Sampling takes periodic samples (voltages) of the original analog signal at fixed intervals and converts them to digital data. The rate at which the fixed intervals sample the original signal each second is called the sampling frequency, or sampling rate. For example, a sampling frequency of 48 kHz means samples are taken 48,000 times per second, or each sample period is 1/48,000 second. Because sampling and the component of time are directly related, a system’s sampling rate determines its upper frequency limits. Theoretically, the higher the sampling rate, the greater a system’s frequency range.

In the development of digital technology, it was determined that if the highest frequency in a signal were to be digitally encoded successfully, it would have to be sampled at a rate at least twice its frequency. In other words, if high-frequency response in digital recording is to reach 20 kHz, the sampling frequency must be at least 40 kHz. Too low a sampling rate would cause loss of too much information (see Figure 9-1).

Figure 9-1. Sampling. (a) A signal sampled frequently enough contains sufficient information for proper decoding. (b) Too low a sampling rate loses too much information for proper decoding.

Think of a movie camera that takes 24 still pictures per second. A sampling rate of 1/24 second seems adequate to record most visual activities. Although the camera shutter closes after each 1/24 second and nothing is recorded, not enough information is lost to impair perception of the event. A person running, for example, does not run far enough in the split second the shutter is closed to alter the naturalness of the movement. If the sampling rates were slowed to 1 frame per second, the running movement would be quick and abrupt; if it were slowed to 1 frame per minute, the running would be difficult to follow.

A number of sampling rates are used in digital audio. The most common are 32 kHz, 44.1 kHz, 48 kHz, and 96 kHz. For the Internet, sampling rates below 32 kHz are often used (see Chapter 14). To store more audio data on computer disks, submultiples of 44.1 kHz usually are used, such as 22.05 and 11.025. Multiples of 44.1 kHz (88.2 kHz and 176.4 kHz) and 48 kHz (96 kHz and 192 kHz) are used for greater increases in frequency response.

The international sampling rate—32 kHz—is used for broadcast digital audio. Because the maximum bandwidth in broadcast transmission is 15 kHz, the 32 kHz sampling rate is sufficient. For compact disc and digital tape recording, 44.1, and 48 kHz are used. Generally, standards for the digital versatile disc (DVD) are 48 and 96 kHz. DVD consists of several formats, however, some of which use higher sampling rates (see Figure 9-2; see also “Digital Versatile Disc” later in this chapter).

Depending on the comparative sampling frequencies, there may or may not be a significant difference in frequency response. For example, 44.1 kHz and 48 kHz sound almost alike, as do 88.2 kHz and 96 kHz. But the difference between 48 kHz and 96 kHz is dramatic. Among other things, 44.1 kHz and 48 kHz do not have the transparent response of the higher sampling rates. Transparent sound has a wide and flat frequency response, a sharp time response, clarity, detail, and very low noise and distortion.

While sampling rate affects high frequency response, the number of bits taken per sample affects dynamic range, noise, and distortion. As samples of the waveform are taken, these voltages are converted into discrete quantities and assigned values, a process known as quantization. The assigned value is in the form of bits, from binary digits. Most of us learned math using the decimal, or base 10, system, which consists of 10 numerals—0 through 9. The binary, or base 2, system uses two numbers—0 and 1. In converting the analog signal to digital, when the voltage is off, the assigned value is 0; when the voltage is on, the assigned value is 1.

A quantity expressed as a binary number is called a digital word: 10 is a two-bit word, 101 is a three-bit word, 10101 is a five-bit word, and so on. Each n-bit binary word produces 2n discrete levels. Therefore, a one-bit word produces two discrete levels—0, 1; a two-bit word produces four discrete levels—00, 01, 10, and 11; a three-bit word produces eight discrete levels—000, 001, 010, 011, 100, 101, 110, and 111; and so on. So the more quantizing levels there are, the longer the digital word or word length must be. (Word length is also referred to as bit depth and resolution.)

