3.1 Pickup types
Any pickup is a local representation of what is going on in acoustic sound production. Compare it to viewing a movie through a hole, exposing a patch of the screen. Designers move the hole around to get the best impression of the movie by a single patch. Obviously standing back and seeing the whole screen is the best, but then you will also see the cinema and the stray light from the other viewers. This is the equivalent of the microphone. Picking up sound signals is all about getting the whole picture and nothing but the picture. This section is about the viewing holes.
Electromagnetic pickups consist of wound wire around a magnet. Each winding is producing an electrical voltage when the magnetic field strength is changed. These changes are made by movement of the metal in the string, which is magnetized by the field and attracts magnetic field lines. If the string moves in the magnetic field the field is changed and a current appears in the coil. To enhance the output, many windings are applied, requiring thin wire. The long and thin wire creates an electrical resistance which affects the high frequency response, since the resistance and the capacitance of the wire cast a low pass filter. The alternative is to use thicker wire and less windings. The output is brighter, but weak. Pickup designers have been pushed back and forth between these choices.
The pickup and the cable to the amp have capacitances which together with the induction of the coil create an electrical resonating system. The coil resistance may under-, critically- or over-damp the signal, affecting the sound. Fig 3.1.1 shows some characteristic response curves of pickups, demonstrating that slight under damping is often chosen, resulting in a 3-5 dB bump at the resonance frequency. This gives extra presence in the 2-5 kHz range. Guitar strings do not produce strong frequencies higher than about 4 kHz (1 kHz fundamental, 3rd harmonic). Some pickups are labeled humbucker. The humbucker has double coils which were introduced to suppress electronic noise, picked up from the surroundings, by means of subtraction. The signal from one coil is inverted and summated to that of the second coil. Simple summation results in the ‘hot humbucker’. Summation with a phase shift results in a differential signal with a different transfer function, labelled ‘active humbucker’. This curve is tilted 6 dB/octave relative to the hot humbucker, pointing at differential processing. The high frequencies are amplified at the expense of the low frequencies.
Fig 3.1.1. Pickup response curves used as a template by Fishman for design purposes (from http://www.premierguitar.com/articles/20082-unwound-fishman-rethinks-the-electric-guitar-pickup?page=2).
Electromagnetic pickups register the string motion and produce coloured sounds, which are part of the sound design of fully electric guitars, but less suitable to amplify acoustic instruments, since the acoustic sound shaping is missed. With a little trick, some of the body vibrations can be picked up as well. The pickup is mounted on the soundboard and thus vibrates relative to the strings. The motions of the plate and of the string thus add up, combining string vibration with body vibration, or more precise: local soundboard vibration. If the pickup has a significant mass, it may change the tuning of the soundboard.
Contact pickups, with the piezo pickup as important type, convert deformations of a crystal, like compression, shear or bending, to a voltage. The sensitivity for deformation is exceptional and provides nanometer resolution. That makes them ideal to measure deflection of a plate or pressure on a bridge. The frequency response is excellent and the output signal is strong, be it with a high output resistance and low capacitance. Both properties make a dedicated preamp necessary. Piezo pickups register local vibrations, not the guitar sound. The small dimensions of the pickup make them easy to apply and record specific vibrations, such as the stimulus from a single string, the resonance at a particular spot on the soundboard, the stress on the neck and so on. The most popular use is as an under saddle pickup, but this is an underestimation of its potential. In particular plate bending is of interest, which involves stretching at one side and compression at the opposite side. Gluing the piezo cable over an area monitors bending in that area. Another useful property is that piezo sensors can measure the relative distance between two plates, at both sides of the crystal or polymer. Any motion which these points share is not recorded and the sensor is thus less sensitive to acoustic feedback, etc. If the size of the monitored area is increased, more vibrations are included, but also the vibrations from outside, causing the plate to bend: feedback. The monitored area can be expanded using piezo coax cable (http://www.meas-spec.com/piezo-film-sensors/piezo-cable.aspx), which is relatively cheap and can be arranged in a great variety of shapes.
A special case of the contact pickup is when it is used as an accelerometer. One side is glued to the surface of the guitar and the other side to a small mass. If the surface vibrates, the inertia of the mass induces a force on the crystal which is measured as an electric signal. Accelerometers register local vibrations and contact noises alike, but are not sensitive for low frequencies.
