Chapter 9

Sound Matters

In this chapter, I concentrate on one of the most important aspects of an editor’s life, and it’s strangely not the vision. Whereas the eye can be very forgiving about what it sees, the ear is always supercritical of what it hears, which means we just have to get the sound right.

Stranger still, you’re about to find out that audio throws up many more fiddly, technical issues than video.

The reason for this is that video, including its levels, content, and technical range, is largely handled by the software. With audio you have the ability to mix different sources together that can sound absolutely dreadful, be technically untransmittable, or indeed just be downright bad, and the software doesn’t give a damn and will happily let you get on with it!

It is for this reason that this chapter is very detailed and comprehensive.

The chapter is divided as follows:

• 9.1 Get the Room Right First
• 9.2 Sound Levels
• 9.3 Sound Monitoring
• 9.4 Split Tracks
• 9.5 Wild Tracks
• 9.6 Sound Effects
• 9.7 Sound Mixing
• 9.8 Sub-Frame Editing
• 9.9 Sound Processing
• 9.10 More Specialised Sound Processing
• 9.11 Additional Dialogue Recording
• 9.12 Foley Sound Effects
• 9.13 Key Points—Sound Matters

9.1 Get the Room Right

Traditional nonlinear editing rooms, especially if they’re primarily used for cutting offline material, are too often equipped with less than perfect sound monitoring. It is essential before you start doing any serious alterations to the sound that you make sure your monitoring is set up to do the job.

You can do more harm than good if you start rebalancing sound using indifferent equipment. This could not only harm the programme, but it could also harm your reputation.

To start with, the room in which you are working should be sound-treated, at least with carpets, soft furnishings, and curtains, so that unwanted reflections and standing-wave effects are minimised.

Go on, insist on better speakers, get them onto a proper pair of stands, and put some curtains up!

 Neighbours—Everybody Needs Good Neighbours (and Keeps Them!)

It is equally important that your room is sound-proofed to some extent so that you can not only be free from distracting external noises but also be able to turn the wick up without disturbing the neighbours. Computers, drives, and air conditioning are all sources of noise in an edit suite which ideally need to be minimised before serious sound balancing can be attempted.

Alternatively, you can use a pair of decent headphones, but be careful, as they can cause you to produce oddly balanced mixes on some material. Definitely stay clear of noise-cancelling headphones, as they all have some form of inbuilt sound processing that could influence your mix strangely. If you have to use headphones, make sure that at some stage in the process you listen to your mix on decent monitoring loudspeakers. In good dubbing suites you will also find a pair of very ordinary speakers, so that the seven-channel mix can be listened to on the equivalent of TV speakers. Not a bad idea for the edit suite.

9.2 Sound Levels

Location sound recording is one of the most difficult things to get right on a shoot. Very often, the conditions for filming are perfect on one day, but too easily turn into a perfect storm the next. Even on the same day, shooting in different directions can produce wildly different sound backgrounds. Even an innocent bunch of trees gently blowing in the wind will produce different textures when recorded from different angles, and these differences will only reveal themselves when you attempt to cut between those shots in the edit suite. Some of these differences are simply down to level or volume and are easily corrected, but others are more difficult to cure.

 Peak Practice—Adjusting Sound Levels within Limits

Never be frightened to push volume levels around. However, as you would expect, this cannot be done without limits. Sound, like vision, has to exist within certain electrical limits of level. This used to be a voltage in the old analogue days, but now it’s a ‘bit’ number.

In either case, if limits are exceeded, the result will be peak distortion. Here, the sound becomes hard-edged and painful to listen to. In the digital world, the binary word representing the analogue voltage is at a maximum of all 1s or all 0s so that an even higher input level will continue to produce the same all 1s or all 0s result. In these circumstances, the signal is said to be squared-off. The great feature of digital signals is that their levels are defined and aren’t subject to the unwanted variations in gain or attenuation as suffered by their analogue equivalents when passing through the circuits of tape machines and other equipment in an edit suite of a few years ago.