The longer the digital word, the better the dynamic range. For example, the number of discrete voltage steps possible in an 8-bit word is 256; in a 16-bit word, it is 65,536; in a 20-bit word, it is 1,048,576; and in a 24-bit word, it is 16,777,216. The greater the number of these quantizing levels, the more accurate the representation of the analog signal and the wider the dynamic range (see Figures 9-3 and 9-4).^[1]

Figure 9-3. Quantizing. (a) As the number of quantizing levels increases, the digital sample of the analog signal becomes more accurate. In 16-bit systems the audio spectrum is divided into 65,536 values in a single sample. A 20-bit system captures 1,048,576 values in a single sample. A 24-bit system captures 16,777,216 values in a single sample. (b) The addition of a single bit doubles the data that a digital device can capture.

This raises a question: How can a representation of the original signal be better than the original signal itself? Assume the original analog signal is an ounce of water with an infinite number of values (molecules). The amount and the “character” of the water changes with the number of molecules; it has one “value” with 500 molecules, another with 501, still another with 2,975, and so forth. But all together, the values are infinite. Moreover, changes in the original quantity of water are inevitable: Some of it may evaporate, some may be lost if poured, and some may be contaminated or absorbed by dust or dirt.

But what if the water molecules are sampled and then converted to a stronger, more durable form? In so doing, a representation of the water would be obtained in a facsimile from which nothing would be lost. But sufficient samples would have to be obtained to ensure that the character of the original water is maintained.

For example, suppose the molecule samples were converted to ball bearings and a quantity of 1 million ball bearings was a sufficient sample. In this form, the original water is not vulnerable to evaporation or contamination from dust or dirt. Even if a ball bearing is lost, they are all the same; therefore, losing one ball bearing does not affect the content or quality of the others.

A memory recorder is a portable digital recorder that has no moving parts and therefore requires no maintenance. The storage medium is a memory card, a nonvolatile memory card that can be electrically recorded onto, erased, and reprogrammed. Nonvolatile means the card does not need power to maintain the stored information.

Several models of memory recorders are available. Depending on the design, the storage medium, recording configurations, recording times, bit depths, and sampling frequencies vary. An example of a memory recorder and its features is displayed in Figure 9-5.

Figure 9-5. Zoom H2. (a) This model includes four cardioid microphone capsules, which allow (b) front 90-degree, (c) rear 120-degree, and (d) 360-degree pickup. The 360-degree pickup facilitates conversion of a recording to 5.1 surround sound. It comes with a 512 MB Secure Digital card but can accommodate up to a 16 GB card, allowing for as much as 24 hours of recording time using the 16-bit/44.1 kHz WAV format. The data formats include WAV (44.1 kHz, 48 kHz, and 96 kHz at 16- or 24-bit), MP3 to 320 Kbps, and variable bit rate (bit rate is not constant). Other features include a metronome, a guitar/bass tuner, a low-pass filter, voice-activated recording, and a USB interface. It weighs 4 ounces.

Memory cards have taken portability in digital recording to a new level. They are quite small and lightweight, and some models easily fit into the palm of a hand. They hold a substantial quantity of data for their size, which facilitate long recording times, and memory cards have fast read access times. They are a robust recording medium, as are the recorders, which makes the technology highly suitable for production on location. Most memory recorders provide flexibility in recording formats and supported audio formats; selectable sampling rates and bit depths; wide frequency response; and USB connectivity. Many models include a built-in stereo microphone or two microphones for separate mono or stereo pickup; some models also include editing features.

Examples of memory cards include flash cards such as CompactFlash, flash memory sticks (a family of formats so named because the original cards were about the size and the thickness of a stick of chewing gum), Secure Digital (SD) memory cards, and SmartMedia. The PCMCIA card (named for the Personal Computer Memory Card International Association) is yet another recording medium used in memory recorders. Depending on the recorder, the flash card may be removable or fixed.

Digital recorders also use fixed and removable hard disks. Compared with memory recorders, they usually provide better sound quality and greater recording flexibility (see Figure 9-6). They are available in portable and rack-mountable models.

Figure 9-6. Sound Devices 788T and its control features. (a) This model is about the size of a small paperback book and weighs about 3.6 pounds without the battery. It includes eight mic/line inputs, a selectable sampling rate up to 98 kHz with selectable sampling rate converters on each digital input, peak/VU meters, a word clock, time code, and internal hard drive and a Flash slot, FireWire and USB connections, and a separate USB keyboard input for control. (b) Remote fader controller includes two 2-position toggle switches; programmable to control record, start/stop, and other functions; and LEDs to indicate record and power status. It can be mounted to a boom pole.