Another type of pickup is the capacitance pickup. Like tuning channels in vintage radios, displacement of an electrically charged platelet may cause a change in capacitance. The capacitance is set by the physical distance between two conducting parts and if they move relative to each other, the motion can be converted to an electrical signal. This type of sensor picks up the motion of a key on a keyboard better than the tiny vibrations in the guitar body. The largest motion is found on the string, relative to the guitar body. Capacitance pickups could serve as an alternative for electromagnetic pickups, without the notorious electromagnetic interference (hum).
Small microphones are used frequently. The microphone technology is so advanced that even small and cheap mikes may record the full spectrum with great accuracy, such as electret microphones. These hold a dielectric, containing electrically polarized particles, of which the common electric field is shaped by the force due to sound pressure. The difficulty with internal microphones is in the exact placement. The sound field is strongly location dependent. Also, mikes pickup feedback and must be mounted in a flexible way to avoid contact noise. Both these properties make the use on stage in presence of PA systems troublesome. Some pre-amps are fitted with an adjustable notch filter. The notch is very selectively damping a particular frequency and adjusting the notch to a feedback frequency can effectively suppress feedback.
Because none of these sensors are capable of recording the full acoustic guitar sound, a combination may give the best results, as is done in some high end products. Examples of combined sensors may be found at LR Baggs’ TRU•MIC technology (http://www.lrbaggs.com/pickups) or at Fishman’s build in preamps (https://www.fishman.com/products/series/aura/aura-pro-onboard-preamp/).
3.2 Pickup position
In principle we have four types of sound information. These can best be recorded with specific pickups:
– String vibration (magnetic and under bridge pickups)
– Local wall vibrations (piezo acceleration or bending pickups, strain gauges)
– Wall area vibrations (Piezo cable)
– Sound wave information (microphones in or outside the body)
String vibration is affected by the connection with the soundbox. In fact, the soundbox feeds back on the string, which is easily verified by plucking one string. The resonances of other strings merge into the tone because the bridge acts as a transducer. The string vibration is the stimulation for the soundbox, but it is a poor representation of the acoustic sound and thus only interesting if something else is done with it, like sound processing or adding the contact noises of the playing style.
Local wall vibrations are associated with one or a selected group of resonances. Which resonance is picked up where can be read from the Chladni patterns. A location mid under the bridge is on the nodal lines of many vibration modes, certainly for classical guitars. The bridge is stiff, relative to the plate and little can be recorded there in terms of variation in plate vibrations. However, bending of the bridge is still relevant as Fig 2.4.1 demonstrates. In general, bending sensors provide more information than acceleration sensors. However, with acceleration sensors a specific resonance can be selected. In principle the acoustic spectrum can be reconstructed from a considerable number of local vibrations. The challenge is to find a limited number of locations that represent the sound as desired. As far as published, nobody tried to cover the whole plate with a sheet of piezo (or a mesh for that matter). That would be interesting! Al those higher vibration modes would not suffer from projection problems and could be mixed and moderated to obtain the signal one likes.
Area vibrations may be recorded by routing piezo cable in a dedicated way. The cable consists of a core, covered by a piezo spiral, a flexible metal mesh and insulation. When the cable is bent, the piezo crystals are compressed and produce a voltage. A common application is to measure in line with the bridge and record the (0,0), (2,0), (2, 1) and (3, 2) modes, but hardly the important (0,1) or (1,0) modes, neither any of the higher modes (check Fig 2.4.1). Measuring perpendicular to the bridge is even more selective, recording (0,0), but hardly any other since the axis is very often a node and in other modes a flank, thus with zero bending. However, a loop around the bridge is very effective in recording many nodes because there are always several nodes and antinodes in the loop. Fig 3.2.1 demonstrates this for the saddle between four adjacent antinodes. All four antinodes contribute to the signal. Following the same logic, a ring around one antinode is useless.
Fig 3.2.1. In a ring around the saddle between four antinodes, stretching is optimally used to record various vibration modes. Black dashed lines run from antinode to antinode, blue interrupted lines are nodal lines. The ring is not fully closed: one end is open and the other end is connected to the leads.