In order that we all keep within the speed limit, a reference or line-up level is established. This reference, usually in the form of a burst of tone, is put, or should be put, at the start of any camera recording, whether on a file or on a tape.

 The Sound Barrier—What Is a Decibel?

We were bound to get caught up in technicalities at some stage, and it’s funny that the seemingly simple subject of sound brings these problems to our attention. To talk sound, we have to talk decibels (dBs). A decibel is a unit of a logarithmic scale used to describe voltage or power relationships. A logarithmic scale of volume is chosen to better match our hearing because our ears hear sound levels logarithmically.

In electrical terms, doubling the voltage of any analogue signal causes a gain increase of 6 dB.

Mathematically it is:

where G_dB is the gain, V₁ is the new higher voltage, and V₀ is the original voltage.

I hope that wasn’t too bad; just take away the fact that we describe sound in decibels with respect to an absolute maximum level.

As we have seen, the absolute maximum (all 1s or 0s) is dangerous or fatal to our signal quality. We therefore need a buffer zone to keep us from getting near this limit. To do this, a working maximum (or peak) is set up to be 10 dB lower than the absolute clipping maximum. In this way, a buffer zone is created (which is known as headroom) and allows for level transgressions that will not result in peak distortion and will therefore be capable of correction. It follows that no sound on your timeline should be louder than minus 10 dBFS (FS standing for full scale).

The reference level I referred to earlier in the previous section is usually defined as minus 18 decibels (–18dB) under the digital maximum level of all 1s. This reference level is more accurately described as: –18dBFS.

Some manufacturers take this reference level to be at –20 dBFS (or even –12 dBFS), so make sure your software matches your friendly broadcaster’s specification. Most editing software allows you to set this reference level to whatever you require.

I did warn you this would be comprehensive, didn’t I? Pulling all this together:

• The acceptable range of normal audio levels runs from: minus infinity dBFS (or silence) to a working maximum (or peak) of minus 10 dBFS.
• Line-up level is at minus 18 dBFS (or 8 dB under the working maximum).
• A headroom buffer is defined from minus 10 dBFS (the working maximum) up to 0 dBFS (the absolute maximum).
• To check that our levels are in the acceptable region, we need a meter of some kind.

9.3 Sound Monitoring

 See Hear—Looking at Sound

The bar graph level meters available on much editing software are okay, but a good pair of PPMs (peak programme meters) really helps highlight any level problems.

The PPM scale is easier to read than most other meter types. On a PPM, the needle pointing straight up (PPM 4) represents the reference level, about which I have already spoken. This is variously called line-up level, zero level, or minus 18 dBFS. PPM 6 represents the peak level (minus 10 dBFS), so no sound in your cut should go over PPM 6.

Some really good meters can be switched to monitor the sum (left plus right) and difference signals (left minus right) of stereo sound.

Needle colours on such meters are: red for left (L), green for right (R) (the same as port and starboard), white for the sum (M), yellow for the difference (S).

Monitoring the sum and difference signals can also alert you to any part of your mix, which, if it is out of phase, would produce abnormally high S (or difference) signal levels.

It is very unusual to get ‘out of phase’ material any more, except maybe from microphones, but it is worth you being aware that this can still happen.

You’ll generally find that studio-based material is much more accurately controlled, as far as level is concerned, than location material. Either way, a burst of tone placed before the programme material is so helpful when setting correct levels within your edit.

Most editing software can generate reference tone at whatever level and frequency you like. Normally, if this is set at 1 kHz and at a level of minus 18 dBFS (PPM 4), you won’t go far wrong.

 See No Evil—Displaying Sound on Visual Displays

Stereo sound can also be displayed on a screen, whether that’s an oscilloscope or computer monitor, as a two-dimensional waveform. This is done electronically by connecting the left signal to the horizontal direction on the display and the right signal to the vertical direction.

It follows that two signals of equal amplitude would generate a display that flickers on a line at 45° to the horizontal. This can be considered as a mono direction, or axis, and therefore equivalent to a sound that appears to come from the middle of the space between your two loudspeakers. Variations of amplitude and phase between the channels would take the trace (and your loudspeaker image) away from that one-dimensional mono centre line to the stereo two-dimensional world.

