Music Engineering CD
Part 1 Physics of sound and psychoacoustics
Reference level sine wave tone*
1kHz at –18 dB below digital FS (–18 dBFS). The simplest musical sound, the sine-wave, is produced when an object (or in this case, the current in an electrical circuit) vibrates backwards and forwards, exhibiting what physicists call simple harmonic motion.
Track 2
Frequency sweep*
5Hz to 22 kHz. The pure tone, as illustrated in Track 1, actually sounds rather dull and characterless. But we can vary the number of cycles of oscillation which take place per second. Musicians refer to this variable as pitch – physicists call it frequency. This track also illustrates the possible range of human hearing, which in most adult individuals is limited to about 20 Hz to 15 kHz.
Track 3
Square waves at various frequencies*
A square wave may be synthesised by an infinite number of sine-waves but it’s much easier to arrange a circuit known as an astable multivibrator. Just such a circuit was used to generate the sounds on this track.
Track 4
Piano range; 27.5 Hz to 4.19 kHz
The relationship between written musical pitch (on a stave) and frequency (in Hz) is illustrated in Figure 2.5. Remember that the frequency components of the sound produced by each of these instruments extends very much higher than the fundamental tone. Take for instance the highest note on a grand piano. Its fundamental is about 4.19 kHz but the fourth harmonic of this note, which is certainly seen to be present if the sound of this tone is analysed on a spectrum analyser, is well above 16 kHz.
Track 5
The harmonic series, up to the twentieth harmonic The vibration of the open string produces notes which follow a distinct and repeatable pattern of musical intervals above the note of the open string. These are illustrated in this track. They are termed the harmonic series. A similar pattern is obtainable from exciting the air within a tube as is the case with a pipe organ, recorder, oboe or the clarinet.
Part 2 Microphones and acoustic recording
Piano recorded in nearfield. The idea behind near-field microphone technique is to suppress the effect of the acoustics of the room on the signal received by the microphone. This is accomplished by placing the microphone as close as possible to the original sound source.
Track 9
Classical microphone technique has ‘shied-away’ from nearfield method and it is for this reason that recording venues for classical music must be more carefully selected than those for rock and pop recording.
Track 10
Classical session, crossed-pair, stereo recording
Whilst it’s possible in principle to ‘mike-up’ every instrument within an orchestra and then – with a combination of multi-track and electronic panning – create a stereo picture of the orchestra, this is usually not done. Instead, most recordings of orchestras and choirs depend almost exclusively on the application of simple, or ‘purist’ microphone techniques where the majority of the signal that goes on to the master tape is derived from just two (or possibly three) microphones. In this track, the technique of a coincident crossed pair (sometimes referred to – somewhat incorrectly – as Blumlein Pair) was used.
Part 3 Electric instruments
The most famous electric piano is, without doubt, the Fender Rhodes. This, and its many imitators is actually more of an electronic Glockenspiel (or Vibraphone) than an electronic piano because the sound-producing mechanism is formed from struck metal bars, the hammers being actuated via a conventional keyboard mechanism. An adaptation of the electric guitar type pickup is utilised so that the piano can be amplified.
Track 13
There are two basic types of electronic organ. The divider type uses a digital top-octave generator (one oscillator for each semitone) and chains of divide-by-two bistables to provide the lower octaves. However, this approach tends to produce a ‘sterile’ tone disliked by musicians. The alternative is known as a free-phase electronic organ. Theoretically the free-phase organ has a different oscillator for each note of the keyboard. In the case of the Hammond each oscillator is mechanical.
The Clavinet was, commercially and artistically, the most successful keyboard produced by German Company Hohner, who designed it to replicate the sound of a Clavichord.
Part 4 Electronic effects
Unprocessed guitar. The earliest electric guitars were created by attaching a contact microphone to the top sound-board of a conventional acoustic guitar, the resulting signal being fed to an external amplifier. However, the modern electric guitar was born with the invention of the electromagnetic pick-up and a typical instrument (Fender Stratocaster) is used in this and the following examples.
Track 15
The Slap (or Slap-back) echo was first heard in Scotty Moore’s chiming lead guitar on the early Elvis records.