The amount of data that memory and hard-disk recorders can encode is impressive given their size. But all technology has its limitations. When using these recorders, especially in the field, it is essential to know their storage capacities in advance so that you do not get caught shorthanded (see Figure 9-7).

Like many digital audio recorders, a digital audio workstation (DAW) records, edits, and plays back. But unlike digital audio recorders, DAWs have considerably greater processing power because of the software programs they use. Generally, there are two types of DAW systems: computer-based and integrated.

The recordable compact disc (CD-R) has unlimited playback, but it can be recorded on only once. The CD-R conforms to the standards document known as Orange Book. According to this standard, data encoded on a CD-R does not have to be recorded all at once but can be added to whenever the user wishes, making it more convenient to produce sequential audio material. But CD-Rs conforming to the Orange Book standard will not play on any CD player.

To be playable on a standard CD player, the CD-R must conform to the Red Book standard, which requires that a table of contents (TOC) file be encoded onto the disc.^[4] A TOC file includes information related to subcode and copy prohibition data, index numbering, and timing information. The TOC, which is written onto the disc after audio assembly, tells the CD player where each cut starts and ends. Once it is encoded, any audio added to the disc will not be playable on standard CD players due to the write-once limitation of CD-R. It is therefore important to know the “color book” standards with which a CD recorder is compatible. Details of these standards are beyond the scope of this book but are available on the Internet.

Compact-disc recorders are available in different recording speeds. For example, single-(1x), double-(2x), quad-(4x)—up to 16x speed. Single-speed machines record in real time; that is, at the CD’s playback speed, 2x machines record at twice the playback speed, reducing by half the time it takes to create a CD, and so on. For the higher recording speeds, the computer and the hard drive must be fast enough.

Playing times vary with the CD format. CDs for consumer format use a 63-minute blank disc. Disc length for the professional format is 63 minutes (550 MB), 74 minutes (650 MB), or 80 minutes (700 MB).

The rewritable CD (CD-RW) is steps better than the CD-R because it can be recorded on, erased, and used many times again for other recordings. If the driver program supports it, erase can even be random. Like the CD recorders, CD-RW drives operate at different speeds to shorten recording times.

After the Orange Book, any user with a CD recorder drive can create a CD from a computer. CD-RW drives can write both CD-R and CD-RW discs and can read any type of CD.

CDVU+ (pronounced “CD view plus”) is a compact disc with interactive content. It was created by the Walt Disney Company to reverse the decline in music CD sales. In addition to the music, it includes multimedia material such as band photos, interviews, and articles relevant to the band or the music, or both.

The digital versatile disc (DVD) is the same diameter and thickness as the compact disc, but it can encode a much greater amount of data. The storage capacity of the current CD is 650 MB, or about 74 minutes of stereo audio. The storage capacity of the DVD can be on a number of levels, each one far exceeding that of the CD. For example, the single-side, single-layer DVD has a capacity of 4.7 billion bytes, equivalent to the capacity of seven CD-ROMs; the double-side, dual-layer DVD with 17 billion bytes is equivalent to the capacity of 26 CD-ROMs.

The CD has a fixed bit depth of 16 bits and a sampling rate of 44.1 kHz. DVD formats can accommodate various bit depths and sampling rates. The CD is a two-channel format and can encode 5.1 (six channels) surround-sound but only with data reduction. DVD-Audio (DVD-A) can encode up to eight audio tracks without data compression.

In addition to DVD-Audio, other DVD formats include: DVD-Video (DVD-V), DVD-Recordable (DVD-R), authoring and general, DVD-Rewritable (DVD-RW), and another rewritable format, DVD+RW, DVD-ROM, and DVD-RAM. Of these formats, only DVD-V has met commercial expectations. Moreover, with the advent of the highdensity optical format, the DVD is being supplanted.

High-density disc technology is another entrant into the competition to meet the demands of high-definition media. The most familiar format at this writing is the Bluray Disc. (BD). Another high-density optical disc format, HD DVD, was developed to compete with the Blu-ray Disc but lost out and is no longer marketed.

