Audio in Film and Video: Post Production

Audio in Film and Video: Post Production
By Bernie Allen

Too often when people think of audio engineering or audio related work their thoughts go immediately to music, be this live sound engineering or studio production work. This, however, fails to recognize the importance of audio in most other forms of popular media. Try watching your favourite TV show on mute or plugging your ears next time you’re in a movie theatre and you will quickly understand the importance that audio plays in our ability to not only follow a storyline, but to enjoy the overall viewing experience.

In this article I will try to provide some insight into the world of audio post-production as it relates to film and video.

Sound Editorial Team:

A film’s sound editorial team is comprised of many different individuals all with important editorial duties. Depending on the budget of a production, teams of individuals may be responsible for everything ranging from dialogue editing, to mix engineering and everything in between. The editorial team is typically led by a supervising sound editor who has the unique role of interacting with the film’s director, producers and picture editor, as well as overseeing all sound preparation responsibilities. It is ultimately the supervising sound editor’s vision and tastes more so than any other sound editor that will determine the personality of the finished soundtrack (1).

The supervising sound editor directs the work of the sound editorial team in a way that allows for the final product to be brought together as a cohesive whole.

Dialogue Editing

Dialogue editing is the process of taking the recorded dialogue from a video shoot and making sure that it is as clear and audible as possible, at all times. This can mean removing set noises, disruptions, pops, clicks and anything that might interfere with the overall quality of the soundtrack, or anything that may prevent an audience from appreciating the dialogue at its very best. The result is hundreds of cuts and pieces arranged and blended in a way that is meant to be undetectable (2). Other responsibilities of the dialogue editor may include making quiet passages louder or loud passages quieter; anything that helps improve the overall balance of a soundtrack and make the film’s dialogue seem more believable.

Dialogue editing is largely an invisible art, which means the only time an audience will notice the editing work is if the work is done poorly. The dialogue track is the most critical audio component of a film (3) (4), and so it is important that the job is done correctly. If the work is done well, an audience should not notice the work at all.

Dialogue editors often work by isolating different speakers on a soundtrack and allocating them their own (audio) track so that their dialogue can be mixed and blended separately from other speakers on the soundtrack. Dialogue editors may also be given audio outtakes from alternate takes to work with (5). This is done so that if a line is compromised or slurred in the main soundtrack it can be replaced with a line from an outtake. In the event that the original soundtrack audio cannot be salvaged, it is then up to the dialogue editor to prepare the soundtrack for Automated Dialogue Replacement, otherwise known as “ADR.”

ADR

When there is a dialogue problem that cannot be easily repaired the next step in the post-production process is usually Automated Dialogue Replacement. Otherwise known as “ADR”, this process involves re-recording the actor’s dialogue in a controlled studio environment. This is can be necessary for a number of reasons ranging from technical problems, better reading of lines, or excessive noise on set while filming. Technical problems might include microphone related problems like distortion or pops, while “better reading” could mean the director prefer an actor’s lines be read differently in post-production. Noise problems might include something like wind in a convertible scene or voices on set that cannot be easily edited out. Whatever the reason, ADR is often a critical part of achieving a professional final product.

A typical ADR session involves an actor performing his or her lines in a recording booth while video footage of their original performance is played on a screen. It is very important that the actor’s lines are timed to match the filmed performance. In fact, emphasis is often placed on timing even over performance (6), as poor timing can drastically take away from the believability of the final product. However, there is now software that can automatically fit an actor’s performance to the required timeframe (e.g., Synchro Arts’s “VocAlign” or Audiosuit plugins). This can be a time saving investment well worth the money.

Automated Dialogue Replacement is usually less preferred to that of the original soundtrack audio recorded on site. This is because the cost of ADR can be great, as well, the realism in an ADR performance can be lacking. Because actors must re-record their dialogue away from set and in isolation from other actors, it can mean a different feeling from that originally captured on film. Also, in the studio environment actors are expected to perform in exact time with the performance on film, which can diminish from the believability of the performance. However, ADR is often critical to achieving the best final result for the finished production.

