| Home | Acoustics | ||||
| Emotions and Music |
Representing Emotions in Music
Introduction
When this project was set, we were tasked with producing a number of pieces based on a common theme. The theme chosen was emotions and our chosen emotions were guilt and paranoia. In this essay I will write about our composition. I will describe how the emotions are to be represented and how the piece was recorded. I will discuss studio and on location recording techniques and technology. I will discuss creative and corrective procedures used in sound recording as well as good practice.
The theme was chosen by group discussions. Several themes were suggested but in the end the only one, which everyone in the group agreed on was emotions. Then the discussion turned to which emotion each of us would work on. Rhys chose guilt and it was later decided that we would work together and combine our emotions as they complemented each other nicely.
I chose paranoia for a number of reasons. Firstly, I thought it would be an easy one to represent by creating an unsettling atmosphere and by using sound effects to create “suspicious” noises, which could bring on an air of paranoia in the listener. Secondly, I chose it because I knew I would enjoy creating such an atmosphere and a soundscape. Thirdly, I had been writing lots of traditional pop songs and thought that something slightly more atmospheric and less traditionally structured would be a nice change. Few other emotions would give me the same freedom. Happiness would have been a totally different requirement. Depression would have worked but others in the group were keen to do this (although I believe in the end they changed to something different).
Anger, again would have been a different atmosphere and feel to the music altogether. Paranoia gave me a perfect opportunity to design a musical environment, which the listener could become absorbed into and that could easily be created using sound effects and musical devices. And finally, the idea of being able to unsettle the listener and leave them with the effect after the event appealed to my sense of humour.
Lewis (2000) lists some defining characteristics of emotions. First are the external triggers. Second are the internal brain receptors, which react to these stimuli. Third are the changes in somatic and psychological activity, which are the result.
We thought about situations, which could give rise to the extremes of these emotions. These might include childhood traumas such as an overly strict parent and remorse over a crime committed. We came up with very similar images of the situation we wanted to portray. We settled on a lone man in a spotlight in a dark room suffering from paranoid delusions about everyone being out to get him. On the floor is a trail of blood leading into a corner of the room where can be seen a dead body. The “lyric” we decided would be the paranoid rantings of this man. We will use chromatic harmonies (chords which do not fit into any recognised major or minor scale) and minor chords to give an unsettling feeling, which will leave the listener without a feeling of resolution. Sounds effects will be used to increase the degree of disquiet.
The form we decided on was basically six sections; three softer and three more chaotic. The idea is that the man will start out relatively calm but on the lookout and will work himself into a frenzy before realising that what he fears isn’t actually happening and so he will settle down again. And so the music will start out calmly and build into a cacophony, which at its climax will drop back down to relative serenity again. This will then again build to another “panic attack” before settling down a final time. This will then build to a “bloody” ending.
Sound effects will be used both as general noise to add to the unsettling elements of the piece and the chaos when required, and as specific effects such as phantom laughter and imagined, accusing voices to suggest the guilt and paranoia. Effects will be used to encapsulate the listener in an auditory environment in which they will become part of the events unfolding around them.
Initially I began the piece in Logic Audio using virtual instruments. These provided many sounds that we would not otherwise have access to. Many of these sounds were treated with delay and distortion to make them sound more random, chaotic and unworldly.
This was all done in the computer room where there are many computers available. This minimised time in the studio of which the college has only two to share between many students.
Mediums here include CD, DVD, SHVS (ADAT), minidisk, hard disk, or super 8 tapes. Digital signals are not actually recorded as audio signals but are encoded and this code is stored. One advantage of this is that any noise created by the recording medium is not heard because it will not be part of the decoded signal on playback. There are various standards in digital recording primarily determined by two factors: the sample rate and the word length.
An audio signal is fed into a digital to analogue converter. This converter will then take samples of this signal at regular intervals. If very short samples are taken at the rate of one per second what will be heard is a series of clicks interspersed with one- second gaps. This is clearly not sensible. If these samples are taken often enough however, the ear can be fooled into believing it is hearing a continuous signal. CDs for example sample the signal 44,100 times per second (otherwise known as a rate of 44.1KHz). This rate determines the frequency range of the sample.
“Nyquist’s theorem: A theorem, developed by H. Nyquist, which states that an analog signal waveform may be uniquely reconstructed, without error, from the samples taken at equal time intervals. The sample rate must be equal to, or greater than, twice the highest frequency component in the analog signal.” (Definition: Nyquist’s Theorem, 1996)
This means that CD’s can play back signals up to roughly 22KHz (the PC .wav format also uses the same standard). This is good enough for the human ear. However there are now schools of thought that say higher frequencies, even if they are outside the human hearing range, still influence that which can be heard. Therefore many modern systems record at 48KHz, 96KHz or even 192KHz.