In MIDI, digital refers to a set of instructions in the form of digital (binary) data that must be interpreted by an electronic sound-generating, or sound-modifying, device such as a synthesizer or computer that can respond to the directions. MIDI does not create or communicate sound, it communicates instructions. Instructions to a device or program may include creation, playback, or alteration of sound or control function parameters. In other words, the process is not unlike that of a piano roll and a player-piano. The roll itself does not make any sound. When inserted into a player-piano, it instructs the piano to play the programmed sound.

Interface is the link permitting the control signals generated by commands from one synthesizer or controller to trigger other synthesizers and equipment. Thus, one person can “play” several “instruments,” thereby having the capability to create an infinite variety of combined sounds that would otherwise be unachievable.

With MIDI, different voicings from various MIDI devices can be layered to reproduce virtually any sonic structure; multiple hardware and software electronic instruments, performance controllers, computers, and other related devices can communicate and be synchronized with each other over a connected network. Moreover, most MIDI synthesizers are compatible with most others because the entire electronic industry adopted the MIDI specification (see “How MIDI Works” in the next section).

MIDI software is available in a number of categories: (1) performance—software that allows composition, orchestration, arranging, and performing music; (2) productivity—programs that transcribe, data base, and print music using any MIDI setup; (3) editing—for editing digital samples; (4) patching librarians—for storing settings or “patches”; and (5) instruction—software for learning MIDI operations.

A detailed discussion of all of these elements is beyond the scope of this book. It is useful, however, to have some idea of how MIDI works and what a typical MIDI setup includes.

MIDI enables hard- and software-based synthesizers, computers, rhythm machines, sequencers, and other signal-processing devices to be interconnected through an interface. The interface is based on a standard convention, or protocol, called General MIDI, devised by the International MIDI Association (IMA) and agreed to by manufacturers of MIDI hardware and software. General MIDI defines a set of minimum standards among MIDI devices. These standards have been expanded in the General MIDI 2 protocol.

MIDI data is communicated digitally throughout a production system as a string of MIDI messages. MIDI messages may be grouped into two categories: channel messages and system messages. A channel message applies to the specific MIDI channel named in the message. A system message addresses all the channels.

A basic MIDI facility typically includes a MIDI controller, sequencer, hard and/or soft synthesizer(s), computer, MIDI computer interface, sampler and sample CDs, loudspeakers, and appropriate audio and MIDI cables. Other equipment may include a mixer, recorder, and drum machine.

MIDI instruments are connected using a standardized cable with five-pin DIN connectors at each end. (A DIN connector is a connector originally standardized by the Deutsches Institut für Normung [DIN], the German national standards organization.) There is also a five-pin connector that provides MIDI phantom power.

While MIDI devices share the same type of jack, there are three types of MIDI connectors on electronic devices: MIDI IN accepts MIDI signals from another device; MIDI OUT sends signals generated within a device to the MIDI IN of other devices; MIDI THRU is like MIDI OUT, but it passes information arriving at a device’s MIDI IN connector to other devices without regard for internally generated MIDI data. Figure 9-11 displays an example of signal flow in a MIDI setup.

Figure 9-11. Example of signal flow in a MIDI setup.

The following is a typical MIDI signal flow, assuming the use of a software-based synthesizer: MIDI controller; MIDI cable from the controller’s MIDI OUT to the interface’s MIDI IN; MIDI interface; FireWire, USB, or sound card (PCI [Personal Computer Interface]); MIDI driver that facilitates the recording software to transfer data to the interface; sequencer; synthesizer; FireWire, USB, or PCI connection; MIDI interface; and loudspeakers.

The sequencer is the brain of a MIDI setup. It can be a stand-alone unit, a computer that runs a sequencer program, or a circuit built into a keyboard instrument. A sequencer resembles a multitrack recorder for MIDI data. It does not record audio, it receives information from MIDI devices and stores it in memory as separate “tracks.” Once information is in a sequencer’s memory, it can be edited and transmitted to other MIDI instruments for playback.

The advantages of MIDI sequencing over conventional recording are (1) performance and orchestration are completely shapeable in MIDI form, (2) there is no generational loss in copying or manipulating MIDI data, and (3) the amount of data needed to represent MIDI performance is comparatively inconsequential compared to that of digital audio.