Studio Monitor Glossary

The acoustical controls are designed to compensate for commonly encountered issues seen in listening rooms:

Low Cut: Compensation for a lack of low-frequency damping

Bass: Acoustical loading from a wall

Low-mid: Acoustical loading from a desktop

LF Para EQ: Compensation for other deviations in the loudspeaker/subwoofer response

Mid: Midrange harshness caused by room acoustics

Treble: High-frequency harshness or over damping caused by room acoustic

As the amount of the effect can vary, each control has four settings. Adjusting these controls can make a huge difference to the performance of the system, so read the operating manual for suggested settings depending on the loudspeaker’s positioning in the room.

Sound travels out from the loudspeaker in more than one direction. As the listener moves off axis, the level of higher frequencies reduces more than the level of lower frequencies. The reduction in level off-axis compared to on-axis depends on the size of the driver and the surface into which it is mounted (a wall, panel, or loudspeaker cabinet). The larger the combined diameter and surface, the smaller the radiation angle: lower frequency, longer wavelength, more omni-directional behavior; higher frequency, shorter wavelength, more directional behavior; midrange frequencies sit somewhere between these two extremes. It is not practical to control the dispersion of a bass driver using an acoustical horn as it would have to be very large to control the long wavelengths involved. This frequency-dependent dispersion leads to problems around the crossover frequencies. In a two- way loudspeaker system the woofer plays up to the crossover point, typically 2–3 kHz. The wavelength of this frequency is quite small in relation to the diameter of the bass driver, which is why the directivity of the bass driver is narrow in the kHz frequencies. Just above the crossover frequency the tweeter starts to radiate sound.

The wavelengths of these frequencies, while short (10–20 cm), are quite large in relation to the diameter of the tweeter, and therefore the dispersion is wide. So the dispersion has changed from gradually narrowing with increased frequency in the bass driver region, to becoming wide again before narrowing with increased frequency in the treble driver region. This creates a non-smooth power response (the total sound energy radiated from the loudspeaker) which has consequences when the loudspeaker is placed into the listening room – see next paragraph. Even if the on-axis response is flat, the loudspeaker can have a tonality which is more apparent with decreased quality of acoustical treatment in the room. Neumann avoids this off-axis problem by using waveguides (acoustical horns) in front of the midrange driver and tweeters. Moving onto the sound in the room, even if the listener is positioned on-axis, the total sound at the listening position is the sum of the direct on-axis sound plus the reflected off-axis sound. Firstly, the direct sound should be a flat as possible. Secondly, the off-axis sound is reflected by the surfaces in the room (equipment, furniture, and walls).

This reflected sound will be colored in some way and attenuated in level due the reflecting surfaces’ acoustical properties, although it is possible with good acoustical design to minimize the adverse affects of this. The off-axis sound from the loudspeaker should not be colored otherwise the acoustics of the room will have to cope with this additional sound quality factor, which is not practical nor even possible in some cases. Finally, if there is a non- linear reverberation time in the room, the result at the listening position will also be colored. 

To measure the way in which sound radiates, many measurements are required around the loudspeaker. To avoid the need to measure every angle in all directions, typical just the horizontal and vertical planes are measured with a resolution of 5 deg. From this data, interpolation can be used to derive frequency responses of the angles that were not measured. Furthermore, symmetry can be used to further reduce the number of measurements required. 

When a Fourier transformation is taken of a time series the result is a complex frequency response. In the case of loudspeakers, the time series to be transformed is the impulse response. The impulse response is the response of the loudspeaker to an impulse input (a very short and loud sound – like a handgun or balloon burst). The impulse response gives a complete description of a linear time- invariant system (in this case, a loudspeaker playing at moderate levels). In acoustics, a complex frequency response is not very useful in its raw form, so the magnitude and phase responses are calculated to give the “magnitude of the frequency response” and “phase of the frequency response”. 