Word length is the definition of each sample. CDs use 16 bits, i.e. a 16 digit binary number to encode the signal. The largest number that can be represented by a 16 bit digital number is 65,536. Therefore there can be 65,536 levels of differentiation between the quietest and loudest parts of the signal. If a signal tries to go above this digital distortion is recorded. Unlike analogue distortion this is not a gradual and desirable effect but produces a horrible digital noise. The 16 bits mean the dynamic range of a CD is roughly 96dBs, much larger than analogue. However because of the undesirability of distortion it is common practise to record at much lower levels to allow headroom, i.e. a space where the signal may rise without reaching the distortion level. This then means our dynamic range is reduced. With this in mind many newer systems use 18, 20 or even 24 bits for recording. Many of the digital mediums in use today are non-linear. This means that the recorder/player can leap almost instantaneously to any part of the recording.
We decided to use digital recording because of the increased dynamic range which would allow us to contrast our softer and louder sections with a greater amount of freedom and the lower noise floor meant that the noise would be less of a problem in our quieter sections.
The alternative to digital was analogue tape, which could come in several formats. These include multi-track ¼” (normally 4 or 8 track), ½” (4, 8 or 16 track) 1” (8 or 16 track) and 2” (16 or 24 track) reel-to-reel tape and cassette for mastering purposes.
The advantage of the wider tape (in addition to the greater number of tracks) is that there can be a wider space between and therefore more separation between the individual tracks on the tape. This means there is less chance of, for example, the bass drum on track one, leaking over onto the snare drum on track two. This effect is known as crosstalk. 2” tape normally also has a coating on the back to prevent ghosting or print through. This is an effect whereby, when the tape is wound onto the tape spool and stored the magnetism from each layer actually prints its signal onto the adjacent layers creating a ghosted image of previous or subsequent sounds on playback. All of these tapes can also run at different speeds, typically, 7.5, 15 or 30 inches per second.
“In general, background noise and susceptibility to drop outs (areas of tape where the recording material is absent or inefficient) is reduced as the recorded area of tape is increased.” (Robjohns H., 1997)
Magnetic tape, even when blank, also generates its own noise, as it is never completely demagnetised. In addition to using a higher tape speed noise reduction systems are often used to overcome this further. Most commonly it is the high frequency noise, which causes the bulk of the problem. There are various such systems. One type of system that has been commonly used are encode/decode systems such as the Dolby system. This boosts the higher frequency bands at record time therefore masking the noise on the tape by boosting the signal volume in relation to the background noise.
Then at playback time, it reduces the same frequencies by the same amount. This means that not only is the natural tone restored but while reducing the high frequency content of the material it is also reducing the high frequency content of the tape noise. Later systems split the signal into multiple frequency bands and process each separately. These require very accurate alignment of the recording and playback machines in order to be sure that the signals are encoded and decoded accurately. (Robjohns H., 1997)
Other systems are single ended and involve filtering or equalisation. These monitor the input signal and when the high frequency content falls below a certain level they use filters or equalisers to reduce it further. The better systems do this intelligently by attempting to differentiate between the wanted and unwanted elements of the signal. Expanders and noise gates can also be used. These again monitor the input signal and when the whole signal falls below a preset threshold they either cut it out completely (noise gate) or reduce it to a second preset level (expander).
Another problem with analogue tapes is the dynamic range. This is the range of volumes, which can be recorded before the signal either distorts due to overload or is lost in the sea of tape noise. With analogue tape this range could be pushed into the high 60s dBs i.e. the loudest sound in your recording could be close to 70dBs louder than the quietest one. With analogue tape however, distortion is less of a problem. It is a gradual effect which, depending on how much the tape is overloaded and the effect required can actually be quite a pleasant and in some cases desirable because of the perceived warmth it adds to certain sounds such as drums. Also, analogue recordings are linear, that is the signal starts at one point and runs continuously to the end point. This means that in order to get to a point 7 minutes into your recording you have to wind through the preceding 6 minutes.
I also recorded a number of real world sound effects using a minidisk recorder and a condenser microphone. A condenser microphone uses a pair of parallel magnetic plates separated by an electrically charged insulator. A power source is needed to maintain this charge and this is normally supplied in the form of phantom power from the mixing desk. As the plates move they change the current flow, which is sent through a resistor connected to the power supply. This voltage change is monitored by a high-impedance pre-amp, which derives the required signal.
Microphones come in a number of different types. The main two are dynamic and condenser although other types such a ribbon and boundary mics also exist.
A dynamic microphone uses a moving coil of very fine wire with a thin plastic film diaphragm attached, all enclosed in a permanent magnet. As the coil moves back and forth between the poles of the magnet (when moved by sound waves) it creates changes in the magnetic field, which generate an electric current which is then sent on into the signal chain as an electrical representation of the sound waves.
A ribbon microphone is similar to a dynamic except it uses a thin conductive ribbon in place of the diaphragm and coil.
A boundary microphone is normally attached to a flat surface, which it uses to pick up the sounds. The larger the surface, the lower the frequencies that can be recorded. The idea is that because of the mass of the boundary, sound waves are unable to move beyond it and the reflected sound creates a change in air pressure, which the microphone can detect. (White, P, 1990)
Microphones also come in a variety of pickup patterns. Omni directional picks up equally from all sides. Cardioid picks up more from the front. Hyper cardioid is even more directional that cardioid. And figure of eight picks up from front and back and less from sides.