Track 16
Recordable tape loops originate with the work of Brian Eno and Robert Fripp in the early 1970s, where sounds are recorded over and over onto a loop of magnetic tape on a tape deck which incorporates the crucial modification that the erase head is disabled and an electrical path provided so that sounds may be re-circulated in the manner of a tape-echo device. The sounds are therefore recorded ‘on top of one another’ and one instrument may create vast, dense, musical structures. Importantly, subsequent signals do not simply add and from an artistic point of view this is extremely valuable because it means, without continually ‘fuelling’ the process, the ‘sound-scape’ gradually dies away. The artist may thereby control the dynamic and tonal ‘map’ of the piece. Nevertheless, the control of this process is not comprehensive and many of the results are partially random.
Track 17
In a ‘fuzz’ circuit, the guitar signal is applied to a circuit at a sufficient amplitude that it drives the circuit beyond its available voltage swing. The waveform is thus ‘clipped’. For guitarists this effect is amongst their stock-in-trade. It’s now understood, the manner in which the circuit overloads influences the sound timbre.
Track 18
Wah-wah is a dramatic effect derived from passing the signal from the electric guitar’s pickup through a high Q band-pass filter, the frequency of which is adjustable usually by means of the position of a foot-pedal. The player may then use a combination of standard guitar techniques together with associated pedal movements to produce a number of instrumental colours.
Track 20
Pitch shifting is used for a number of aesthetic reasons, the most common being the creation of ‘instant’ harmony. Simple pitch shifters create a constant interval above or below the input signal, like these ‘harmonies’ at the octave.
Track 21
Pitch-shift, Perfect 4th up; Perfect 5th up
Harmony at a perfect fourth produces only one note which is not present in the original (played) key. However, this extra note is the prominent ‘blue’ note of the flattened 7th. It is therefore often acceptable in the context of rock music. Harmony at a perfect fifth is usable except for the note a perfect-fifth above the leading-note of the scale which forms a tritone with the root note of the scale.
Track 22
George Martin claims the invention of flanging came about due to the slight lack of synchronisation between two ‘locked’ tape recorders. This effect caused various frequency bands to be alternately reinforced and cancelled; imparting on the captured sound a strange, liquidity – a kind of ‘swooshing, swirling’ ring. Of course, such an effect is not achieved nowadays using tape recorders; instead, electronic delays are used.
Track 25
Hymn to Aten – for soprano, orchestra and pre-recorded tape
Hymn to Aten was written for the soprano Jane Searle and the Kingston Chamber Orchestra and was first performed in February 1998. Technically the piece is in two halves. The first part (which is the only part recorded here) depicts the primeval and ritualistic elements of a brief, schismatic faith which blossomed in Egypt in about 1345 BC; based on the benevolent, physical presence of the sun. This part is prerecorded and extends the technique used in Birdsong (Track 52) of using effect electronics to process vocal sounds; sometimes creating noises (for instance the bell-like motif) which appear very different from the original material. Harmonically this section is bi-tonal (each group of chords being generated from a MIDI controlled harmoniser) but gradually resolving to a D minor chord for the entry of the orchestra and soprano.
Track 26
Uncompressed guitar – deliberately played with high dynamic contrast
For engineering purposes, it is often desirable to shrink the dynamic range of a signal so as to ‘squeeze’ or compress it into the available channel capacity. The studio device for accomplishing such a feat is called a compressor. When using a compressor, the peak signal levels are reduced in the manner illustrated in the following track.
Track 27
Note that the peaks are reduced but that the gain is not made up. Obviously this would be of little use if the signal (now with compressed dynamic range) was not amplified to ensure the reduced peak values fully exercised the available ‘swing’ of the following circuits. For this reason, a variable gain amplifier stage is placed after the compression circuit to restore the peak signal values to the system’s nominal maximum level. This is demonstrated in the following track.
Track 28
Same effect as Track 27 but with gain made up Notice that the perceptible effect of the compressor, when adjusted as described, is not so much apparently to reduce the level of the peak signal as to boost the level of the low-level signals; in other words, that the guitar is now apparently louder than in Track 26.
