XL1 Audio Step by Step, Part One
an article provided by Canon USA
The Watchdog Notes: Special thanks to Mr. Tim Smith of Canon USA for providing this information for the benefit of all XL1 Watchdog readers. Part One begins with a basic description of digital audio recording, and the other chapters deal specifically with XL1 audio features and detailed instructions on how to use them. Enjoy!
One of the many advantages of the new DV format is found in its greatly expanded audio capabilities. The Digital Video standard includes Pulse Code Modulation (PCM) audio recording. In conventional analog recording, sound waves are recorded as changes in the magnetic field on the tape. In digital audio recording, sound is recorded as 0 and 1 after it is converted in pulse codes. This is the reason digital audio is referred to as Pulse Code Modulation. The digital audio code (a series of "off or on" signals) is recorded by the drum, on a part of the tape that is separate from the video information. A provision for PCM is part of the 8mm video specifications, however, Canon has not used this optional sound track in any of its 8mm or Hi8 models to date.
Sound waves are vibrations in the air with two basic properties: the first is frequency, from low (bass) to high (treble); the second is amplitude, from soft to loud. Together they form a simple sine wave. The wave’s amplitude is represented by its height; the further the curve swings above and below its center line, the louder the signal. Its frequency can be represented by the number of times per second the wave goes through a complete "cycle". The more cycles per second, the higher the wave’s frequency. The average young human ear can hear frequencies from about 20 cycles per second (20 hertz, or 20 Hz), a very low base tone, to about 20,000 cycles per second. The distance between peaks is the wavelength, which becomes shorter as the frequency rises.
The camcorder’s microphone picks up sound and outputs an analog signal consisting of minute voltage changes. This signal is then passed through an analog-to-digital (A/D) converter. In a digital recording, the original sound wave is measured at thousands of sampling points per second, and records those voltage samples as numbers. In playback, the sampling points are recreated, and the audio is processed by a digital-to-analog converter (D/A).
The quality of the reproduction depends on how detailed the blueprint is, and how well the reconstruction is done at the playback end. The amount of detail mainly depends on the number of samples per second (which controls frequency response) and the number of binary digits, or "bits" per sample (which controls noise and distortion).
Most sound waves are complex mixtures of simple sine waves. We only need to record two points per cycle of such a wave’s highest frequency to be able to reconstruct the wave in playback. The sampling frequency (the number of times the signal is measured per second) must be high enough to ensure at least two samples for every wave of every audio frequency—at least 40,000 samples per second for an audio band going up to 20,000 Hz.
Digital systems measure in steps, but the analog signals they’re measuring are continuous. An analog signal that ranges between +1 and -1 volts goes through an infinite range of values between those points, but a digital system can record only a finite number of those values. The more digits it has, the more steps it can distinguish and the more closely it can match its readings to the variations in the original signal. Because digital systems use finite means to record infinite signal variations, some mismatch is inevitable, and every such mismatch adds noise and distortion to the signal.
Digital systems use the same binary numbering systems found in computers; that is, each digit only has two possible values, 0 or 1. Each digit added doubles the number of possible values the system can handle: a one-digit number has two values (0 and 1); a two digit number has four values (00, 01 10, and 11; a three-digit number has eight possible values; and so on.
Every time a digit is added to a digital recording system, the amount of its inaccuracy—and, therefore, its noise and distortion—is cut in half. This increase in accuracy is equivalent to cutting noise and distortion by 6 dB; so you can roughly gauge a digital system’s dynamic range by multiplying its digits, or bits, by six. For example, a fourteen-bit system has 84dB of dynamic range, and a sixteen bit system (such as the Compact Disc) has 96 dB.
The number of bits in a system limits the dynamic range. Slight signal overloads don’t cause slight increases in distortion, as they do in analog. In digital systems, they cause sudden, intolerable distortion. Weak signals, no stronger than the system’s noise, are simply not recorded at all. Even though the digital system’s dynamic range is firmly limited, its limits are far wider than those of most analog systems. At 96 dB, those limits are wide enough to accommodate the entire dynamic range of music.
When you copy a signal, you degrade it. In analog, this limits frequency response and adds noise and distortion. There is no degradation in digital reproduction.
Digital has another virtue—no wow or flutter. The tiny speed variations that cause wow and flutter in analog tape recorders are also present in digital ones, but you never hear them. As samples are read off the recording, they’re fed into a buffer circuit, which smooths out the speed variations.
The device responsible for changing an analog signal into a series of numbers is the analog-to-digital converter. It measures (or "samples") the strength of the changing voltage at regular intervals, generating a steady stream of numbers. Two parameters directly affect the quality of the resulting audio: sample rate and bit depth. The converter’s sample rate dictates how often it measures the signal to generate a new value. The more frequently the converter measures the signal the more accurate the resulting data. Sample rate corresponds directly to frequency response—the highest frequency a digital system will capture is exactly one-half the sample rate. To capture the full audio spectrum up to around 20,000 cycles (or 20kHz), a sample rate of 44.1kHz is common. Higher sample rates make for increased treble response and a more "hi-fi" sound. Low sample rates sound duller and darker.
Bit depth affects how many bits the converter uses for each numerical measurement of the signal. More bits equal a more accurate measurement, which explains why 16-bit CD audio sounds so much better than an 8-bit multimedia sound file. A low bit depth allows the converter the measure the sound with a yardstick marked only in inches. A higher bit depth allows the converter much greater accuracy (a yardstick marked in 1/8th inch increments, for example).
See also Part Two, Part Three and Part Four
of XL1 Audio Step-by-Step by Canon USA
Back to the XL1 Articles Menu
Thrown together by Chris Hurd