These are normally abbreviated simply to “frequency response” and “phase response”. Finally, the magnitude is normally converted to decibels as our perception of sound pressure is logarithmic and as a consequence easier to interpret on a log scale. 

Ideally, an accurate loudspeaker should have a flat magnitude of the frequency response in anechoic conditions across the entire audible frequency range (20 – 20k Hz). Technically this is impossible to achieve but we can get close. There will be a low frequency roll-off that limits bass extension. There will also be some deviation away from a perfectly flat response - this deviation should be minimized. 

Less capable loudspeakers would have less bass extension and a greater deviation of the frequency response away from flat. Examples of the causes of deviations are poor low-frequency alignment, cabinet and port resonances, crossover alignment, edge diffraction, and inappropriate driver selection. Sometimes the frequency response is made to be deliberately non-flat in anechoic conditions. This occurs when the typical loudspeaker positioning is known, and there is no possibility for the user to adjust the response on the loudspeaker. 

The low-frequency extension of a loudspeaker can be extended by using a subwoofer. This brings some additional advantages. Some of them are: increased SPL, reduced distortion from the main loudspeakers, and a decrease of the frequencies where group delay increases. There is one disadvantage of increased group delay around the crossover frequency. 

When a Fourier transformation is taken of a time series the result is a complex frequency response. In the case of loudspeakers, the time series to be transformed is the impulse response. The impulse response is the response of the loudspeaker to an impulse input (a very short and loud sound – like a handgun or balloon burst – also known as a “Dirac”). The impulse response gives a complete description of a linear time-invariant system (in this case, a loudspeaker playing at moderate levels). 

In acoustics, a complex frequency response is not very useful in its raw form, so the magnitude and phase responses are calculated to give the “magnitude of the frequency response” and “phase of thenfrequency response”. These are normally abbreviated simply to “frequency response” and “phase response”. 

The phase response can be further processed to give the group delay (it is the negative slope of the phase response, -dφ(ω)/dω). Group delay is the time it takes for the electrical input to pass through the loudspeaker system and become an acoustical output. Ideally group delay should be zero at all frequencies, i.e. all of the signal takes the same amount of time to pass through the loudspeaker, and that this time is a minimum. In practice, group delay increases with reduced frequency due to electronic (infrasonic) protection filters and the natural roll-off seen in vented cabinets (equivalent to a 4th order filter) or sealed cabinets (equivalent to a 2nd order filter). The higher the order of the filter, the higher the group delay. The lower the corner frequency of the filter, the higher the group delay. In this example the group delay increases towards low frequencies but psychoacoustic tests show that the values indicated are on, or just below, the threshold of audibility. 

Loudspeakers suffer from linear and non-linear distortion. Linear distortion can be seen as a non-flat frequency response and/or a non-flat group delay. This is discussed in other sections. Non-linear distortion adds new frequencies to the acoustical output that were not present in the electrical input signal. To measure this, a frequency (the fundamental, say 100 Hz) is played into the loudspeaker at a particular sound pressure level and a frequency above that test frequency is measured. Double the frequency is the second harmonic (200 Hz), triple the frequency is the third harmonic (300 Hz), etc. Sweeping the fundamental frequency allows a graph of frequency-dependent harmonic distortion to be plotted. Total harmonic distortion (THD) is the relation between all sound coming out of the loudspeaker compared to all the additional sound in the output that was not present at the input: second + third + fourth + fifth + etc. 