Track 29
Compressed highly distorted guitar
Unfortunately, compression brings with it the attendant disadvantage that low-level noise – both electrical and acoustic – is boosted along with the wanted signal. Notice the noise floor is unacceptably high. The solution is a primitive expansion circuit known as a noise-gate, the effect of which is to suppress all signals below a given threshold and only ‘open’ in the presence of wanted modulation. The effect of this circuit is illustrated in the next track.
Part 5 Synthetic sound
In white-noise, all frequencies are present (at least stochastically). There is therefore an analogy with white light.
Track 33
Often composers need a sound which is modified in some way. Examples of this include variations of low-passed filtered noise; so-called pink or red noise because again of an analogy with light.
Track 34
Band-pass filtered noise, if generated by a swept band-pass filter can be made to sound like rolling waves or the sound of the wind through trees.
Track 35
The power of the analogue synthesiser lies in its ability to cause each of its individual components to interact in amazingly complex ways. Fundamental to the whole concept is the voltage-controlled oscillator. This may be controlled by a switched ladder of resistances; perhaps by means of a conventional musical keyboard, as in this example, or by means of a constantly variable voltage, thereby providing a sound source with endless portamento like the Ondes Martenot and the Theremin. Alternatively it may be controlled by the output of another oscillator; the resultant being a waveform source frequency modulated by means of another. And perhaps this resultant waveform might be made to modulate a further source! By this means, the generation of very rich waveforms is possible and herein lies the essential concept behind analogue synthesisers. Some examples are given in the following tracks.
Track 40
Sampling; used to generate novelty backing!
Digital sampling systems rely on storing high quality, digital recordings of real sounds and replaying these on demand as shown in this simple example.
Track 41
The tough problem sampling incurs is the sheer amount of memory it requires. Sampling is well suited to repetitive sounds (like drums and other percussion instruments) because the sample is mostly made up of a transient followed by a relatively short on-going (sustain) period. As such, it may be used over and over again so that an entire drum track could be built from as few as half-a-dozen samples.
Track 43
Sampling is great until long, sustained notes are required; like the sounds generated by the orchestral strings or voices. The memory required to store long sustained notes would be impossibly large, so sampled-synthesis systems rely on ‘looping’ to overcome the limitation of any non-infinite memory availability.
Track 44
This particular track demonstrates Roland’s proprietary Structured Adaptive Synthesis (SAS) which is an eclectic blend of techniques, honed to give the most realistic piano sound possible.
Track 45
Composite synthesised track used to create slightly frenetic yet, nonetheless, ‘classical ensemble’ sound: Harpsichord – SAS; Recorder and Bassoon – LS Sound Synthesis/Wavetable; Strings – Yamaha Dynamic Vector Synthesis.
Track 46
Another composite ensemble with sampled drums; demonstrating a mix with conventional guitar and vocal track. This is a very cost-effective production technique because the majority of the ensemble can be prepared in advanced in a MIDI programming studio, making the acoustic recording stage very simple and fast.
Part 6 Theremin
One of the earliest electronic instruments, this monophonic (single tone) melodic instrument was originally developed in Russia in about 1920 by Leon Theremin. Magically, the Theremin player does not touch the instrument and has only to bring their hand or body within a small distance of a special aerial to control the pitch produced by the instrument. The Theremin is thus able to produce an endless range of frequencies from the subsonic to the inaudibly high in long sustained glissandi. Despite being very difficult to play, the Theremin has achieved undeniable artistic success. It may be heard in several orchestral pieces and has been used on many film and early TV soundtracks. Furthermore the Theremin remains the emblem of experimental electronic music. A status that it perhaps enjoys because it’s one of the very few instruments designed in historical memory to employ a truly novel playing technique.
Track 47
Unprocessed Theremin; some reverb added during recording
The Theremin used on this track has a particularly pure (close to sine-wave) output due to the purity of the RF oscillators and a linear mixer and demodulation stage. The original Theremin had an output nearer to the sound of a violin (i.e. with a large degree of even-harmonic distortion). Waveform distortion is achievable by various means, including reducing the purity of the original RF waveforms or arranging a non-linear detector circuit. However the preferred technique (utilised, for instance, by Bob Moog) is to start with a linear Theremin and to distort the waveform afterwards in the audio domain. Such a technique was used in the following track.