This sort of display can offer you extra information about a stereo image that a pair of meters would not be capable of showing. Also, it can look very attractive, as the illustration demonstrates. Note that the top left plot represents a sound which is placed at the centre of the image and is offering roughly equal amplitude levels to both the left and the right outputs.

As displays and software sound processing techniques get more sophisticated, the range of what can be displayed also increases, but the principle remains the same.

9.4 Split Tracks

Location sound can be recorded with a combination of different microphones. Personal radio mics and a boom mic is a common example of a ‘belt and braces’ approach often used by sound recordists on location. These different versions of the sound will come to you either on different tracks on a tape (A1, A2, A3, A4), or incorporated with the vision file, or indeed as individual WAV files which you’ll have to synchronise with the vision file yourself.

Modern WAV files (or BWAVs) have timecode information encoded with the audio data, so this synchronising operation can be performed with ease. Once married up to the vision, you can choose the best and most appropriate sound from the selection offered to you by the sound recordist.

Variations in editing software offer different solutions to this selection problem. To save you time, and especially if your programme will be delivered to a dub when you have finished, you can retain the split nature of the various recordings in your timeline and simply mute selected tracks with any unwanted alternative sound. Any panned nature of the original material can also be ignored by putting those selected tracks in your timeline to ‘mono’, which means that audio on these tracks will come out equally loud from your two loudspeaker monitors.

Again, this is not as complicated as it sounds, I promise.

9.5 Wild Tracks

Wild tracks (sometimes referred to as buzz tracks) divide into two main types—background wild tracks and specific wild tracks.

 Noises Off—Background Wild Tracks

Background wild tracks are primarily recorded to help you get over the problems stated earlier about the variations in the quality of background sounds, which are caused by different microphone positions necessary during shooting. Trees in a breeze, a nearby road, or a children’s playground can all sound different depending where the microphone, which is primarily picking up the actor’s voice, is pointing. To get over this problem, a sound recordist on location will, with the crew and actors silent, record chunks of this background noise (or atmos) before the camera is moved to the next set-up. As it is impossible to remove atmos from speech once recorded, all you can do is bring the level of the quietest atmos up to match that of the loudest by adding some of the appropriate wild track to the mix at various points.

If your scene takes place in a noisy pub, the director will often shoot the scene in dead quiet, with all the extras miming the chatter, and once the dialogue is shot, get the sound recordist to record a wild track of that noisy pub, now clean of scripted dialogue, for later use. Keeping the dialogue separate in this way, will save you having to edit the noisy pub at the same time as editing your often less noisy actors. This technique only works if the actors are constantly reminded they are talking over the absent noise of the rest of the pub and that their speech and body language must reflect this.

 Making Tracks—Specific Wild Tracks

Examples of specific wild tracks are sounds like a car driving off, a baby crying, or a telephone ringing.

If recorded ‘clean’ on location, these wild tracks can often save a huge amount of time in an edit session or dub, as the sounds recorded in this way are already in the right acoustic and will require very little fiddling around to make them sound perfect in the final mix. This is so much easier than finding an alternative sound effect from an effects library.

9.6 Sound Effects

The term sound effects does not only apply to specific additional sounds, but also, and rather confusingly, to the alteration of existing sounds with processes like equalisation or compression.

I will talk about these sound effects later on, but for now let’s concentrate on sound effects. I hope that’s clear!

Just like wild tracks, recorded sound effects can also be divided into two main categories: background effects and spot effects. Thus, in recreating the sound of a stormy night, the rain and wind would be background effects, and the thunder would be a spot effect.

 Stormy Weather—What Are Background Effects?

Background effects are sounds that don’t have to synchronize with the picture but indicate the setting to the audience, such as a forest atmosphere, restaurant chatter, or traffic.

 For Whom the Bell Tolls—What Are Spot Effects?