Harmonic distortion can be expressed in decibels or as a percentage. Second-order harmonic distortion is generally caused by asymmetries in the system. Third-order harmonic distortion is generally caused by “clipping” in the system, and this can come from the electronics or the acoustics, for example short voice coils. Odd-order harmonics generally sound a lot worse than even-order harmonics. Higher order harmonics will be lower than the second- and third-order harmonics, and should be at reasonably low levels in a well-designed system. Ideally the lower the harmonic distortion, the cleaner, or more transparent, the loudspeaker will sound. Less than -30 dB (3%) at low frequency and less than -40 dB (1%) at mid-high frequencies is normally considered to be good, lower values than these is of course better. Harmonic distortion is non-linear with level, in that an increase of 10 dB in the test signal typically results in a far greater increase in the level of harmonic distortion. As a result, one should check the test conditions before comparing measurements of different loudspeakers. In general, larger loudspeakers suffer from less harmonic distortion than smaller loudspeakers when played at the same level. Additionally, three-way loudspeakers will suffer from less harmonic distortion than two- way loudspeakers as each driver has less work to do. Both these effects can be seen together in the two examples below which are tested at the same sound pressure level. 

Unfortunately, the level of harmonic distortion does not correlate well with subjective sound quality, for example, an audio system with high levels of second order harmonic distortion can sound quite pleasant, whilst the same system with same level of third order harmonic distortion would sound rather poor. However, harmonic distortion plots are very useful for design engineers to use as a tool to trace problems in their designs. Note also that even if second order harmonics created by a monitoring system might sound good (like a tube distortion, exciter, compression, etc.) the signal is changed by this distortion. The acoustic output therefore differs from the electrical input signal, which must be avoided as much as possible. 

Loudspeakers suffer from linear and non-linear distortion. Linear distortion can be seen as a non-flat frequency response and/or a non-flat group delay. This is discussed in other sections. Non-linear distortion in the loudspeaker system adds new frequencies to the acoustical output that were not present in the electrical input signal. One measure of this is harmonic distortion, where a single frequency is played into the loudspeaker and the level of each harmonic measured. Another measure of distortion is to play a more complex signal into the loudspeaker and measure the additional frequencies in the output. The problem with this is that there are no internationally agreed standards for the test signal so comparison of different measurements is impossible unless the test conditions are identical. Since an intermodulation distortion measurement test signal consists of many discrete tones played at the same time, this type of broadband measurement signal is representative of real world usage of the system unlike the pure sine tones used for THD measurements. The test signal is commonly known as a multi-tone because it consists of a collection of tones. It sounds like someone leaning on a church organ with both forearms. In the example below there are 43 tones logarithmically distributed between 40 and 20k Hz. All the detail on the graph between these tones is the intermodulation and harmonic distortion, and this is what should be minimized. 

Like harmonic distortion, a lower intermodulation distortion results in a more transparent, cleaner sound quality. Unlike harmonic distortion, intermodulation distortion levels do correlate well with perceived sound quality, the lower the better. Intermodulation distortion is non-linear with level, in that an increase of 10 dB in the test signal typically results in a far greater increase in level in the intermodulation distortion. As a result one should check the test conditions before comparing measurements of different loudspeakers. In general, larger loudspeakers suffer from less intermodulation distortion than smaller loudspeakers when played at the same level. Additionally, three-way loudspeakers will suffer from less intermodulation distortion than two-way loudspeakers as each driver has less work to do. In the graphs below, one can see the advantage of introducing a midrange driver. The midrange intermodulation distortion is reduced by 10–15 dB. 

Typically maximum sound pressure level (SPL) figures are quite simply specified as a number. This means nothing unless the measurement conditions are known: frequency range, test signal, measurement type, distance of the microphone to the loudspeaker, duration of the signal, location of the loudspeaker, etc. As an example, here is the specification for the O 410: Maximum sound pressure level in half space at 3% THD at 1m averaged between 100 Hz and 6 kHz = 120.0 dB. Changing any of the measurement conditions will give a different value. To measure the maximum SPL, a sine wave is played into the loudspeaker and it is increased until the distortion reaches some value, say 1% or 3%, and then the SPL is noted. Sweeping the sine wave’s frequency allows a graph to be plotted or an average to be calculated. Alternatively a broadband signal can be used, for example, pink noise, IEC-weighted noise, or “program material” (whatever that means!). As with most aspects of acoustics, maximum SPL is frequency-dependent. Lower frequencies are harder to reproduce at high levels as the displacement of the drivers must be high (four times the displacement required with each halving of frequency). This is seen as a drop in maximum SPL with reduced frequency. 