The microphone I used was cardioid. This was most suitable because the effects I was after recording were focused in a small area (e.g. someone laughing) rather than spread over a large area. I therefore wanted a microphone, which would focus on a specific area without picking up lots of background noise.
Once I had these recordings I returned to the computer room and edited them in Cool Edit Pro. This involved identifying the sections required and cutting them out, normalising them (maximising their volume) and using noise reduction algorithms to try to filter out extraneous noise.
The particular filter we use takes a digital footprint of the noise to be removed. This means you need a section of the signal which contains only the unwanted noise and no desired signal. The software then analyses this “footprint” and designs an algorithm to remove all of those elements from the signal. This is then applied to the desired section of the recording and all of the background elements are removed, hopefully without affecting any of the required elements.
Other filters use preset algorithms or simply filter out the higher frequencies in quieter sections. This relies on what is known as the signal to noise ratio. Basically, in the loudest sections it is assumed the signal will be louder than any background noise and so will cover it up. When the signal becomes quieter the filter kicks in and removes higher frequencies. The quieter the sound the more filtering is applied. It is assumed that the background noise is likely to be in the high frequency range, possibly because this was largely the case with analogue tape noise. To get rid of lower frequency noises such as mains hum other methods will be needed (we didn’t encounter mains hum during our recording so for us this was not an issue).
Parametric equalisers (which allow you to choose not only the central frequency affected but also the width of the range of frequencies around that central frequency) can be used to focus on a very narrow frequency band for just such a purpose. Very narrow cuts are rarely noticed by the human ear which is much more sensitive to boosts than cuts.
Noise gates are another option for removing noise. These are similar to noise filters but when the signal falls below a certain level, rather than applying a filter they cut off the signal altogether. This is good for removing the noise in silences between sounds and therefore useful to apply to individual instrument tracks. They must be set carefully however as otherwise they may cut out wanted, quieter notes in the performance. They may also cut short the decay portion of sustained notes.
Once edited the files were saved in ‘.wav’ format for use in logic. Once I had my MIDI recording and my location recordings it was time to go into the studio. Here we needed to, record guitars, drums, and vocals all directly into logic on the G4 mac in the main studio through two MOTU 2048s. These have audio inputs (fed from the patchbay) and convert the audio signals into digital signals, which are sent down a single cable into a PCI expansion card inside the mac. This card acts like any other soundcard and interfaces between the computer and outside audio hardware. We also had access to a MOTU Timepiece, a device used for synchronising different parts of the studio (e.g. 2 computers, a computer and a tape machine, a tape machine and a video recorder) but we had no need of this as everything we required was available within logic.
It was decided that the guitar would be recorded through a single amplifier (a Behringer GMX100) but using two microphones. It was decided that two microphones placed at different distances would give us two slightly different tones, which we could then use in combination to provide a thicker texture. We could have used more microphones but this would increase the likelihood of phase problems, the worst of which is phase cancellation.
Phase cancellation is a situation where multiple microphones are picking up a single sound source but because of the different distances there is a slight delay between the signals. When these signals are combined it may be the case that the waveform of one signal is rising as another is falling and they therefore cancel each other out. That is the falling waveform subtracts itself from the rising one and in the worst situation creates complete silence. IT is more likely however that they will just sound very thin. Some mixers allow you to reverse the phase of the signal passing through them, which means you can adjust the falling waveform to be a rising one. While this can be a solution it is best to avoid the problems at source where possible. And more than two microphones would have made things more complicated without any significant benefit.
One microphone was placed closer to the amplifier than the other in order to pick up more of the direct sound. The other was placed slightly more distant to gather some of the room ambience.
We used dynamic mics on the guitar amp (and on most of the drums). Owing to the loud volumes involved, delicate nuances of more advanced mics are not an issue and dynamic mics are very good at handling large volumes. We could have used a condensor mic but as well as the potential sound pressure level problems (i.e. the high volumes overloading the mics) these would have given a warmer sound which was the reverse of what we were trying to achieve. Specifically we used an SM57 (an accepted standard for guitar recording) and a Sennheiser E845 (equivalent to the other Shure standard the SM58 which were all unavailable at the time of our recording).
We experimented with which to use as the close and which the distant and eventually chose to have the E845 as the ambient mic as the SM57 seemed to get a crisper effect on close miking than the Sennheiser.
Guitars were recorded through effects pedals before being fed into the amplifier. This is because the texture was the most important element and so getting the effects at source influenced what was played. Generally studio practise dictates that all sounds are recorded flat (i.e. without effects or other treatments) and for all other instruments this was the case. But because the sounds were so integral to the performance it was decided that these effects should be applied at source in order that the guitarist could actually give the required performance.