Spot effects are sounds such as door slams, telephones ringing, and glass breaking. Some spot effects have to be specifically recorded, whereas others are manufactured. There is nothing to stop you from layering several sound effects together in order to create a new and more interesting sound out of two or three old ones.

 On the Shelf—Use of Effect Libraries

Large libraries of commercial sound effects are readily available on the Internet nowadays (in WAV or MP3 form), but on some projects, as we have seen, sound effects may be custom-recorded for the purpose.

In the early days of film and radio, library effects were held on analogue discs, sometimes even 78 rpm discs, and an expert technician could play several effects on separate turntables into a live broadcast in a matter of seconds. Today, with effects held in a digital format, it is easy to sequence as many effects as you like into a timeline, and therein lies the problem: it’s far too easy to overdo things.

 In the Night Garden—Adding Sound Effects

I know the final soundtrack of most programmes is finished in a dub, but it can be the case that the soundtrack you produce in the edit suite goes on the air. A ‘good ear’ is required to produce a professional result, but no more so than when it comes to adding sound effects.

There is a huge danger here that you’ll be tempted to plaster on added effects that are too loud; after all, you want the audience to hear the ticking clock effect you’ve just added, don’t you? NO YOU DON’T! Remember, be subtle. No viewer should be made aware of the effects themselves, just what they do to enhance the realism of the drama.

In my opinion, many modern films are guilty of making the effects and music track too loud, and if you add this to the ‘realism’ of allowing your actors to mumble, the dialogue can become somewhat unintelligible.

There has recently been a flurry of viewer complaints about unintelligible dialogue in recent TV productions in the UK; Quirke (BBC) (2014), Jamaica Inn (BBC) (2014), and Broadchurch (ITV) (2015) have all come under fire recently for the poor quality of some of the dialogue.

 I’m All Right Jack (1959)—Funny Sound Effects

One of my favourite scenes in which sound effects were used to their comic best is in the film I’m All Right Jack (1959), directed and produced by John and Roy Boulting and edited by Anthony Harvey. The scene is set in a sweet factory.

Our hero, Stanley Windrush (Ian Carmichael), is being shown around the sweet-making machinery on the shop floor by the supervisor, played by Ronnie Stevens. Both the actors have to shout above the din of the machinery, which is whirring, bleeping, burping, wheezing, farting, and belching as the machines turn out vats of liquid cake mixtures and half-finished sweets. Windrush is, of course, asked to sample all the confectionary delights, with inevitable consequences.

It is a triumph for the sound department, where Chris Greenham is credited as the sound editor. Incidentally, the design department didn’t do so badly either, with all the machines having face-like features with loads of flashing lights and moving parts. A siren sounds, a bell rings, the eyes light up, and the mouths vomit the next load of a sugary, gooey liquid into a duct for the next process.

Incidentally, the film contains one of the great lines of the movies, when Fred Kite, a left-wing union leader played by the great Peter Sellers, is asked whether he’s ever been to Russia. Kite replies that he would like to visit one day, saying, ‘Ah, Russia. All them cornfields, and ballet in the evening.’

 Silent Movie—Essential Sound Effects

There are certain sound effects that are used time and time again, such as mobile phones, cars starting, tick over, traffic and so on. The list can become very long, but with the advent of USB memory sticks, or even phone memories, it has become possible for an editor to carry these effects with them. Personally, I’m so pleased not to have to carry that heavy box of 78s around with me anymore. I would strongly advise you to build up a library of your own, which would contain your most commonly used effects. Very often, the wild tracks recorded on location and in the studio from your current production can be saved and used more than once to help other productions whose sound recordists were not so thorough in covering every option. It’s so easy to keep a couple of phone rings to hand that, while they might not end up in the transmitted programme, they will allow you to time a cut properly.

 Laughter in Paradise—Save Your Laughs

In my line of work, it is essential to keep a store of laughs to act as sticking plaster over troublesome joins. While we all try to be honest, it sometimes becomes necessary to repair laughs that become unusable, for whatever reason, with laughs from another source. To clear my conscience, I always try to use laughs recorded in the same studio and, more importantly, from the same show. Replacement laughs have to be as nonspecific as possible, especially if they have to be used more than once in a programme. By nonspecific, I mean they must not contain any individual characteristic, such as a cough or a cackle, just a good, well-rounded laugh with a clean tail.