The loss of headroom at low frequencies can be made up by adding a subwoofer to the system: 

Note these measurements are for a subwoofer located in half-space. Subwoofers are typically located in quarter-space and so the subwoofer part of the curves (below about 100 Hz) would be 6dB higher. Adding a subwoofer also brings a decrease in the low-frequency cut-off of the system, reduces harmonic and intermodulation distortions (even in the midrange), and changes to the group delay. Adding a subwoofer to a smaller loudspeaker gives a more dramatic improvement to the maximum SPL. 

A waterfall plot is a set of frequency responses that are plotted onto a 3D axis. Each frequency response occurs a little bit later in time so that a picture is built up of how the loudspeaker behaves once the sound is turned off. Resonances (left plot) are easy to see in the form of decaying ridges extending towards the front of the plot. Resonances can occur in loudspeakers and in rooms. In both cases these resonances should be minimized. Loudspeaker resonances are reduced by careful design. Room resonances are reduced by adding acoustical treatments to the room. 

The acoustical axis is a line normal to the loudspeaker’s front panel along which the microphone was placed when tuning the loudspeaker’s crossover during design. Pointing the acoustical axis, in the horizontal and vertical planes, towards the listening position or center of the listening area will give the best measured and perceived sound quality. 

A loudspeaker is said to be in “free-field” when it is placed in a space where there are no reflections or boundaries to restrict the sound radiation. These conditions can be found high in the sky (not so practical for listening or measurements) and in anechoic chambers. In practice, a loudspeaker placed on a stand far way from a room’s walls can be considered to be in free space. 

A loudspeaker is said to be in “half-space” when it is placed next to a large solid surface such as a wall or on the floor. The surface (acoustical boundary) limits the sound radiation to half of what it is in the free field. This has the effect of boosting the frequencies that are radiated omni-directionally, i.e. the low frequencies. The amount of boost depends on the solidity of the surface (theoretically 6 dB, in practice about 4 dB), and the frequency range depends on the size of the loudspeaker. This can be corrected with a suitably shaped filter, called “Bass” on Neumann loudspeakers. In practice, half- space is experienced when a loudspeaker is placed next to wall or flush mounted into a wall. 

Similarly, a loudspeaker is said to be in “quarter-space” when it is placed next to two large solid surfaces such as a wall and a floor, or a front wall and a side wall. The surfaces limit the sound radiation to quarter of what it was in free-space. This has the effect of boosting the low frequencies twice as much as seen in half-space. This can be corrected with a suitably shaped filter (“Bass” and possible some “Mid” too), or in the case of subwoofers, attenuating the output. 

“One-eighth space” is seen in the corner of rooms, and a very large boost is seen at low frequencies. This is a good location for subwoofers as the entire passband is boosted. This can be corrected by simply attenuating the output. For the same reason, loudspeakers generally sound very bassy in corners and so it is positioning that is not recommended for critical listening.

Active crossovers can be analog or digital. Analog crossovers are low-level electronic circuits positioned before the amplifiers in a loudspeaker system. As it is a low-level circuit, the filters can be precisely defined in terms of their frequency and phase responses. Digital crossovers are even more tightly defined and can offer additional signal processing that is either not practical or possible in analog electronics. Additionally, active crossovers do not heat up and change their properties as the loudspeaker is being used – this is a large source of distortion in passive systems. Klein + Hummel launched the world’s first commercially available three-way loudspeaker with built-in active crossovers in 1967, the OY. These days the professional broadcast and recording industry has almost entirely converted to active technology. 

Analog is a continuous-time representation of a signal that is represented using an electrical voltage or an acoustical pressure. An analog input is required for analog signals. The electrical interface can be balanced or unbalanced. Balanced interfaces can be electronic or transformer based. There is a maximum input level beyond which clipping will occur. Various connectors are in use, for example RCA, Jack, and XLR. 