The effects in question were an Electro-Harmonix Holy Grail reverb on the “Spring Hall” setting and a Boss DD3 delay pedal. These effects were a key part of the sound and greatly affected what was played during the recording. An e-bow (a device which causes the string to vibrate without being plucked and sustains the vibration for as long as it is held over the string) was also used to get very long sustained notes, which would evolve over time as they were played adding to the atmospheric effect.
We recorded several takes of guitar, which were layered on top of each other to create an ambient wash of sound.
The mics were fed through a multicore into the control room. Here they were patched, though a patchbay into two of the mixer’s mic inputs. Personally I prefer the mic pre-amps in my Mackie mixer at home as they are quieter and give a much truer sound with less colouration that the ones on this Soundcraft desk. But there wasn’t really a choice here as there is only the one mixer in the main college studio and it is the best mixer in the department.
The gain levels on the desk inputs were set so as to get the optimum volume level into the mixer without incurring distortion. This is done to maximise the signal to noise ratio in the mixer in the hope that the signal coming in will drown any noise generated by the mixer’s internal circuitry. The meters were watched carefully as the guitarist performed. If during the loudest part the meters went into the red section of the meters the gain was turned down as this indicates distortion (if necessary the microphones may need to moved or amplifiers turned up or down but this also affects the sound quality as well as the volume).
Once in the mixer the sounds were routed to output subgroups. These are hardwired into the soundcard inputs on the studio mac. We routed our guitar mics to group outs one and two and panned one hard left and the other hard right. This does not actually affect the left right placement in this particular instance but merely differentiates between subgroup 1 (left) and subgroup 2 (right).
Logic was set to receive signals from input one and two and levels were then set using the subgroup level faders in the same way as before i.e. the volumes were increased as far as possible but if Logic’s meters went into the red the levels were dropped. This again means maximum signal into the device without distortion.
During recording we used two auxiliary sends on the mixer (outputs which take an additional feed from signal channels which are then combined and routed out of the mixer) to feed a headphone amplifier. This was used to feed a rough mix into headphones so the performer can hear what is already recorded and if necessary their own performance. The relative levels of the elements in the recording are determined by the amount being fed into the auxiliary bus from each channel. This is determined by the individual channel auxiliary send controls. These sends can also be used to allow many channels to share a single effects unit such as a delay or reverb. These are no good however for effects where the entire signal (as opposed to a parallel copy of it) need to be treated e.g. a noise gate.
The quality of a sound recording is influenced by many factors. These can include noise, the standard/medium used to record, the equipment used, the sound source itself, the recording location, the techniques used and the crosstalk and print through mentioned above. Noise can have many sources. Analogue tape produces its own noise. Equipment such as mixers, amplifiers and cables all produce noise. And the environment can also produce background noise.
Using the best quality equipment is essential. The quality of cabling and components used will determine the amount of noise created by them. Controlling background noise can also help although with location recording this may not always be possible. Noise reduction (as mentioned above) can also help. In studios, large sums of money are spent on sound isolation in order to reduce background noise. If doing a “live” take instruments will be placed in separate rooms as much as possible or at least separated by acoustic screens. Vocalists will almost always be put into a separate vocal booth.
All monitoring was done on closed back headphones in order that microphones will not pick up other instruments from monitoring.
Sound sources should always be prepared as best they can. This may mean new guitar strings, tuning drums, etc and getting the best microphone placement. Although sounds can be treated later there is no substitute for getting the sound right at the source. In our piece this will be vital on the guitar as the sound is the most important element, more so than the actual notes played.
When in the studio there are a number of things that can help the recording process. Being well rehearsed is important as studio time is expensive and shouldn’t be wasted learning songs or parts. Being well prepared with chord charts, lyric sheets, spare strings, guitar tuners, etc. is also important. The producer engineer should maintain clear and accurate track sheets so everyone can see at a glance what is recorded where and what needs to be done on each track to complete it. Finally, having trusted personnel is important. If the band all know they can rely on each other and if they have faith in the producer and engineer things will run much more smoothly.
Rehearsal was not a huge issue for us as most of the piece was written on the computer and the sounds and effects were an integral part of the composition and were created as the piece developed. The guitar and drum parts were largely improvised although some planning and discussion took place beforehand. All tracks were clearly named with what they contained
Various recording options are also available. Some instruments, such as keyboards and drum machines will sometimes be recorded directly from their line out to a line in on the mixing console. In more professional setups the instrument will be connected via a DI (Direct Injection) box. This takes the signal from the line out and reduces it to a microphone level signal. This is the fed along a microphone cable to a mic input on the mixer. This has two advantages. One is because mics use what are called balanced cables, which are less susceptible to interference and noise. In a balanced cable two versions of the signal are run along the cable, one of which is 180 degrees out of phase with the other i.e. where one wavelength is rising the other is falling by the same amount. When the signal reaches the mixer it is turned back into phase so that the signals match. When these newly in phase signals are combined, any noise or interference picked up along the cable run will be turned out of phase and cancel itself out. The other advantage of DI boxes is that generally the mic pre amplifiers on mixing consoles are better quality than those on the line inputs.
