttering within the music which makes it difficult to discern detail. With continued listening this becomes fatiguing. This is caused by interference within and between speaker components, compounding as the power increases. Primary inter-modulation is where a single speaker is used beyond 3 octaves or full range.
The large cone movement (excursion) for low-frequencies, modulates the middle and higher frequencies, causing them to sound dirty. Lobe and node distortion is caused by high-frequencies creating secondary vibrations and chaotic resonances within the speaker cone, causing it to sound harsh and screechy.
Cinema sound that is harsh. Venues and Live concerts where not a single word can be understood. We have become so accustomed to this we have accepted it.
6. Dynamic power response Is the ability for the sound fidelity to remain intact between low and high power. It is not possible for one amplifier driving different speakers (passive) or one speaker used beyond 3 octaves to achieve this accurately. The majority of sound systems are passive due to cost and the fashion for systems to be small. For a system to sound consistent and accurate at all power levels it must be active. Each speaker driven by its own amplifier and matched in efficiency, power and dispersion. This also eliminates cross interference within and between the speakers (inter-modulation).
Professional sound systems use horns and compression drivers for extra power and efficiency.
More information in Horn Systems page.
Speaker Construction
The construction of speakers is approached in the same way as musical instrument making. Fine tolerances and attention to detail make large differences to performance. Large musical instruments and speakers suit low frequencies and vice versa. Each speaker and instrument can only function efficiently with linearity within 3 octaves (octave is ratio 1:2). Theoretically a single speaker would have to change diameter from (1in - 24ft) (20mm - 8m) to maintain similar level and dispersion over the frequency spectrum.
The majority consist of paper or plastic moulded into a cone shape, loosely suspended in a frame so as to easily move back and forth to vibrate the air. Glued to the back of the cone is a coil of wire (voice coil) within a strong magnet field. Passing electricity through wire causes a magnetic field around the wire, which attracts or repels, causing the cone to move back and forth. The larger the magnet and voice coil (BL) the greater the power and efficiency if well made. Externally vibrating the cone will cause the voice coil to generate electricity. A speaker can work well as a microphone especially for bass drums.
The energy of the magnet is conducted through the mild steel pole plates and pole piece and concentrated (north - south) across the gap. Hopefully the voice coil has been perfectly centred in the gap. The clearances are very very small, less than half a bees dick. The smaller the gap - the more intense the magnetic field - the greater the efficiency. The slightest variations in alignment during manufacture, cause large variations in performance. No two speakers or musical instruments can be identical.
Voice Coils. Passing electricity through wire causes a magnetic field around the wire. Changing polarity of the electric current through the wire also changes the polarity of the magnetic field created around the wire. The interaction of the two magnetic fields, causes the voice coil to be pushed out of the gap forward or backward, depending on the polarity of the electricity through the voice coil.
Speaker polarity. The standard test for a speaker, is to put a battery across the speaker terminals. When the + of a battery is put on to the + marked terminal of the speaker the cone should move out. This is also the correct test for checking the polarity of speakers in stereo systems. 1.5 Volt battery is safe to use. Warning. Never use this test on a compression driver.
Most voice coils are double layered, wound with enamel coated copper wire around a former then bonded with an epoxy compound and backed in an oven. Voice coils can also be wound with rectangular aluminium wire to achieve less weight and greater accuracy. A cone speaker is approx 2% efficient therefore approx 98% of the electrical energy is wasted as heat.
Voice coil Length. At bass frequencies the voice coil has to move back and forth a long distance, especially at high power, compared to the high frequencies. During movement, the % of voice coil in the gap must remain constant. The voice coil can be long and the pole plate thin or the voice coil short and the pole plate thick to achieve the same outcome. There are argued advantages and disadvantages both ways. Mid and high frequency speakers cones only move a small distance, compared to the large movement of bass speakers. The voice coil length and pole plate thickness are similar.
Mid and high frequency speakers are approx 6dB to 10dB more efficient than bass speakers. In small passive systems the mid and high frequency speakers are attenuated to match the less efficient woofer. In active systems the bass speaker is driven with higher power. An example is two 15in woofers driven with 400 Watts to match a 12in mid speaker driven with 200 Watts.
Voice Coil Diameter. On the same diameter speaker a small voice coil has less control over the cone compared to a large voice coil. With a small voice coil the cone is able to be more resonant compared to the same size cone with a large voice coil. Some small voice coil speakers may appear to be more efficient but this extra efficiency may be only at the one resonant bass note. At frequencies above this resonant bass note the speaker may be less efficient compared to the same cone with a larger voice coil. Cost and performance of speakers increase with voice coil size.