Sound wave information is the closest thing to the final output as we perceive it. In the section on projection it was explained that the low frequencies are like a pressure in the guitar box, which is uniform inside and which is omnidirectional outside the guitar. The higher frequencies, with shorter wavelength than the size of the guitar, bounce around in the soundbox and cast a more or less diffuse sound field, which can be recorded in the inside, but which becomes directional outside. A microphone inside may thus pickup all vibrations, whereas an outside microphone (on a stand) is more selective and also sensitive to movement of the player. Practical limitations on the frequency range and dynamics of microphones hardly exist anymore. For the location often a place near the soundhole or under the bridge is chosen. Microphones cannot be hard mounted to the top, since contact noises are recorded too strongly.
Mounting a certain distance away from a wall makes the microphone record both the incoming and reflected sound wave, producing double output. However, these interfere. The recorded sound shows a dip in the spectrum which corresponds to the distance to the wall. If the distance to the wall is a couple of mm, the dip falls above the relevant range of the spectrum. This setup is known as a Pressure Zone Microphone. However, in a guitar body the opposite plate is only centimeters away, causing a dip somewhere around 2-3 kHz. Every location may have pros and cons, but side mounting might be an option.
3.3 Acoustic independence
The sound system of concert halls is usually not under the control of the player. Microphones are installed on a stand and the sound is mixed and filtered to avoid feedback in the microphone. Feedback through the guitar body is not controlled by the technician and playing loud may become problematic. Feedback is referring here to positive feedback, which in a loop is causing uncontrolled loudness. There also exists negative feedback, which is indispensable in achieving stable, distortion free amplification. A more reliable setup is when the player can offer a signal for the sound system through a cable (or wireless for that matter), which is preset for optimal sound quality. This eliminates one route for feedback. It also hands the control over the sound to the player, who can hold some extra volume in reserve and adjust the tone to compensate for peculiarities of the hall. The internal pickups and further processing need to produce the acoustic sound which is targeted.
This does not automatically resolve the issue of feedback through the guitar body. The noise impact on the guitar can erroneously be picked up and amplified. If the gain in the loop output – sound system – guitar body – pickup – processor – output becomes larger than 1 the signal keeps growing until it hits a limit. The delay in the loop defines the pitch of the scream. The gain can be reduced by blocking the pathway somewhere between the incoming pressure waves and the pickup. To counteract acoustic feedback sometimes the soundhole is covered, reducing also the bass response of the soundbox (see Fig 7.5.1 in the section on modulation of low and mid resonances). The method is used in combination with a magnetic pickup to replace the acoustic sound with mostly string sound (contradictory to the aim to produce acoustic sound) or with an internal microphone. A more drastic way to suppress feedback is to use the under saddle pickup as a signal source and modulate it electronically to mimic the resonances in the body. This so called sound image can be taken from the same guitar or any other and also may include the virtual microphone characteristics.
3.4 Tuning the perceived sound quality
Despite careful design, the sound of the guitar may be less balanced than desired. A reason could be that other requirements are in the way of the desired sound, for instance because the size of the instrument is too small for deep bass, materials need to be heavier for strength reasons than acoustically desirable, etc. The acoustical balance may be restored by synthesizing the desired frequency spectrum. Recalling section 2.2 on sound quality, emphasis on certain resonances may be enhanced by selecting the pickup that measures that resonance best and process the signal appropriately. As an example, the most important quality aspect as analyzed by Meyer (1983) was the level and Q-factor of the resonance peak around 400 Hz. This peak is associated with different vibration modes in different guitars. In the Gibson Hummingbird it is a complex (0,1) resonance of the top, in a Conrad the (1,1) and in Martin’s D28, D35 and the classical Kohno the (0,2) of the top or a complex resonance called Long Flush, which are producing this peak. With local pickups these resonances may be recorded and pairs of local pickups may enable down selection to precisely the right resonance. The signal can be enhanced by passing a resonance enhancement filter to increase the Q-factor and subsequently mix it with other signals. This is less complicated than Griffin et al’s actuator driven processing of the resonance characteristics described in their US patent US 6320113 B1. What they do is mechanically enhance the resonances, foremost the 400 Hz peak. Yamaha (patent US 20110226118 A1) uses a computation intensive method (convolution from the time to the frequency domain, filtering and deconvolution) to enhance the low end of the resonance frequency spectrum. This is signal post-processing. Some less sophisticated proposed systems simply add the sound of a good instrument to that of the actual instrument to mimic improvement.
Treating the 400 Hz feature, the breathing mode and anti-breathing mode and filtering the overall frequency response curve may optimize for perceived sound quality according to Meyer’s criteria. Other quality criteria may be handled, depending on the style and interpretation of the music. Switching between styles is an option.