Sound Design and Foley:

Sound design is the practice of creating and applying sound effects to a film in order to enhance the storyline and give greater depth to what is viewed on screen. Sound Designers (or sound effects editors) do this by creating a film’s soundscape (background ambience) such as traffic, birds chirping or children playing, as well as developing sound effects for everything ranging from gun shots and robots, to spaceships and animals. Most people do not realize that when a film goes to post production most of the sounds, other than dialogue, need to be recreated; so that a sound designer or sound effects editor becomes responsible for all the audio elements in a film or TV production other than dialogue.

While the roots of sound design go back as far the early 1900s, the term “Sound Design” did not become popular until the 1970s (7) when it was used to describe the work of guys like Walter Murch, Randy Thom and Ben Burtt. Ben Burt, in particular, is the man responsible for the sound effects heard in pioneering movies like Star Wars, Indiana Jones and a long list of others. Ben would spend months and months preparing special sounds for the movies he worked on, something which ultimately lead to receiving an academy award for his work. This is a matter of significance because until that time academy awards were only given to production sound engineers and re-recording mixers. Ben also received an Oscar and a Special Achievement award for his work – on Stars Wars, specifically – which helped to bring the term “sound design” into the popular lexicon.

Rerecording Mixer and the Final Mix:

Re-recording mixers are the people responsible for treating all the different audio elements of a film and bringing them together to create a finished soundtrack. This work means ensuring that the different pieces fit accurately together in a way that is refined, dynamic and believable.

Referenced Material:

1) Yewdall D.L. – The Practical Art of Motion Picture Sound 4th Edition (2011) p. 293
2) Yewdall D.L. – The Practical Art of Motion Picture Sound 4th Edition (2011) p. 363
3) Purcell J. – Dialogue Editing for Motion Pictures: A guide to the Invisible Art (2007) page. ix
4) Yewdall D.L. – The Practical Art of Motion Picture Sound 4th Edition (2011) p. 363
5) Holman T. – Sound for Film and Television 3rd Edition (2010) p. 159
6) Evans R. – Practical DV Filmmaking 2nd Edition (2005) p. 261
7) Holman T. – Sound for Film and Television 3rd Edition (2010) p. 146

All About Microphones

All about Microphones
By Bernie Allen

While there are many makes and models of microphones they all serve the same general purpose. That is, to convert acoustic energy or vibrations into electrical energy. This energy is then amplified and/or recorded (and perhaps mixed and mastered, down the road).

Microphones work through the use of three main mechanical systems: 1) A Diaphragm 2) A Transducer and 3) A Casing. … Air pressure, the result of acoustic energy, vibrates the diaphragm of the microphone which is connected to a transducer, which then creates a small electrical current. This current is in essence the electrical equivalent of the converted acoustic energy.

The three most widely used microphone types are: Dynamic (or Moving Coil) microphones, Condenser (Capacitor) microphones and Ribbon microphones. Of these three types, Dynamic and Condenser microphones are by far the most widely used… and of these two, Dynamic microphones are the most used.

Dynamic Microphones:

Dynamic microphones (also called “Moving Coil” microphones) are the most durable, least likely to distort, relatively least expensive and most widely used of the three main microphone types. In fact, even the cheap computer microphones we use are an example of a dynamic microphone. Small diaphragm dynamic microphones are also the most popular microphones for live recording and live performances; so that at live shows these mics are the ones you are most likely to see used.

Dynamic mics not only offer a good reproduction of sound, but they do not need a voltage source (like condenser microphones), which makes them good in most practical situations. Another reason these microphones are so popular is because of their resonant frequency (exaggerated frequency range), which sits between 1 and 4 kHz (1) – Right in about the same frequency range as the human voice. Unfortunately dynamic microphones have poor frequency responses at about 10 kHz, which means when recording something like vocals you will not get the kind of full, rich sound that you would with a microphone that better accommodates higher frequencies, like a condenser microphone. Dynamic mics also add tonal coloration starting between 5-10kHz (2), and tend to sound sparse when placed more than a foot from a sound source. As a result they generally work best in close-mic situations.

While there are many examples of good small diaphragm dynamic microphones (not to be confused with large diaphragms… which are used almost exclusively when miking loud bass-heavy instruments), two especially popular makes are the Shure SM57 and SM58. Other good dynamic microphones include makes from Sony, Sennheiser and Audio-Technica.