The larger the voice coil the better the control over the cone, and therefore improved damping and linearity. Large voice coils are expensive to make (limit approx 4in 100mm). Assembling the speaker is also more difficult, tolerances taken to greater accuracy. The larger the voice coil diameter - the larger the area of the magnetic gap. To keep the flux density of the magnetic energy (per unit area) in the gap the same (for the speaker to retain the same efficiency) magnet size and therefore mass of the speaker must be increased. The major advantage of larger voice coils is greater power-handling.
Square Wave. It is not possible for a speaker or musical instrument to produce an acoustic square wave. A square wave can only exist in electronics where 'physical' mass and force do not exist. A speaker will average the electrical energy of a square wave as the fundamental period of a sine wave, with discordant third harmonic resonances.
A speaker responds only to the resultant energy of an electrical signal, but not to the separate characteristics of phase between voltage and current within that signal. Also to understand this demonstrates how much of the audiophile and professional audio myths about 'electronic time alignment' are simply marketing fabrications. However acoustical wave-length time alignment between the physical distances of speaker components (woofer and tweeter) is important and should be taken into consideration.
What governs the 3 octave rule?
Sound is a wave motion of air having length and co-responding height, described as frequency. At frequencies where wave-lengths are equal to and become longer than the speaker diameter, decreasing in frequency, the frequency response is smooth, and dispersion widens (directivity).
Theoretically to maintain the same power, the cone must move 4 times the distance, for each octave decrease. Cone mass, and reducing air resistance as the frequency decreases, cause this. The same as a musical stringed instrument (constant velocity).
The upper limit for high frequencies is where wavelengths become smaller than the speaker diameter. This causes node and lobe distortion within the cone (chaotic resonances). Dispersion narrows to a beam and the frequency response becomes chaotic, sounding harsh and screechy.
The lower limit is where wavelengths become 10 times the speaker diameter. Cone movement is so great it becomes inefficient, non-linear and generates inter-modulation (affecting other notes). The fashion for bass speakers to be small compromises performance regardless of marketing claims.
An ideal system will have 4 speakers, each covering 2-3octaves. Eg.
(1in for the harmonics). (4-5in for upper voice). (8-10in for lower voice). (12-15in for bass).
professional systems use horns and compression drivers for extra power and efficiency.
Speaker power
Cone Speakers are approx 2% efficient. power and efficiency is dependant on BL, which is magnet and voice coil size. B is the total magnetic energy in the gap. L is length of voice coil wire. Also mass and area of cone, suspension compliance and frequency.
Bass power. The maximum distance the cone can move limits bass power. To maintain the same acoustic output the cone must move 4 times the distance for each octave decrease and vice versa. 1/4 distance, for each octave increase. Two times due to cone mass changing direction. Two times because air is slippery. As frequency decreases air slips sideways decreasing resistance. As frequency increases air doesn't have time to slip sideways, increasing resistance against the cone (radiation resistance).
At wavelengths greater than 10 times the speaker diameter, cone movement is so great it becomes non-linear and inefficient. Some manufacturers make the suspension tight to limit cone movement, compromising bass performance for a higher power rating.
High frequency power is limited by excessive heat that can destroy the voice coil.
Speaker manufacturers have difficulty specifying power handling. In marketing there has been a modern trend to sell speakers as being capable of 1,000 Watts plus, similar to selling a motor vehicle as being capable of going at the speed of light. At bass frequencies the cone exertion is limited by the tightness of the suspension and may be at maximum between 20-100 Watts. Above maximum exertion the bass notes become compressed.
Cone movement can be so great that at the point where the voice coil and cone are joined it can hit the pole plate, making a loud cracking sound and therefore warning. The voice coil may be capable of 50-200 Watts before it burns out (depending on time). At mid and high frequencies, cone excursion does not limit power handling. The voice coil will keep getting hotter until it burns up, and stop abruptly without warning. Sometimes excessive heat will cause the voice coil to become loose from its former. This will be heard as an annoying mechanical rattling sound.
Speaker Impedance (Z) ohms
A speaker voice coil is specified in Impedance, because its Resistance changes with frequency.
Impedance (Z) is Resistance that changes in reference to frequency (simple description).
Resistance (R) may also be symbolised as (ohm) or the Geek letter omega (Ω)
8 ohm is the standard Z for speakers, measured at 400Hz.