Part 7 Music mixing and production
The result of a commission to produce a dance piece for children for the Rainforest Foundation, Hey Bulldozer involves spoken word, a children’s choir, sound effects and a standard rock-band arrangement plus amongst other things; pan pipe samples and Spanish guitar. Interestingly, none of the piece’s component parts – the choir, the band, the narrators etc. – were able to be at the same place at the same time!
The first step involved making a gash mix from the original MIDI files and dubbing this onto a multi-track tape as a working cue track. The multi-track was then loaded in the back of the car along with an array of microphones, mixer, power amps and fold-back speakers and driven down to the choir. The children recorded several different ‘takes’ whilst listening to the backing via several fold-back loudspeakers carefully set so that they were loud enough for them to hear them, but not so loud as to cause too much ‘spill’ of this signal onto the final choir microphone signals. A week later, after the final tweaks to the MIDI sequencer data and samples were complete, the complete vocals track was formed by selecting the best bits from the various ‘takes’. The narrated speech and choral speaking excerpts (which were recorded a month earlier in a London dance studio complete with traffic noise!) were then carefully edited and ‘topped and tailed’ using a digital audio hard disk editing system and spun-in to a track on the multi-track by hand (or do I mean by mouse?). The two acoustic guitar tracks were then added, then the electric guitar track and then a backwards electric guitar which involved turning the multi-track tape over on the deck and having the guitarist play along to the other tracks backwards (something you can only do with an analogue multi-track machine!) And all that was left was to add the sound effects which involved a combination of real jungle sounds (flown direct from South America and on analogue cassette – arrrgh!) and library FX. Hey Bulldozer received several performances during the summer of 1991 and has been performed several times since in different versions.
Track 52
Birdsong demonstrates how sounds may be cut-up and rearranged using digital audio editing software. Here, the technique has been exploited to produce a musical composition based on the minimalist technique of taking tiny sound ‘bites’ and building them into larger structures. In this case the building ‘bricks’ are taken from a recording of the renaissance choral composer Victoria’s Responses. These are layered, reversed and mixed to produce a collage of sound. The title is from the book by Sebastian Faulks.
Part 8 Stereo techniques and spatial sound processing
Classical, crossed-pair stereo recording
When listening to music on a two-channel stereo audio system, a sound ‘image’ is spread out in the space between the two loudspeakers. The reproduced image thus has some characteristics in common with the way the same music is heard in real-life – that is, with individual instruments or voices each occupying, to a greater or lesser extent, a particular and distinct position in space. Insofar as this process is concerned with creating and re-creating a ‘sound-event’, it is woefully inadequate. Firstly, the image is flat and occupies only the space bounded by the loudspeakers. Secondly, even this limited image, is distorted with respect to frequency. (There exists an analogy with chromatic aberration in optics.)
Track 54
Track 53, FRANCINSTIEN processed
Happily there exists a simple techniques for both the improvement of existing stereophonic images known as FRANCINSTIEN and for the creation of synthetic sound fields in a 360 degree circle around the listening position (OM processing). These techniques are illustrated in this and the following track.
Track 55
Stand-In for an Echo was written during my studies with the Advanced Music Technology Group at Surrey University and was my opportunity to re-write history and re-score some of my favourite moments from a few classic ‘Noir’ films.
Technically, the music incorporates the discrete phasing technique which Steve Reich introduced in Clapping Music in 1972. All four instrumental parts are derived from the set of patterns played by piano 1. Piano 2 copies these patterns but drops on quaver-beat every 30 quavers, the flute drops a quaver every 20 quavers and the marimba drops a quaver every 9 quavers. The process is extended to incorporate several ‘tacit’ bars before any instrument starts. In other words, the phasing has ‘already begun’ before the music starts. Over the instruments are superimposed ‘sound-bites’ taken from the films of books by Dashiell Hammett and Raymond Chandler. These, monaural, sound sources are treated by means of the OM spatial sound processor to appear outside of the area bounded by the loudspeakers and even behind the listening position – without the use of extra loudspeakers. The title is taken from one a jotting in one of Chandler’s note-books; ‘I called out and no-one answered, not even a stand-in for an echo.’
All tracks on this disk © Richard Brice 1997. All rights reserved.
*These tones were not digitally generated and are not intended for precision measurement.