Drums, vocals and guitar amplifiers will normally be recorded via microphones. Guitars are almost never plugged straight into the recording equipment, as the amplifier and speaker used are normally an integral part of the guitar sound. In some cases guitars may be plugged directly into a modelling amplifier, which will attempt to emulate the amplifier and speaker effect before going straight to the mixer either directly or via a DI box.
Drums can be recorded in a number of ways. The simplest method is a simple pair of stereo mics (see below). More typically this is supplemented by several spot mics on particular drums such as the snare and bass drum. Some producers will mic every drum individually and even have mics on both top and bottom skins of each drum and on the hi-hat cymbals along with the stereo pair. This gives more flexibility in the balancing of individual drum levels. The setup we will use is similar to this latter description.
There are various methods of stereo recording. Two microphones can be placed at a distance on either side of the source or even on opposite sides of a room. Alternatively the two microphones can be close together with their bodies crossed in an x shape. This is to mimic the spacing of the human ears. Some producers actually have a dummy head and place two small microphones inside the ears of the dummy. Another alternative is to use a single microphone with a figure of eight pickup pattern (see below).
The only stereo recording we did was for the drum overheads. We used AKG C1000s for our overheads and the D112 for the bass drum. We used condensers for drum overheads as they pickup the harmonics of the cymbals better and are good at recording more distant sounds and will therefore get a better overview of the whole drum kit than would be possible with the more limited pickup range of a dynamic. All other parts on the kit (including the bass drum) used dynamics. Toms were miked using Audio Technica Pro 25s and the snare an SM57 above and below with the same on the hi-hat. The SM57 under the snare is useful to pick up much more of the crispness of the snares whereas the one above gets more of the click of the stick on the upper skin. Experimentation was used to determine the correct balance of the two mics to get the overall required snare sound.
The overheads were placed approximately one meter above the kit and one metre apart. This meant they were far enough away from the kit so as to allow the sound waves to blend and not allow any one element of the kit to dominate by being to close to the mic. It also allowed sufficient separation of the left and right sides of the kit without sounding so far separated as to appear unnatural.
In the mix they are panned to 10 o’clock and 2 o’clock for this same reason. Panning refers to the left right placement of sounds within the stereo field. The sounds were routed and balanced as with the guitar.
We wanted the drums to be atmospheric and add another element to the collage in the quieter sections and to add to the chaos in the louder sections. To this end we did not want the drums played in rhythm with anything else. However, the drummer struggled with this and kept playing in time no matter how mush we told him not to. In the end we took the backing track out of his foldback (headphone) mix and just told him, through the headphones, when to start and stop and when to get louder and softer, faster and slower as Rhys and I listened to the backing track in he control room. This worked well.
For vocals I worked alone in the control room on headphones. I used one of the C1000s and heavy compression (on the recorded track after recording) to get a more intimate sound. A compressor is a kind of automatic volume control, which is used to even out large difference in volume levels in a performance. It monitors the signal and when the sound goes above a defined threshold it reduces the amount of level increase by a defined ratio. E.g. a ratio of 2:1 means that for every 2 dB increase in input above the threshold the output will only increase 1 dB.
During all recordings loose cables were taped down to avoid tripping hazards and no food or drinks were allowed near equipment. All equipment in the college is also regularly tested for electrical safety.
We used cardioid pickup patterns throughout as the items we recorded were not spaced over a large area as a choir or orchestra might be.
Another method of recording is bouncing down. This was initially used by people such as the Beatles to compensate for only having four tracks to record on. They would fill three tracks with instruments and then re record those three tracks through the mixer onto the fourth track freeing up the original three tracks to be reused. The disadvantage of this of course is that the original three tracks cannot then be rebalanced against each other later. On newer computer based systems bouncing down can also be used to free up some processor power when large numbers of virtual instruments are used. The instruments can individually be bounced down to audio tracks and the virtual instrument can then be bypassed. In this case the audio tracks can still be balanced and treated individually and the processor power which was used for the virtual instrument is freed up for use elsewhere.
Creative methods of recording can be used to create special effects. In the early days of analogue tape producers would sometimes turn the tape over backwards and record instruments (hopefully realising that the track numbers were now reversed). When the tape was turned back the right way around the instruments so recorded appeared, to play backwards. Other effects used have included recording taps and toilets cisterns to achieve a stream effect.
Peter Gabriel has recorded African tribal drums and used them as the basis of a song. He also used African vocal chanting as one layer of his song “Biko”. The vocals start and after around 30 seconds studio recorded drums come in in a totally different tempo and with no attempt at synchronisation creating a polyrhythmic effect. This is similar to what we attempted to do with our live drummer (see above).
Kraftwerk in the song Authobahn synthesised the sounds of car engines and horns and rhythmically recreated the sound of engine pistons etc.
“Walk in the streets and you have a concert, cars playing symphonies, even engines are tuned, they play free harmonics. Music is always there – you just have to learn to recognise it.” (Hütter R., n.d cited in Gill A. 1997) We used synths to create more abstract effects and also location recordings of laughter and voices.