As a series progresses, I tend to save laughs that fit the above requirement and build up a library of suitable and varying laughs.

If you find some really good laughs, you might be tempted to store them on a longer term basis, to help another series along.

When you are adding laughs, the golden rule is less is more, so don’t become what I call ‘laugh-happy’. If 300 people in the audience on the night have decided the line isn’t funny, then there is a good chance the line really isn’t funny. Over-egging these indifferent lines leaves you with nowhere to go when it comes to the genuinely funny ones. You’ll never hear of a viewer complaint along the lines that they wished that the studio audience would laugh more; it’s usually quite the opposite.

9.7 Sound Mixing

When you are mixing sound you can only go by what you hear; that’s why good monitoring is essential. Just as you try to make shot changes not look awkward and clumsy, you must try to make the addition of extra sounds sound as natural as possible in order that no undue attention is brought to them individually.

 The Whisperers—Make It Believable

An added phone ring must appear to come from the phone as used by your actor in the scene. If this has been sourced from a wild track recorded at the time of filming, it is much easier to balance this and make it feel totally natural, but if the effect comes from a commercially available recording, then some adjustment will have to be made to the ambience of that sound other than just getting the level right. The reason for this is that most commercially available effects are recorded at very close range and, as a result, are as dry as dust acoustically. To help with this problem, some form of sound processing is employed, such as equalisation or reverberation, to make these added effects sound believable.

The main thing is get the level right, and then turn it down a notch from there. Modern software offers what is called rubber-banding, where levels of individual tracks can be raised and lowered with the use of markers placed on that particular track in the timeline. Sometimes these levels can be set through a fader box (what luxury!) and altered in a ‘fader live’ mode…serious stuff.

 Mixed Blessings—Adding Sound Effects

Here’s an exercise involving mixing sound effects together; it’s called Ex 46 Mixing Sound Fx.

You have several effects to play with here: a clock ticking, Philip making tea in the kitchen, a plane flying overhead, the projector noise, and the bleep of the VHS machine receiving the play command from a remote control. In addition, you have ‘clean’ footsteps from Philip, so that you can replace the footsteps on the actual shot, which are covered in dialogue, with the clean ones coming from another part of the scene.

It’s quite an assembly job to start with, even before you add any of the effects. Remember, when Mark runs the film, we need to see his face come alive with wonder as he sees the flickering images of his childhood, but at the same time, give us a chance to see dad’s film as well. You need to balance the POV shots of the film with Mark’s face, certainly to start with, then the film can run in vision a bit more after that.

As to the effects, the results must be a subtle balance of realism and perceptibility. To best judge the levels of the individual sound effects you have just added, try listening to your mix at a wide range of volume levels and alter individual elements if necessary. Also, listen to the mix on headphones to give you an idea what your mix sounds like on different kinds of equipment.

A brief note here to those of you not working on Avid: because you had to import my timeline via an EDL, you will not get any of my track level adjustments (rubber-banding, in other words). However, you can still hear my final mix in the MOV file of my cut (in this case Ex 46-CUT.mov), which I have provided for you.

9.8 Sub-Frame Editing

Sound edits in most video editing applications can only be achieved every 25th of a second, or 29.97th of a second with 60 Hz based material. Occasionally, there is a need to edit without this restriction, for example in the case of click removal. A click can be microscopic compared with the limit of 25th of a second (that’s 40 ms), lasting as it does only a few audio samples. The removal of these individual samples is called sub-frame editing.

If your editing software does not include a sub-frame facility, a section of your ‘clicky’ waveform will have to be exported as a WAV file and imported into specialist sound editing programs like Sony’s Soundforge or Audacity.

