Module 42-B

Digital Audio 

The decade of the '80s saw the introduction of digital audio signal processing. This not only opened the door to a vast array of new audio techniques, but it represented a quantum leap in audio quality.

For example, the following technical problems have been a headache for audio recording engineers for decades: (Don't worry if you don't understand what these things mean — or, then again, you could memorize them and impress your friends by tossing these terms around in a conversation!)

All of these problems and even a few more are eliminated with digital audio.

This is possible because of the precise timing pulses associated with digital audio and the fact that digital signal is only comprised of  "0s" and "1s." These represent simple positive and negative voltages that are not close to each other in value (so as to get easily confused or muddled along the line).

As long as equipment can reproduce just these two states, there is an audio signal.

However, with an analog signal there are an unlimited number of associated values, providing ample opportunity for things to get out of whack.

Technically speaking, the background noise of a digital signal can be as bad as 20dB (which is a lot) and the digital signal will still survive. In the case of an analog signal, this would translate into intolerable noise.
 

Copying vs. Cloning

Each time you make a copy of an analog audio segment you introduce aberrations because you are only creating a "likeness" of the original. With digital technology you are using the original elements to create a "clone."

If we are using the original uncompressed digital data, we can fully expect to end up with an exact clone of the original, even after 50 generations (50 copies of copies).

With analog data copies of copies quickly result in poor audio quality. Before the event of digital technology, such things as nonlinear editing (which we'll talk about in Module 56) were not possible.

If you have the option, you'll want to convert analog data into digital as soon as possible and leave it that way until you are forced at some point to convert it back to analog.
 

Converting Analog to Digital

The same sampling and quantizing principles that we discussed in digital video apply to digital audio. With both audio and video the analog signal is typically quantified or sampled 48,000 times per second.

That means that every 20 microseconds a "snapshot" is taken of the analog voltages. This instantaneous snapshot is then converted first to a base-ten number and from there to a computer-type binary ("0" and "1") form.

The number of data bits used to encode the analog data determines the resolution and dynamic range possible.

A 16-bit encoding system has 65,536 voltage steps that can be encoded. Obviously, the higher the data bits the better the quality — and the more technical resources required to handle the signal.

Such high rates demand a high degree of timing (synchronization) precision. Without it things fall apart with stunning speed.

Just as in video, a synchronizing signal is used to keep things in lock step. This signal or synchronizing (sync) pulse in digital audio is typically sent out every 0.00002 of a second.
 

Quantizing Error

In audio production signals must be converted back and forth from analog to digital and from digital to analog. Since we are dealing with "apple and orange" types of data, something called a quantizing error can result.

In the analog-to-digital conversion process, a voltage midpoint is selected in the analog values to use as the digital equivalent. This midpoint is a close, but generally not a perfect, reflection of the original analog signal. Thus, the error, and the need to minimize the number of digital-to-analog (as well as analog-to-digital) conversions.
 

Optimum Digital and

Analog Audio Levels

The optimum audio levels for digital audio signals are different than those for analog signals.

Whereas the 0dB peak setting is the standard operating level (SOL) for analog systems, for digital equipment the maximum level (in North America) is typically -20dB. 

With both analog and digital signals it comes down to something called headroom.

Headroom is the safe area beyond the SOL (standard operating level) point. With a SOL of -20dB, (which is typically the standard in North America) this leaves 20dB for headroom. European countries tend to allow for-18dB of headroom. 

Okay, this is a bit technical, but just keep in mind that the maximum audio level for analog signals will generally be different than it will for digital signals.

Analog audio systems often use an analog meter, such as the one shown on the left.

With digital signals, however, a digital meter, such as the one shown on the right, or a PPM meter (to be discussed below), is used. In the case of the digital meter on the right, when the signal touches the red area, we're entered the headroom area.

If a digital signal were to go to the very top of the scale, clipping would occur. Unlike analog audio, where exceeding the maximum level will result in signal distortion, in digital audio you might not notice the elimination of audio peaks .

Actually, an occasional full-scale digital sample (to the top of the red range above) is considered inevitable; but, a regular string of "top of the scale" occurrences means that the digital audio levels are too high and you are losing audio information.

VU meters respond in different ways to audio peaks.

In the case of a ▲ standard VU meter the needle tends to swing past peaks because of inertia. At the same time, this needle will not quickly respond to short bursts of audio. Thus, this type of meter tends to average out audio levels.

Because of the limited headroom with digital audio signals a faster responding peak program meter (PPM) or the previously discussed digital meter is preferred. On the outside a PPM looks like the animated VU meter.

Before you can really get serious about maintaining correct audio levels throughout a production facility, you must see that the audio meters throughout the facility accurately calibrated to a standard audio reference level.

Although, facilities can adopt their own in-house standards, typically, a 1,000Hz audio tone should register 0dB on analog equipment and -20dB on digital equipment. 

At the same time, production facilities can set their own internal standards as long as they remain consistent throughout the facility and everyone knows what they are.
  

Digital Standards

In 1985, the Audio Engineering Society and the European Broadcasting Union developed the first standard for digital audio. This is referred to as the AEB/EBU standard. This standard was amended in 1993.

Before this standard was adopted digital audio productions done in one facility could experience technical problems when moved to another production facility.
 

Digital Audio Time Code

Although we will cover time-code when we talk about video editing, we need to mention at this point that digital audio systems make use of similar system of identifying exact points in a recording.

This is essential in the editing process in order to identify and find audio segments, as well as to keep audio and video synchronized.

But, as we will see when we talk about video time code, in the process of converting frame rates between the 24, 30, and 29.97 (the different video standards), timing errors develop.

Unless the audio technicians are aware of these differences and take measures to compensate, after a few minutes video and audio can get noticeably out of sync. You've probably seen movies where the lip-sync was out and the words you were hearing didn't exactly match the lip movements of the actors.

People working with digital audio should at least be aware of the potential problem, and before a video project is started, consult an engineer about the possible problems that could arise in the conversion process. It's much easier to head off these problems before a projects starts than to try to fix them later.

In the next section  we'll talk about audio control devices.