4 ohm is the standard Z for car speakers (because of the low supply 12V battery, explained in amplifiers).
Back E.M.F. At fundamental resonance (Fs) a reactive effect called Back EMF, causes the speaker (Z) Impedance to rise many times. E.M.F. Electro Motive Force is a term that applies to electric motors and generators. Externally moving the cone causes the voice coil to generate electricity. A speaker can work well as a microphone especially for bass drums.
At 'Fs' fundamental resonance, the cone vibrates readily, almost of its own accord, and the voice coil partially generates electricity. It appears from the outside, that the voice coil Impedance Z has risen. At 'Fs' the speaker is also at maximum efficiency, requiring only the smallest amount of input signal for the cone to vibrate. Another way to understand this is to hold the speaker in your hands (not connected to the amplifier) and tap the cone. It will sound like a drum skin.
If you put a screwdriver or a piece of wire across the speaker terminals and tap the speaker cone it will sound dead. Also while keeping the short circuit across the speaker terminals if you gently but firmly grip the cone between your fingers and try to move the cone back and forth it will noticeably resist moving. For large machines, when the electric motor is turned off, a short circuit may be placed across the terminals. With the power off a spinning electric motor becomes a generator and shorting the input terminals will cause the motor to brake (stop almost instantly).
Efficiency
We can now review speaker efficiency by including radiation resistance, frequency, wave-length and fundamental resonance. Efficiency of a speaker is frequency dependant as shown in the below pic. Efficiency specifications of speakers are very misleading but this is not meant to be intentional. This olso reflects the reason that musical instruments are not specified with efficiency. Simply too complicated.
The efficiency specification of a speaker is typically referenced at 1 Watt at 1K Hz measured at one point at a distance of 1 meter 3ft. (dB/mW) This measurement shows the efficiency of the on-axis directivity of the speaker only. This measurement does not give the actual efficiency of conversion of electrical energy to sound energy, across the frequency spectrum.
In reality speakers are not measured at 1 Watt, but with a fixed voltage of (2.6 V) that would give 1 Watt if the speaker was exactly 8R ohm. At Fs fundamental resonance the speaker Z impedance could be 32R ohm, and with 2.6 Volt this would be 1/4 of a Watt. At the higher frequencies the speaker impedance also rises, and gives a misleading efficiency reading. At higher frequencies depending on the cone diameter the directivity of the speaker increases to a beam, making it appear that the speaker is increasing in efficiency.
The efficiency specification of a speaker is relative to the application of listening at one point on-axis only and not to the overall sound energy it produces at all frequencies. However if a speaker is fed with a constant power of 1 Watt (allowing the voltage to vary as the impedance changes) the efficiency measurement would look similar to the impedance graph.
Tonal Quality
The tonal quality a speaker has, when it is physically tapped is also the tonal quality it reflects in the music. There is no such thing as a speaker without its own tonal colour. Musical instruments and speakers share similar characteristics of the physics that governs tonal quality and sound colour. This physics has not yet been specified. Research was conducted in the 1980s by Dr Simon Marty in the department of electrical engineering Sydney University using holographic technology studying classical guitars. Recently, he received a grant, to continue this research into loudspeaker tonal colour. Hopefully this will enable faithful reproduction, of electrified classical instruments and we eagerly await the outcome.
Resonance Q
The springiness of the suspension and mass of the cone, cause it to have a natural fundamental resonance like a drum skin. Q is a number specifying Quality of resonance (duration of sustain time). This resonance is often tuned to the lowest bass notes, to help improve their efficiency.
Some manufacturers deliberately make the cone very heavy, shifting the fundamental resonance well below the musical range eg. 20Hz. Doping compound (tacky tar like substance) can be painted around the surround, damping the cone movement, reducing its Q at resonance. If this is overdone it can reduce speaker efficiency so it becomes virtually useless. There is a compromise between over damping the cone to flatten its response causing it to be inefficient and sound dead, or under-damping, allowing the cone to resonate (sounding alive) but will exaggerate the resonant note. This choice is personal. A Q of 1.4 tends to be the optimum for musical enjoyment.
Valve Amplifiers and Speaker Q
Complications regarding back EMF and fundamental cone resonance are often not fully understood by amplifier and speaker manufacturers. When technology changed from valve to solid-state it was noticed that solid-state amplifiers lacked warmth and bass performance and had to be twice as powerful as valve amplifiers to sound as loud.