Here the waveform can be stretched out into individual peaks and troughs, and the click, when located, can be cut out and the waveform rejoined. In this way, because the WAV file is only shortened by a fraction (microseconds perhaps), no further action is necessary when this WAV file is inserted back into the edit, and it drops in perfectly. It’s as though nothing has happened, but at least that annoying click has gone.

Other forms of audio degradation, such as peak distortion, can be treated in the same way, and individual peaks that have been ‘squared-off’ can be surgically removed, leaving an unclipped but slightly shortened waveform. This can take a long time, but the results can be impressive.

9.9 Sound Processing

‘Free’ with much modern editing software are sound processing tools, often known as digital signal processing (DSP), that will help you tailor any aspect of the sound to meet the needs of your project. Yes, I know there might be a subsequent dub (am I repeating myself?), but some processing effects may have to be added in the edit suite. Another good reason for doing some of this work in the edit suite is that it is difficult to judge a cut, however well-crafted visually, with a soundtrack lacking in appropriate effects or full of inappropriate ones.

The range of processing effects and how they are applied are totally dependent on the software in use. Here are the most common, and therefore the most essential and powerful, sound processing effects.

 Acoustic Café—A Look at Equalisation

Equalisation (EQ) is audio ‘posh-speak’ for bass and treble adjustment where different frequency bands are attenuated or boosted to produce some desired spectral characteristics.

Moderate use of equalisation can be used to fine-tune the tone quality of a recording. Extreme use of equalisation, such as heavily cutting a certain frequency or band of frequencies, can create more unusual effects (see parametric equalisation, coming up next).

A common use of equalisation is to transform a commercial track so that it might appear to be coming from a radio on the set, or to ‘distance’ wild tracks so that the tick-over of a waiting taxi doesn’t appear to be coming from inside the room.

Both of these examples can be achieved, to a large extent, by attenuating the high and lows of the audio spectrum and adding a bit of reverb to create distance.

 Pick of the Week—Filtering Sounds

Parametric equalisation is a more flexible form of equalisation, or in other words, equalisation with more knobs. In this mode, you are able to set the range of frequencies you wish to boost or cut and also set the sharpness of this adjustment, which is referred to as the Q.

My advice is to play around when you have a spare moment and listen to what you can do with speech and music when extreme and selective boosting or cutting is used on different frequency bands. Remember to store some of your favourite settings that might be most appropriate for your kind of work.

Parametric equalisation can also be used to repair a recording that has suffered from an external source of interference. Mains hum (50 Hz or 60 Hz), caused by induction from nearby electricity cables, for example, can be tuned into with a high-Q (or very selective) filter that is initially set to boost and then, once you’re on top of the offending noise, switching that boost to cut, with the hope that the offending mains hum will disappear.

Air-conditioning machines produce a range of frequencies of interfering noise, but the most troublesome frequencies are related to the motor speed, and these can be dealt with individually and removed. Beware though not to do too much damage to the rest of the audio spectrum you want to preserve.

I’ve probably gone into too much detail, but Brownie points are available here as you offer a solution to a distraught director worried about inaudibility in a noisy location.

 Open Spaces—Use of Echo and Reverberation

Reverberation (or reverb for short) is where one or several delayed signals is added to the original signal to simulate the reverberation effect of a large hall or cavern. Just clap your hands in a cathedral and count the number of seconds it takes for the sound to disappear, and you’ll know what we are trying to achieve. You might get up to four or five seconds in such an environment; this is known as the reverberation time. That’s just one of the parameters available here in most DSP reverb applications. In addition, you are also able to adjust the quality (or EQ) of the reflected sound to allow for different surfaces in the room space whose sound characteristics you want to mimic. By adjusting this and other parameters, such as depth, attack, and release, you can create the ‘sound’ of many different environments.

 On the Ropes—Use of Audio Level Compression

Level compression is not to be confused with audio data rate compression. The audio compression here is primarily the reduction of the dynamic range of a sound or sounds, such that louder elements are reduced in level dynamically, to better match with lower level signals. Most sound reaching the cinema screen or TV is compressed in one way or another. This is done either as it is recorded or somewhere along the chain of processing from microphone to transmitter, which can include the edit suite.