Condenser Microphones:

Condenser microphones (also called “Capacitor Microphones”) work distinctly differently from a dynamic microphone. These microphones use electrically charged plates, one that is fixed, and one that can move which acts as a diaphragm (3) (4); unlike dynamic microphones which utilize a moving coil attached to a diaphragm in a magnetic field. While more delicate than ultra-durable dynamic microphones, condenser mics have a flatter frequency response which produces a more accurate sound reproduction, with less tonal coloration, particularly in the mid to high ranges, exactly where precision is needed when recording vocals. The down side to these microphones, however, is that they produce less voltage (converted acoustic energy in this case) than other microphone types, and as a result require additional power to boost signal levels…. Enter: “Phantom power.”

Phantom power is a 48-volt direct current (DC) power source delivered by a recording console or microphone preamp through the same cable that transfers the audio. Most consoles permit phantom power to be turned on and off because of small problems it may cause with cables connected to dynamic microphones (e.g. loud pops). Phantom power can also come from an external power source or a battery located inside the microphone. If the power source is a battery inside the mic, it is important to make sure it is kept fresh; otherwise the microphone may not provide the best results.

Condenser microphones are usually the best option when recording vocals, acoustic instruments (e.g. guitar and piano), woodwinds and strings. Omni directional (the natural state of) condenser microphones also capture a wider range of frequencies from greater distances than do other microphones, which makes them good for overhead miking or capturing acoustic room ambience. Though condenser mics tend to have flatter frequency responses than other microphones, they are not without their resonant frequency range, which is usually between 8-12 kHz.

Ribbon Microphones

Ribbon microphones work much like dynamic microphones, with one major difference; the diaphragm in these microphones also doubles as the transducer, which makes for lighter moving parts, and in turn a higher frequency response (5). Unfortunately the electrical signal generated in these mics is small, and an output transformer is necessary to boost the mic signal to an appropriate level. Another major difference between ribbon and dynamics microphones is the ultra-fragile nature of ribbon microphones, which are even more fragile than condenser microphones. This fact can make them poor choices when doing live sound.

Like dynamic microphones ribbon microphones tend also to sound sparse if placed too far from a sound source. They are, however, preferable to dynamic microphones in close-miking situations, where they create a warmer more even sound (6). This is because their frequency response is flatter and tends to reach about 14 kHz as opposed to the dynamic’s 10 kHz.

Pickup Patterns

Pickup patterns refer to a microphone’s ability to pick up sound from different directions. For example, some microphones pick up sounds from all directions, while others only pick up sound from a single direction. Pickup patterns can be useful depending on the kind of job you are doing. For instance, when recording atmospheric sounds, live background vocals or room ambience it might be preferable to use a microphone that picks up sound from all directions. However, when recording a vocalist you might want to use a microphone that only picks up sound from one direction, so as not to include unwanted noise. The same is especially true when miking different components of a drum kit. Noise bleed can be a big problem when miking a drum kit so the better one can block out unwanted noise, the better the end results.

Common pickup patterns are:

Omni-Directional: Able to pick up sound from all sides of the microphone

Cardioid (or unidirectional): Picks up sound from directly in front of the microphone, while sounds to either side are barely heard.

Figure 8 Pattern (or bidirectional): Figure 8 microphones have a pick up pattern shaped like a figure 8. These mics pick up sound in the same way from both the front and rear of the microphone, while neglecting sounds from either side.

References:
1) Owsinski B., The Recordings Engineer’s Handbook (2005) p.1
2) Gibson B., Sound Advice on Microphone Techniques (2002) p.9
3) Owsinski B., The Recordings Engineer’s Handbook (2005) p.5
4) Milstead B., Home Recording Power (2001)p.46
5) Owsinski B., The Recordings Engineer’s Handbook (2005) p.4
6) Gibson B., Sound Advice on Microphone Techniques (2002) p.11

Understanding Common Audio Meters

Understanding Common Audio Meters
By Bernie Allen

Common sayings in audio related work are: “Let your ear be the final Judge” or “Listen, don’t look.” While very true, careful use of meters can be an important part of managing noise and distortion, and ensuring that one’s final mix or master is ready for the world. Meters help to provide greater precision and act as cautionary tools in the audio environment.

While I will not cover all the different meters in this article (that is, specialty meters and such), I will cover the more common meters one is likely to encounter, and hopefully provide a better understanding of these practical tools.