Digital is a discrete-time representation of a signal that is expressed using numerical values. A digital input is required for digital signals. The input interface has parameters that must be appropriate for the signal: format, bit-depth, and sample rate. For example: AES3, 24-bit, 192 kHz. A practical example is, if the digital signal is always 48 kHz, as broadcasters always use, then a 48 kHz interface is sufficient. Various connectors are in use, for example RCA, BNC, and XLR.

Analog loudspeakers have an analog crossover made using discrete components such as op-amps, resistors, and capacitors. The input to an analog loudspeaker can be analog or digital. Digital signals are immediately converted to analog using a digital-to-analog converter (DAC). 

DSP loudspeakers have a crossover made using a digital signal processor and firmware. The input to a digital loudspeaker can be analog or digital. Analog signals are immediately converted to digital using an analog-to-digital converter (ADC). Digital signals can remain in their native sample rate and the internal filter coefficients changed to suit that sample rate. One coefficient set required for each sample rate so this is not very practical for all the sample rates now in use. These days a sample rate converter (SRC) is used to convert the digital input signal to a fixed internal sample rate. Additional hardware is required for this but the internal DSP implementation is much easier as only one set of coefficients is required. In the past, sample rate converters were not of sufficient quality to be used in high quality loudspeakers, but now they are. 

Bass management moves the low-frequency portion of a signal to a loudspeaker other than the one normally used to reproduce that channel. For example the bass of the center channel is reproduced by a subwoofer. This reduces the work that the center loudspeaker is required to do thereby allowing a smaller loudspeaker to be used or for the same sized loudspeaker to play with reduced distortion or at a higher level. Additionally, the LFE-channel is routed to subwoofers and/or loudspeakers that have the capacity to reproduce this high-level low-frequency signal. Finally, greater flexibility is available when positioning the source of the low-frequency energy in the room which can result in a better sound quality.

The human auditory system is a logarithmic device, so the numbers used to describe audio perception can easily become quite small or large and thus tricky to handle. Converting these logarithmic-spaced numbers to a linear scale is more convenient to work with and read. Decibels have no dimensions as it is the ratio of two values having the same dimensions. The classical Bel calculation gives values that are quite small for practical purposes, so a factor of 10 is used to give decibels. 

W = 10 log10 (P/Pref) dBW 

where Pref = 1 Watt 

Some simplified examples: An amplifier of 100 W can be expressed as 20 dBW. If a loudspeaker driver has a sensitivity of 90 dB/W/m, the maximum output with a 100 W amplifier would be 90 + 20 = 110 dB SPL. If the listening position is at a distance of 4 m, the maximum SPL would be 110 – 12 = 98 dB SPL as sound pressure halves with each doubling of distance. 

An additional factor of 2 is used when the ratio values are pressure or voltage, rather than energy or power. 

SPL= 20 log10 (p/pref) dB SPL where pref = 2x10-5 Pascals 

Some simplified examples: A sound pressure of 2x10-5 Pa, which is the threshold of hearing and used as a reference level, is easier to express as 0 dB SPL. Then 1 Pa is 94 dB SPL and 130 dB, which is the threshold of pain, is 64 Pa. If two loudspeakers are playing at 94 dB SPL, the total level would be 100 dB SPL.

There are a number of ways to describe the low and high cut-off frequencies of a loudspeaker: The −3 dB points are defined as the frequencies where the frequency response is 3 dB lower than the average level of the pass band of the loudspeaker. These are the most often quoted values and are easily comparable from one loudspeaker to another, even models from different manufacturers. An 

example is 30 – 24k Hz ±3 dB. 