The dynamic range of sounds in nature far exceeds a comfortable range we would expect to experience when we watch a film or TV programme at home or in the cinema. If the threshold of hearing anything at all is 0 dB SPL (sound pressure level) and the threshold of pain is 120 dB SPL, our ears are capable of hearing a dynamic range of 120 dB. Given that, first, listening environments are never totally quiet (typically they give values in the range 20 to 40 dB SPL), and second, we would not want to get near the pain threshold, any soundtrack content must be encouraged to fit into an acceptable range of levels from 40 to 90 dB SPL. This gives a dynamic range of only 50 dB, which is somewhat smaller than the range of sound levels that we can experience outside the cinema. It follows that some form of level compression is required in the chain from microphone to speaker; if this was not done, we would constantly be reaching for a volume control because it would be too loud or too soft all the time.

Most studio-based sources of sound are compressed as they are recorded, so no further action is required by the editor or the dubbing mixer unless a specific problem arises, such as with the addition of some newly recorded dialogue that hasn’t yet had this processing applied to it. Again, with most software, this is made easier with the ability to call on several presets within the compression parameters. As these are numerous and complex, it is good practice to remember and store the ones you like.

If your edit is going to a dub (oh no, not that again!), it is just as well to remove your processing effects prior to export, or at least let the dub have access to the original sound. The dub will almost certainly be able to call upon cleverer techniques which might well produce a better end result.

 Trouble in Store—Keep Storing What You Like

The advantage of modern DSP applications is that you can store any processing effect you like, name it, and keep it forever. In this way, when you’re asked to make the dialogue sound as though it was recorded in an empty 25,000 cubic metre oil tanker about 10 nautical miles off the coast of Newfoundland—again—you can look very smug and pull it immediately off your USB stick.

Seriously, it’s good practice to store certain favourite effect parameters, as they tend to come up time and time again.

9.10 More Specialised Sound Processing

 Mixing It—Flanging or Phasing Effects

Flanging (or phasing) is often used to create unusual sounds. Here, a delayed signal is added to a proportion of the original with a continuously variable delay, usually shorter than 10 ms. This effect was originally created by playing the same recording on two almost-synchronized tape machines and then mixing the signals together. Filtering the delayed signal prior to mixing can produce a range of weird and wonderful effects. There are some Beatles and Beach Boys tracks of the 1960s that are good examples of this technique. Phasing can be used to give a synthesised or electronic effect to natural sounds, such as human speech.

Interestingly, the severe modulation of a sound wave that was necessary to produce the sound of a Dalek’s voice was done in a circuit called a ring modulator. Today, what any hardware used to do can be replicated with DSP technology.

 The Choir—Chorus Effects

Chorus is where a delayed signal is added to the original with a short and constant delay. If the delay is too short, it will interfere with the undelayed signal and create a phasing effect as above.

Chorusing can be used to good effect to make your 20 extras sound like 200 in a pub or club. However, a better way to achieve this might be to duplicate the wild track in the timeline (more than once if you wish), separate the copies in time by a few seconds, and apply different reverb effects to each layer, thus giving the impression of a much larger group of people.

 Tune-Up—Pitch Shifting Effects

Pitch shifting shifts a signal up or down in pitch—yes, really. Here the frequencies within the wave are multiplied by an adjustable amount so that their pitch goes up or down but they occupy the same time scale. One cruel sound supervisor (namely Laurie Taylor) at the BBC always used to do this to the producer’s audience warm-up speech whenever the sound department was mentioned. The producer’s voice kept swinging from a squeaky cartoon voice to a deep, deep bass. Well, it was very funny at the time, and the audience loved it!

One application of pitch shifting is pitch correction, the modern salvation of many indifferent singers. However, the result can sound strange, especially if you have to change the pitch too far.

 Time Bandits—Time Stretching Effects

Another DSP effect available with most editing software is time stretching. This is the opposite of pitch shift because here the duration of an audio signal is changed without affecting its pitch. This is often used to fit and fine tune new dialogue from an additional dialogue recording session (discussed next) to better match the lip-sync of existing pictures.