Peak Meter:

Peak Meters are used to display immediate (instantaneous) audio signal levels and warn of levels that may exceed a predetermined limit — 0db on a digital system, for example. A signal that exceeds a predetermined level may result in what is called, “clipping,” or distortion. If an audio file is too badly clipped there is virtually nothing that can be done to restore it! Though, audio restoration software does exist which attempts to repair things like clipping, they are usually not without their compromises.

While peak meters can be good time-accurate indicators of audio signal levels, they are not always good indicators of loudness. For example, two platforms both reaching 0dB on a peak meter can have as much as a 10dB + difference in loudness (1). For this reason it is always important to let your ear be the final critic or use averaging meters (such as RMS or VU) in conjunction with a peak meter.

Digital Peak Meters (ones which most readers will be more familiar) make use of three main functions 1) Clip Indicator 2) Peak Hold and 3) Peak Level:

Clip Indicator – A Clip Indicator (also called an Over Indicator) lights (usually red) when the audio signal exceeds the clipping point. With digital audio this usually means that 3 or more consecutive 0dB overs are present in the audio signal. With analog audio the over level may actually be set above 0dB.

Peak Hold – This is a line that is held for some time on the meter, indicating the highest peak reading. This feature lets the user know how far below 0dB the highest peak of the signal reaches. The duration of the Peak Hold can be set to off, held for a few seconds, or even held forever.

Peak Level – Peak Level is simply a numerical display of the highest peak… a number reading.

VU Meter:

VU (Volume Unit) meters are voltmeters with a famously slow response time (2) (3). Of questionable value in the digital domain (4) VU meters are considered to be better indicators of loudness than are Peak Meters – This is because the prolonged time they take to rise and fall better corresponds to our natural perspective of loudness (5). For this same reason the VU is also considered to be a poor representation of peak signal level. The immediate (instantaneous) changes of Peak Meters better represent signal level.

VU meters typically run on a scale between -20dB and 3dB, with 0dB being the common ‘reference level’ (6) (7). Engineers dealing with analog mediums generally try to keep the VU meter as close to 0dB as possible, as this tends to maintain good headroom which will keep the tape from distorting. In the digital domain, however, far more stringent measures must be applied to ensure that at no time do signal levels ever exceed 0dB. There is no headroom beyond 0dB in the digital domain, and so a signal level that exceeds this threshold is automatically clipped.

RMS Meter:

RMS (Root Mean Square) meters are used primarily in the digital domain to measure the loudness of an audio signal. While good, as they display the amplitudes that occur in a complete cycle of a frequency (8) – or average levels as opposed to just peak levels – they tend also to be sensitive to lower frequencies. That is, lower frequencies can give the illusion of greater loudness or power when reading these meters.

Phase Correlation Meter:

The Phase meter is used to compare the degree of similarity between the left and right channels of a mix. This meter is usually scaled between -1 and +1, with +1 signifying that both channels are in perfect phase and -1 indicating that the two channels are phase–inverted. A positive reading is usually the desired reading, suggesting that one’s mix is good. Though, there are some instances where other readings may be preferable, namely where phase inverting effects are used.

References:

1) Katz B., The Secret of the Mastering Engineer (1999) p.9
2) Watkinson J., The Art of Digital Audio 3rd Edition (2001) p.70
3) Hoeg W. and Lauterbach T., Digital Audio Broadcasting (2003) p.122
4) Self D., Audio Engineering, Explained (2010) p.126
5) Katz B., The Secret of the Mastering Engineer (1999) p.9
6) Izhaki R., Mixing Audio – Concepts, Practices and Tools (2008) p.95
7) Self D., Audio Engineering, Explained (2010) p.126
8 ) Sonar X1 (Producer) Help Index (2011)

Digital vs. Analog Audio

Digital vs. Analog Audio
By Bernie Allen

Digital Audio began for most people with the introduction of the CD (Compact Disc) in the 1980s. At this time it was expected that the digital contents of a CD would be immediately converted to analog upon use in a CD player – which relied on analog components in order to work. Today digital technology allows audio engineers, producers, sound designers and musicians the freedom to do things that were once difficult or impossible with analog technology, at a price that is within the reach of most people. Equipment that once easily took up an entire room worth of space can now also be neatly saved on a computer hard drive and conveniently accessed and utilized through the use of a digital-audio workstation or audio-editor, often with the same or better results.