To create the original gated reverb sound Hugh Padgham got Phil Collins to play the drums in a very live room. As well as close miking the drums he also placed two very distant mics on the far side of the room to get a very ambient sound. Realising this ambience was making the mix very muddy he used a noise gate to cut the reverb off once it fell below a certain level so as to create space between the drum hits. This effect was so successful at creating a powerful drum sound; the effect is now built into almost every effect unit.
"“They set me up in the Townhouse's stone room, and I played a bit while they fooled around with miking combinations, compressors, noise gates. All of a sudden, I heard this sound...."
That sound--that heavily gated, open-and-closed effect--inspired the drummer. He started playing a pattern that timed perfectly with the opening and closing of the gated reverb. "I just started fooling around with this beat, and the next thing I know Peter is shouting over the intercom, 'Don't stop! Give me that for ten minutes--just that!” (Miller W.F. 1997)
On Queen’s A Night A The Opera album, Roger Taylor vocally recreates various brass instruments.
“This track resembles the upper-class atmosphere of Lazing on a Sunday Afternoon and is really 'over the top' if you don't like the humour of it. It also features a weird brass section, which is no-one else than Roger Taylor, playing with his voice again.” (Jaap De Haan J., 1999-2005)
And on I’m Not In Love; 10CC recorded 144 tracks of vocals (presumably including some track bouncing) to create the choir vocal effect in the background. This is similar to what we did with our guitar parts, layering them on top of each other to create a dense wall of sound.
On early Who recordings, poorly setup compressors resulted in pumping cymbal sounds, which have been said to add to the dynamics of the recording.
Once recorded sounds can be treated in many ways. Special effects such as delay, reverb, chorus, phasing, etc can be used to achieve particular effects. Cut and paste can be used to move and copy elements around the piece e.g. a duff bass drum part could be replaced with a better take from elsewhere in the piece. Individual parts can even be edited out completely such as extraneous noise at the beginning or end of a take.
Reverbs and delays will be a vital part of creating the atmospheres and chaotic environments we intend and will be used to some degree on most elements of our piece. It is likely that the effects, by determining the sounds will also affect what is played using those sounds.
Noise filters/gates can be used on individual tracks to automatically take out noise on individual tracks, even between notes if correctly setup. E.Q. can be used to achieve a particular effect, such as the telephone voice, or for reducing unwanted resonant frequencies.
Pitch correction can be made using devices/software such as Antares Autotune, which will monitor an input signal and automatically pull all pitches to the nearest pitch within a preset scale. Samples, for example, can also be adjusted in pitch to fit in with other instruments by speeding them up or slowing them down. This will of course also affect their tempo so my not be so useful with rhythmic sounds.
Time correction can be used to adjust the tempo of loops for example or sounds, which are slightly out of sync, can be manually edited into place. Adjusting the tempo will of course affect the pitch of the loop, which could be a problem. Beat slicing (where each individual beat of the loop is made into a separate sample, which can then be re-spaced played back at different tempos without affecting the pitch, can also be used.
One other creative effect uses the side chain of effects such as compressors and noise gates. The side chain allows the effect to be triggered from a source different to that, which is being processed. For example a bass drum sound could be used to trigger a noise gate, which has a bass guitar fed into the input. This would result in the gate only allowing the bass guitar to be heard when the bass drum played resulting in a bass line very tightly allied to the bass drum.
During the mixing process all of the recorded, sampled and other sounds are combined into a stereo mix. At this point the relative levels of the sounds are adjusted and their placement within the mix is set, i.e., left-right panning and rear-front placement. By adding reverb a sound will appear to be pushed to the rear of the sound field. Special effects such as timed delays, phasing, harmonisations using pitch shift, etc. can be added. Noise gates, filters and expanders can be used to reduce the noise on individual tracks and also the overall mix.
Physical editing can also be done e.g. to add silence at the beginning and end of the finished recording, to change the order of sections or to splice together different takes.
The final stage of the process is postproduction. Here an engineer will adjust the final mix to fit the medium on which it will be distributed. Processes used may include compression and limiting to balance the overall level of the piece. This would normally be a multi-band compressor, which would split the signal into different frequency ranges and process each separately. Equalisation may also be used and an overall reverb may be added to give the mix the feel of having been made in a single, uniform environment.
Our piece was mixed in the small studio in order to free up the larger one for others. The mixing was done internally in logic using the virtual mixer and mix automation (this is the only automation available in the college although some facilities have hardware mixers where settings can be recorded to a timeline and in some cases the physical controls will even move as this timeline is played back).as well as the internal virtual effects. Lots of delays were used. Some filters and distortions were used. The guitar tracks were cut up quite extensively in order to take out sections of noise inbetween notes. Also at one point in the drum track a loud click occurred owing to drummer error. Fixing this involved cutting out a small section of audio from each of the drum tracks. However, while this was fine on most tracks cutting it out on some caused a noticeable dropout in the track. On these tracks we therefore used logic’s track automation to do a more controlled fade at this point. This seemed to work well allowing the drop off and return to be more gradual and also less total dropping down low enough for the click not to be heard without cutting the audio altogether.