9.11 Additional Dialogue Recording

For whatever reason, sometimes your actors have to be asked back to attend an additional dialogue recording (ADR) session so that dialogue that was impossible to use from the original recordings is re-recorded or new lines are added.

 The Chat Show—Replacing Dialogue

ADR sessions are used for several reasons. They include:

• Changing or adding to the original scripted words.
• Altering the inflexion of a performance.
• Clarifying dialogue to make it more intelligible.
• Revoicing dialogue which was impossible to record at the time of filming because of, say, high background noise levels.
• Revoicing dialogue into a foreign language.

9.12 Foley Sound Effects

Just as dialogue is sometimes rerecorded in an ADR session, sound effects are also recreated in a specialised process called Foley work. Some movies are completely revoiced and Foleyed; none of the original sound is used at all, and it only serves as a guide for future rerecording.

In this way, all sounds, from a door squeak to the rustle of a newspaper, have to be recreated from scratch. This would be incredibly time consuming, if not impossible, to do from a sound effects library alone, so it’s simpler to use the expertise of Foley artists to record appropriate sound effects in direct synchronism with the actions on screen. Thus, footsteps, the movement of hand props such as a teacup, and the rustling of cloth are common Foley effects.

A skilful Foley artist is able to put back all the emotion contained within the seemingly simple action of an actor walking.

 The World Service—Preparing for International Sales

The advantage of the ADR and Foley approaches is that, in these days of international sales, the final soundtrack of a film can be distributed with the major elements of dialogue, music, and effects all completely separate. This will easily allow the dialogue to be revoiced into any language and then be remixed with the original and unchanged music and effects tracks. All can be processed separately to suit different outlets, such as Dolby 5.1 for cinema, HDTV, or Blu-ray release, and right back to ordinary two-channel stereo sound for analogue TV transmission.

9.13 Key Points—Sound Matters

• Get the room right first by treating it acoustically.
• Get decent loudspeakers and put them on stands.
• Sort your monitoring out with a good pair of PPMs.
• Make sure line-up levels match from all sources.
• Often, with multitrack location material, you have to choose the appropriate sound source for each of your characters’ speeches. Once selected, pan it to the middle so that it comes equally from both speakers.
• Background wild tracks are primarily recorded to help you get over the problems of variations in the quality of background sounds caused by different recording positions.
• Specific wild tracks are sounds like a car driving off, a baby crying, or a telephone ringing. If recorded clean on location, these wild tracks can often save a huge amount of time in an edit session or dub.
• Build a library of the most commonly used sound effects, such as mobile phones, cars starting, an engine ticking-over, traffic noise, and so on. Encode them as good quality MP3s, or WMAs, or better still, straight WAVs, and carry them with you on a USB stick, flash memory, or your phone.
• Sub-frame editing offers you the ability to do microsurgery on a sound waveform. Even though it’s not included with some editing platforms, it is available on more specialised sound editing programs.
• Digital sound processing (DSP) offers you a wide range of possibilities of tweaking sound.
• Equalisation (EQ) is audio posh-speak for bass and treble adjustment but is considerably more versatile.
• Filtering with a parametric equaliser is a more flexible form of EQ.
• Reverberation can help recreate the acoustic of any location, from a small room to a cathedral.
• Level compression (not to be confused with data rate compression) can make even a softly spoken commentary audible over an already full audio soundscape.
• Phasing/flanging can be used to give a synthesised or electronic effect to natural sounds, such as human speech.
• Chorus can be used to thicken sound.
• Pitch shift and time stretching can be used to repair or fit sounds to match the timing of their surroundings.
• Additional dialogue recording (ADR) allows dialogue, which was impossible to use from the original recordings, to be rerecorded.
• Foley sessions are used to recreate and add nondialogue sounds to a movie or TV programme.
• Some movies are completely revoiced and Foleyed with none of the original sound being used at all. This can make the creation of international versions a lot easier.