Solid-state amplifiers function in voltage drive (zero output impedance) (100% damping factor) and act as a short circuit across speakers' effectively damping their performance. The larger the magnet and voice coil (BL) the greater the effect of damping. This becomes self-defeating for reproducing the lowest bass notes, and the primary reason for boxes to be ported to reproduce bass.
Valve amplifiers naturally function in current drive (high output impedance) (less than 50% damping factor). This allowed speakers to function at maximum efficiency. Valve amps are sensitive to speaker impedance variations. Some speakers had copper caped pole pieces, which helped damp impedance variations. Had this been fully understood at that time, solid-state amplifiers would have had a control to vary between voltage and current drive as a standard function.
Load Impedance
Review. R or Ω or Ohm, are the 3 symbols for Resistance.
This text uses R for Resistance. Other text may use Ω or Ohm for Resistance.
Z Impedance is, Resistance R that varies with frequency.
Zero 0R is a short circut. Infinite ∞ R is an open circut.
Load Impedance is the load speakers represent to the Amplifier. The maximum power rating of an amplifier is in reference to the load impedance. (Eg: a 100 Watt amplifier rated into 4R.) Only with a load of 4R is this amplifier capable of 100 Watts.
With a speaker load of 8R this amplifier will only be capable of 50 Watt. With a speaker load of 16R this amplifier will only be capable of 25 Watt and so on. The higher the load impedance, the cooler and more reliable the amplifier will be, and the lower the internal distortion but less power.
This 100 Watt amplifier must not have a speaker load that goes below 4R impedance. With a load of 2R this amplifier will attempt 200 Watt. The amplifier output transistors will get excessively hot and may/will fail. Heat is the enemy of amplifiers. The more speakers that are connected to the amplifier the lower the load impedance, and hotter the amplifier will get. Accidentally shorting the amplifier speaker terminals, represents a load of 0R and may instantly destroy the amplifier.
One 8R speaker the amplifier will only give 50 Watt run cool and reliable.
Two 8R speakers in parallel is 4R 50 Watt into each 8R speaker (100 Watt).
Four 8R speakers in parallel is 2R 50 Watt into each 8R speaker (200 Watt).
The speakers must not represent a load lower than amp Z rating.
The speakers can represent a load higher than amp Z rating.
Repeat-
An amp must not have a load, lower than the impedance the amp is rated for.
An amp can have a load higher than the impedance the amp is rated for.
Calculations
Speakers in parallel, divide speaker R by the number of speakers.
2 8R speakers in parallel is 8R/2 = 4R.
4 8R speakers in parallel is 8R/4 = 2R.
Speakers in series, x speaker R by the number of speakers.
2 8R speakers in series is 8R x 2 = 16R.
Speakers in series-parallel with even numbers of speakers only.
put each pair of speakers in series put each series pair in parallel with other series pair.
2 8R speakers in series is 8R x 2 = 16R.
2 8R speakers in series is 8R x 2 = 16R.
2 16R pairs in parallel is 16R/2 = 8R.
4 8R speakers in series-parallel = 8R.
The points
- Speaker Z load must not go below amp Z rating.
- Speakers in parallel or series must be identical.
- Speakers in series multiply speaker distortion but do not effect reliability.
- Speakers in parallel do not multiply speaker distortion.
Connecting speakers in series and series parallel may be essential in certain applications where many speakers are required to be connected to one amplifier. However connecting speakers in series causes the distortion of each speaker to be reflected into the others. But connecting speakers in series does not effect the reliability of the speakers or amplifier.
Speaker polarity Test The standard test for a speaker is to put a battery across the speaker terminals. When the + of a battery is put on to the + marked terminal of the speaker the cone should move out. This is also the correct test for checking the polarity of speakers in stereo systems. 1.5 Volt battery is safe to use. Warning never use this test on a compression driver.
Speaker Box
Sealed box. Bass speakers softly suspended often have (outside of box) a free air fundamental cone resonance below the lowest bass notes (eg 25Hz.). In design, the speaker is put in an imaginary infinitely large box. The box is then reduced in size to bring the combined resonance up to the lowest bass notes (eg 40-50Hz).
Low weight cone speakers need a large box whereas heavy cone speakers need a smaller box to bring them up to the same resonant frequency. Heavy cone speakers are less efficient with lower Q.
Above the square root of 2 (1.414) and the inverse of root 2 (0.707) are the most used numbers for academic calculations of speaker systems. Above is an example of changing box volume ratio 1:2 for the same speaker. The box-speaker system resonance changes by the square root of 2 (1.414)
Above are the approx box volumes and crossover points for a 4 way system. These points can be moved to suit the individual speakers.