Finally we bounced the finished mix down into a stereo master, 16 bit, 44.1 KHz .wav file ready for burning to CD. A small amount of mastering reverb and compression was applied to this stereo master before burning using iZotope Ozone and the beginning and end were trimmed in Cool Edit Pro where the track was also normalised (volume maximised). Izotope Ozone is a software plugin to logic. It is designed for the post-production mastering of stereo files. It features several distinct sections such as mastering reverb, to be applied over a w hole piece to give the specialisation a homogeneous feel so it appears the whole track was performed in a single location. It also contains a number of multiband effects. The one we used was the compressor. This divides the stereo file into a number of frequency bands and applies compression separately to each band. This is useful because a sudden loud bass drum beat for example will not pull all of the treble frequencies down in level making the effect highly noticeable. The effect of this multiband treatment is too make the effect less obvious. Treatment applied to one frequency band will not necessarily affect any other band thereby allowing the unaffected bands to mask the effect.
So how do we create these the previously discussed emotional causes and effects musically and thereby trigger these emotions within a sound environment? Attempts to explain the effects of music on emotions are normally split into two types of theories: heteronomist and autonomist. The first school of thought argues that the emotional response is caused by a direct appeal to our inner selves. The second claims that that the effect is due to a unique property contained only in music (Hanslick, n.d).
Music like language is organised sound but does not have the referential basis. Hanslick (ibid) argues that emotions have no isolated existence and that sadness assumes a previous state of happiness. He argues that the aesthetic value of music lies in the co-ordination of intrinsically pleasing sounds.
Emotions however, are unique to language. Each word (anger, happiness, fear, etc.) are understood by speakers of English. Some English emotions however, do not translate into other languages. Similarly words in other languages pertaining to emotion do not translate in English. This is because different cultures have a different perspective on emotions.
This of course makes the universal representation of an emotion in a more universal language such as music more problematic.
Although of course music is not always universal. Different scales are used in different parts of the world. The only thing that is common throughout the world is the octave, which appears almost universally in all cultures. Different cultures however, divide the octave in different ways as a reflection of their culture.
Thompson & Robitalle (1992, cited in Huron D. n.d) claim that composers in the west can communicate emotional states through a monophonic melody.
Plato (cited in DeBellis M. 2004) claimed that the emotional effect of music was as a result of the patterns within the music mirroring speech patterns of people feeling the particular emotion.
This again means the representation of an emotion is close to impossible on a universal scale as different cultures will have different points of reference, different responses and even different emotions. I decided that realistically we would only aim our piece at western listeners as there was little likelihood of it being heard in other cultures. I also realised that by using the kind of chromatic harmonies we planned our music would be unlikely to give a sense of resolution in any other scale . Also the floating, atmospheric environment we intended to create, I hoped would be considered atmospheric in any culture. However, I had to accept that paranoia itself may not exist in some of them and it would be unrealistic to try to represent something that didn’t exist for those listeners. I hoped however, that whatever name it went by, if it was heard in other cultures it would have the effect of unsettling the listener and making the apprehensive about what they heard.
Hanslick (ibid) claims that music cannot express the clear thought that can distinguish definite emotions. Music therefore can only give a vague idea of an emotion. This would tie in with the difficulties of universal translation. The concept of feeling and emotion is universal. It is just some of the specifics that are not. And so it can be argued that music can produce a vague feeling or mood, which is then translated by the listener into a specific emotion for his or her frame of reference. This all ties in with our aspirations above.
Nawrot (2003) carried out studies where adults were asked to listen to pieces of music and match them to photographs of different facial expressions. Most adults were consistent in their choices (although children less so indicating that they have yet to learn the cultural references or the intricacies of musical representation of emotion). Also when asked to produce spontaneous verbal indications of the emotions most adults corresponded in their choices indicating that generally, within a given culture adults produce the same emotional response to a piece of music.
However, it has been shown that patients with a mental disturbance can have different emotional responses to healthy people and that those responses can vary according to the type of mental disturbance
Numerous studies have been done which prove that music can have a physiological effect on things such as heart rate and blood pressure. If music affects the physical state then surely the emotional state is also likely to be affected as the two are interrelated.
For centuries it has been thought that different keys created different moods. Banks took this further and did research into the effects of different tempos, timbres and modes on conscious perceptions of emotion and found that these did indeed have an effect. However, none of this research covered guilt or paranoia.
I decided that a slower tempo would be likely to have the effect that we required. The slower tempo exemplified in the quieter sections of our piece would be likely to slow the listener’s physiognomy make them want to sit and listen and be drawn into our piece. The timbres and harmonies here may entice them by making them curious but would hopefully not allow their heart rate to settle to the point of relaxing them. Even if this did happen, I hoped the intended faster, more chaotic sections (see below for a description of our piece’s structure) would have the reverse effect and get their blood pumping and that the irregularity of the beat in these sections might even throw out the bodies rhythmic responses to create further unsettlement. This hopefully would raise the listener into a state of apprehension before allowing them to settle slightly once more during the following, softer section.