The passband points are defined as the frequencies where the low frequency response passes through the lower pass band tolerance level. This value defines the lowest frequency where the loudspeaker reproduces audio with the same accuracy as the rest of the response. It is not useful for comparing different loudspeakers unless they have the same passband tolerance. If the passband tolerance is less than ±3 dB, the specified frequencies will be inside the ±3 dB point frequencies. An example for the same loudspeaker specified above is 32 – 20k Hz ±2 dB. 

The useable operating frequency range is defined by the two frequencies that are 10 dB lower than the average level of the pass band of the loudspeaker. Whilst this may seem to be a very wide tolerance, it is a frequency range of an unequalized loudspeaker that can be corrected using equalization. This specification is commonly seen in loudspeakers for use in installations and live sound. An example is 58 – 22k Hz for a loudspeaker that has a ±3 dB frequency response of 90 – 19k Hz. 

A measurement of sound pressure level can either be performed at particular frequencies or as a wide band summation over a defined frequency range, say 100 – 6k Hz. The ear is not equally sensitive at all frequencies and so an emulation of this is required to get an idea of the level of a sound. Additionally, the shape of the ear’s sensitivity changes depending on the level of the sound: it becomes flatter at higher levels. The frequency shapes (weightings) are defined in IEC 61672:2003. 

The different weightings were designed to be used for different sound levels but laziness has meant that only the A and C -weightings are typically used, thus creating a ubiquity. 

“Linear” or "Lin" is used for peak SPL measurements. Sometimes LP and/or HP filters are inserted. “A” is used for measuring sounds with a low SPL (40 phon). The low frequencies are heavily attenuated and the high frequencies are moderately attenuated. “B” is used for measuring sounds with a medium SPL (70 phon). It is not used very often. “C” is used for measuring sounds with a high SPL (100 phon). It has a gently rolling off low and high frequency response. “D” and “E” are used for measuring sounds with a very high SPL, like aircraft noise. “Z” or “Zero” is used for peak SPL measurements. LP and/or HP filters are not allowed. Other weightings exist for special purposes like measuring infrasound. 

Additionally, time-domain averaging is used: slow (1 s), fast (1/8 s), and integrate. “Slow” is typically used for tuning sound systems as the display changes relatively slowly thereby allowing time to take a reading of a rapidly time-varying signal such as music. “Fast” and “Integrate” are used by consultants for noise assessments. To measure a sound system, “C-weighting” and “Slow” is generally used. 

The entire front panel is constructed using Low Resonance Integral Molding™ materials (LRIM™). The midrange and treble drivers are mounted into a waveguide – see Mathematically Modeled Dispersion™. The waveguide is injection molded. The rest of the front panel is either a part of the waveguide molding or CNC machined. 

A loudspeaker driver’s motor consists of a coil moving inside a permanent magnet. The magnet has a usable magnetic field where it is required (i.e. where the voice is located), and a stray magnetic field where is not required (i.e. around the outside of the driver). Unless the loudspeaker cabinet made of steel, the stray magnetic field will be experienced outside the cabinet. This will distort the picture on 

CRT screens and potentially destroy data on magnetic storage media, such as hard drives and tapes. To cancel most of the stray magnetic field a bucking magnet (a smaller magnet with the opposite polarity) is glued onto the back of the main magnet. 

A vented cabinet has a tube somewhere in the cabinet. The acoustical system is then resonant at certain frequencies. This has the effect of producing a lot of sound for not much movement of the bass driver’s cone. The advantages are a reduced bass distortion, a higher maximum SPL, and a smaller required amplifier power. The disadvantages are a higher group delay at low frequencies, complex port design required, potential port turbulence and noise, and space is required for the port and its opening. 

A sealed cabinet design has no ports or vents, so there is a limit to the output level at low frequency. The disadvantages are an increased bass distortion, a lower maximum SPL, and a larger required amplifier power. The advantages are a lower group delay at low frequencies, no port design required, no port turbulence and noise, and no space required for the port and its opening. 

Except for the openings of vent in a vented design, both cabinet types must be completely air tight. Small openings can create whistling sounds and compromise the acoustical performance of the system.