Sub-woofer EQ There is a large market for active sub-woofers especially for home cinema and vehicles. Many use 10-12in speakers in very small boxes, approx 1-2 cubic ft or 30-60 litres. The fundamental resonance of these speaker-box systems can be as high as 50-100Hz. Below system resonance the cone cannot increase excursion at 4 times the distance for each octave decrease as previously explained. Below system resonance the cone excursion is kept constant by the box. Efficiency and therefore frequency response rolls off at -12dB/octave (1/16th the power for each octave decrease below system resonance).
Many sub-bass amplifiers have special equalization (EQ) to boost amplifier power, compensating for the decrease in efficiency below resonance. Eg. If the speaker-box system resonance is 60Hz and the system is required to be flat down to 30Hz, then the amplifier will have to put out 16 times more power at 30Hz to compensate. These speakers already have heavy cones and therefore inefficient. Amplifier power requirement can be approx 300 Watts for domestic application. Some systems designed this way work well but often sound un-musical.
Bass Reflex phase Inversion Enclosurers
Ported Box The trend for boxes to be small makes the port a forced compromise. The ported box becomes a Helmholtz resonator (enclosed volume of air with aperture) similar to wind instruments. The resonator generates an artificial bass to represent the lowest notes. These generated notes have a separate tonal quality to the notes above them and are in reverse phase.
Many professional bass speakers are designed for ported boxes. These speakers are tightly suspended. Their free air fundamental resonance can be above the lowest bass notes. This is done to give the speaker a higher power rating. For these speakers the port is the only means by which the lowest bass notes can be generated regardless of box size. The port extends the frequency response below what the box would normally do if it was sealed. But unlike the sealed box which rolls off at -12dB/octave below system resonance, the port rolls off at -18dB to -24dB/octave below port resonance.
Ports also make the speaker vulnerable. Frequencies below what the port is tuned for cause the speaker to behave as though it is outside the box. At high power, at frequencies below the port tuning the cone movement is extreme and can easily damage the speaker. To protect the speaker it is essential to put a filter circuit, in or before the amplifier, preventing frequencies below the port tuning being amplified to the speaker. More details in (Amplifiers Voltage Drive Current-Drive).
Isobaric Compound sub-bass
Isobaric Cavity Drive sub-bass is a special high-powered Helmholtz resonator. The speakers are inside the box and compounded to increase power. The 2 speakers are coupled together to act as a single speaker. Frequency response is limited to within the lowest 2 octaves. Cavity drive can be efficient and generate similar bass energy as a normal sealed speaker box 2-4 times in size.
There are many excellent books and software programs for calculating box sizes and port tunings.
Baffles
The infinite baffle is the simplest of all baffles. It is a theoretical concept referring to a speaker delivering a hemispherical polar response, free of diffraction created by box corners etc. It also refers to a sealed box where sound from the back of the speaker is 100% absorbed, and only the sound from the front of the speaker is heard.
A circular box is the best acoustical container for a speaker but is more difficult to construct. The rectangular box with parallel sides, is the worst container for a speaker to be placed in. The inside parallel walls of the box coincide with wavelengths of musical frequencies, creating standing waves and resonances. It is important to put acoustical absorbent material inside the box, to absorb sound from the rear of the speaker. Fibreglass is best if you are not allergic to it.
Box construction
A speaker box is 6 pieces of wood screwed and glued together. In a rectangular box it is important to place the speakers off centre to minimise external diffractions from the sides. Diffractions created by equal wave-lengths (path-lengths) to corners, sides, cause the most problem. Aspect ratio (height - width - depth) must be un-equal. path-lengths between the sides, top to bottom, front to back, must be un-equal.
Make sure all measurements from the centre of each speaker to any side or top are not equal. In theory the box should have no parallel sides and no right angles. The box should be absolutely air tight (port excepted). Every side should be cross-braced. Also put cross-braces across the baffle, between the speakers. The box should produce no mechanical resonances. A true fanatic would put lead-lining on the inside and a double sand filled outer shell. The speakers should be vertically in line, off center and close together as possible. Fidelity improves if sound appears to come from a point source.
Other types of speakers
Dome the majority of tweeters are dome, similar to the cone speaker.
Ribbon tweeters are expensive, less efficient, not common, have excellent performance.
Electro-static speakers are expensive very fragile and have a niche following, excellent transients.
-- To be continued ----