Major keys are widely accepted to feel happier than minor keys. Major keys would generally be chosen to represent happier, more upbeat emotions whereas minor keys would normally be chosen to represent more negative emotions. Possibly this is in part due to the more chromatic nature of minor keys owing to the number of accidentals required to create the numerous forms of the minor scales. Chromaticism generally creates a more unsettled feeling owing to the lack of resolution and this unsettled feeling can be used to recreate a feeling of unsettling emotions. Chromatic scales and chords sequences, which do not fit into diatonic keys are useful for such representations.
In diatonic harmony, typically we would expect to hear cadences (a short sequence of chords suggesting the end of a phrase) of chords four to one or, most commonly, five to one. Over the centuries these two cadences have been accepted as having a feeling of finality and are used to end a phrase, section or an entire piece. They give a pleasing feeling of resolution and allow the listener to settle having been brought to a satisfying conclusion in their aural experience.
In our piece we have avoided any such cadences and therefore any such sense of resolution so as not to allow the listener to fully settle even after the piece has finished. Throughout our piece we will not allow the music to settle and will hopefully heighten the listener’s sense of unsettlement as the piece progresses.
Dischord also creates an unsettling feeling owing to the unnatural and unresolved combination of sounds, which cry out to be resolved.
Combinations of notes played together produce a periodic beating caused by the interference of the different waves, which alter the overall amplitude periodically. Certain intervals between the base notes cause more beats in the upper harmonics than others. The intervals with the fewest beats are said to be consonant. Dischords contain more a complex relationship in the series of harmonics they generate and possibly this complexity contributes to the unsettling feeling they cause. Indeed there becomes a point where the human ear is unable to distinguish individual harmonics because they are too close together and this is perceived as a very harsh sound.
Diminished and augmented chords for example create a sense of tension, which leaves the listener crying out for resolution and causes the emotional state of the listener to reflect this tension and desire for resolution. When this resolution is not forthcoming the feeling of tension heightens and the emotion builds.
In our piece there is no specific use of dischords although the guitar parts float ethereally above and may inadvertently create dischords against the backing. I decided early on that chords such as augmenteds and diminisheds would be too jarring to the ear and detract from the atmosphere I wanted to create. I wanted to draw the listener into the piece and such dischords may have had the opposite effect. Instead I chose to rely on the chromticism in the chord sequences to create an atmosphere that, while unsettling, would be smooth sounding enough not to discourage the listener from wanting to hear more. Although the guitar may create such dischords the flowing, smooth and ambient nature of the sounds leads the listener to treat it as atmospheric sound effects more than part of the harmony. The sound does not jar because the listener is smoothly and gently led from one long flowing note to the next.
Of course different types of brain waves (Alpha, Beta, Delta and Theta waves) occur at different frequencies and it is possible that different pitches affect the brain in different ways.
In our piece we will employ descending sequences of notes to hopefully create a descent into a downbeat mood in the listener by leading them through the different types of brainwaves into a downward spiral of emotion. This again will be reinforced by cultural references, which equate being “down” with depressive moods and being “up” with happiness.
All of these methods will be used to produce a feeling in our listener that all is not well and to draw them into our story and to help them experience the fears and disturbed and unsettled nature of our character.
It has been shown that music can produce similar emotional responses in adult listeners and that particular kinds of chords, harmonies and scales can influence the kind of response. The use of minor chords for sad music and major chords for happy music, chromaticism and dischord for an unsettling feeling. We used many such elements to represent the unsettled mind of our character.
During the recording of our piece used electronics to create some special effects and instruments such as strings, synth sounds and piano. This is because some of the effects we have found fit so well with our ideas and the versatility of virtual synths over the hardware we have available in the college (e.g. no string sections, few real synths).
Live sound effects were recorded on location using a minidisc machine with an external condenser microphone. They were transferred into Logic along with all other elements ready for final mixdown after editing, in cool edit pro.
We used digital recording equipment, both fixed and mobile, to record our piece, which will include, electronically created sounds, real instruments and real world and artificial sound effects. We decided to use digital recording for three main reasons. The low noise was important during the quiet passages. The wide dynamic range were important to maintain the contrast between the frantic sections and the calm ones. Finally, for the ease of editing, which was useful for unusual effects.
We used sound treatments such as distortion, reverb and delay to add to the chaotic effect required in parts and also the atmospheres required in softer parts. We used chromatic harmonies, minor chords and a lack of cadences or any real resolution to create a feeling of disturbance. And we used a combination of music, sound effects and narrative to express our chosen feelings of guilt and paranoia.
This project inspired a new style of composition from me. Being asked to produce music to a theme in this way inspired us to create music that would not otherwise have existed. It instigated thought processes and creative stimuli that would not otherwise have occurred. It caused me to think harmonically, and technically in ways I would not otherwise have done. While many of the processes are fairly standard this project inspired me to use them where other projects may not have.
Donate to new music
business
Link to this site
This website copyright 2004-2009 Maddening Music except where otherwise stated.