While current analog equipment often looks more advanced than its predecessors, the underlying technology remains, for the most part, unchanged. This is because most innovations in analog equipment has been made, thus further advances are slower in the making (1).

Aside from being a more effectual way to work, digital technology has also brought the field of audio from being an esoteric domain isolated from other fields, to one that now shares much in common with information technology (IT) and even computer science (2). This is because once in the digital domain audio becomes little more than data, which means approaches and methods that were originally developed for other areas and/or purposes can now be easily applied to audio, video and other forms of media. In other words, the sky is the limit for digital audio, as it continues to evolve along with other digital technology.

Cost Effectiveness:

Both high quality modern analog and digital equipment provide good quality professional results, and both are able to do most jobs (3). Still, some make the argument that digital equipment may actually be better for a number of reasons relating to audio quality and greater capabilities (4) (5). Where the two mediums tend to differ markedly, however, is in the area of cost (especially in recent years). With advances in digital technology it is now possible to achieve a level of audio quality and professionalism that only 10-20 years ago was impossible even with the most expensive analog equipment (we’re talking hundreds of thousands of dollars!) (6).

Upgrading one’s equipment in the past could mean buying entirely new systems that might run a person thousands and thousands of dollars. Software, because it doesn’t require special mechanical bits (like nuts and bolts) or production line manufacturing, is almost always cheaper; and because the mechanisms that make software equipment work, exists virtually; there is less chance that your gear will need to be replaced prematurely. Another thing to note is that most software companies that specialize in digital-audio software like ‘Sonar’, ‘Logic’, or ‘Pro Tools’ (for example) attempt to hold on to their customer base by offering deals or discounts to users who own earlier versions of their products when newer updated versions of their product is released.

Convenience:

Today most simple processes relating to audio equipment are carried out in the digital domain. This provides many conveniences in the way of flexible configurations vs. hardwired (analog) configurations. In the case of digital mixers this means that any input can go to any or all channels or that people can have greater capabilities, like unlimited channels or tracks, without having to find the space to house cumbersome analog hardware that can weigh as much as thousands of pounds.

Another major bonus about going digital is that your audio remains in the digital domain at all times, meaning one is not likely to experience audio degradation as a result transferring audio data from point a to b, or bouncing down tracks etc. There is then also the benefit of “Non-Linear Editing” or “random access” (vs. serial access), which means one can move instantaneously to any point of a recording for editing or playback, or have the ability to cut and paste audio where needed without compromising the original audio (i.e. slicing treated tape) or audio quality.

Software vs. Hardware:

As hardware has both digital and analog alternatives (7), I think it is important to unambiguously explore the differences between digital ‘software’ vs. digital ‘hardware’. Much the same can be said of this debate as has been said of that relating to analog vs. digital, in general. That is, where hardware is concerned manufacturing costs are always passed on to the consumer; additionally, the matter of space is an important one. Hardware equipment means a ‘physical product’, one that needs to be maintained and up kept, space allocated to, and often times hardware needs to be lugged around or taken places. Software doesn’t share such problems… because software doesn’t need to be physically manufactured in the same way, the cost is much less, and because software exists virtually it does not take up physical space nor is it a problem to take places (a laptop or USB stick will usually do), making it a more flexible and accessible alternative to either analog or digital hardware.

With that said, there are still many who prefer to use hardware over software, as it allows for a more hands on workflow. Indeed, there are still many things you will not find a software alternative around, or areas where software hardly lives up to its hardware equivalent (like musical instruments). However, for most things, especially audio effects, software is often as good or better (8). With that said, I think it is safe to say that hardware, both analog and digital, will remain around for some time to come.

References:
1. Watkinson J., An Introduction to Digital Audio (1994). P.1
2. Self et al, Audio Engineering – Know it All (2009) p.410
3. Watkinson J., An Introduction to Digital Audio (1994) p.1-2
4. Berman R., Basics of Mixing (1999) p.6
5. Watkinson J., An Introduction to Digital Audio (1994) p.5
6. Milstead B., Home Recording Power (2001) p.26
7. Roland Corp., An Introduction to Digital Mixing
8. Self D., Audio Engineering Explained – For Professional Audio Recording (2010) p.429