Digital Audio
Constant Bit Rate
CBR; Codecs encodes data at a constant rate regardless of density of the audio file
Coaxial
Cable used to transmit data; Inner cable is surrounded by a plastic insulator, which is surrounded by a wire mesh conductor that insulates the internal signal wire from external interference and an outer casing that functions as a ground
CPU Buffering
Channels are processed one at a time and the results are stored on multiple CPU buffers that alternately send data as DAW requests the data for playback; First few seconds are relayed to the buffer prior to processing and playback and it continuously buffers after that; Effective streaming technique but is very susceptible to latency
Decoder
Circuit that interprets the meaning of the symbols as they were chosen and arranged by the encode
Sample-and-Hold
Circuit that seizes voltage values with each tick of an A/D device's internal clock
AoE Formats
CobraNet; EtherSound; Dante; AVB (currently under development)
Codec
Computer program or device capable of encoding and/or decoding a digital data stream with the end result being a reduced file size
Sawtooth Wave
Contains all even and odd harmonics associated with a fundamental tone, making it a rich source for modeling other sounds; Amplitude of each overtone decreases exponentially as a ratio of the harmonic's frequency to that of the fundamental
Conversion Buffering
DAW's and software must buffer when converting or bouncing and this latency can add up if not monitored
Optical Cable
Data is transmitted over fiber optic lines; Uses a TOSLINK connecter instead of an RCA type; Can transmit multi-channel audio; Not susceptible to ground hum and loops; Able to support far higher rates of data transfer over greater distances than coaxial cable
Lossless
Data reduction technique that does not effect quality of original audio; No effect on original quality; Typically around 50% reduction; Exact reconstruction of digital code for the audio signal
Lossy
Data reduction technique that selectively removes original information in order to significantly reduce the file size; Some data is lost; Files can be reduced up to 99% in size (90% with no perceived sound quality loss); Bit rate effects the perceived quality of reduced audio file
Audio over Ethernet (AoE)
Data transmission protocol over which computer network traffic travels; Poorly suited to real-time transmission but numerous attempts have been made to harness the technology because of its flexible routing; Uses simple Cat5 cabling; Resists interference and attenuation
dBFS
Decibels Full Scale
Sample Rate
Defines the number of sample per second taken from a continuous signal to make a discrete signal; Governs the frequency response of digital audio
Logical Format
Describes acceptable data, performances both offered and essential for a disc player, and the complete user experience
Physical Disc Format
Describes various optical disc characteristics including the size and shape of the disc, the size of pits, the speed at which the disc spins, and a multitude of aspects regarding the specifications of the player itself
Jitter
Deviation from a normal, steady pulse or tick of a clock that contributes to misrepresentation of a signal; Result of small timing irregularities that become magnified during the transmission of digital signals as the signals are passed from one device to another
Index of Reflectivity
Difference in brightness between land and pit on a CD Physical Format
Pulse Width Modulation
Digital (binary) measurements of how long each pulse is either on or off; Width of increasing voltage or decreasing voltage is assigned a 1 or 0 respectively
D/A Conversion Signal Flow
Digital Word -> Series of Resistors (each with assigned charges) -> Sample-and-Hold Circuit -> Anti-Imaging Filter (Smoothing Filter) -> Reconstructed Sample
Delta-Sigma Modulation
Digital and analog processing capability is combined on a single microchip allowing for 1-bit resolution at high sample rates
Quantization Intervals
Discrete incremental distinctions made between the value of one sample and the next; Breaks down bit depth into a series of evenly spaced intervals
European Broadcasting Union
EBU
Edit Decision List
EDL; Final list of samples used in the audio editing process; Identified by time code
6 dB
Each bit in the bit depth is equal to a _____ increase in dynamic range
Dolby 7.1
Eight channel digital surround sound system by Dolby
Photoreceptor
Electromagnetic receptor that detects the radiation known as visible light
Anti-Aliasing Filter
Eliminates frequencies above the Nyquist limit from becoming samples; Occurs prior to quantization
Claude Shannon
Father of modern information theory; Solidified the Nyquist Theory by adding the concept that bits per second (binary representation of audio signals) must be at equal intervals to accurately represent data
Harry Nyquist
Found that aliasing was always a problem no matter how fast you sample; Less data recorded but more accurate; 2 samples per wave length.
Significand
Fractional part of a floating-point number; Also called the mantissa; Defines precision
Joseph Fourier
French mathematician that noted that any complex sound can be broken down into a series of component pure tones
dB/FS
Full Scale; Type of metering that measures level in digital recording system; Recording and Mixing levels should NEVER exceed 0dB FS in digital audio or clipping will occur
Nyquist Frequency
Governs the frequency response of a digital system; The highest-frequency component that can be captured with a sampling rate; always 1/2 of sampling rate; Also called the limiting frequency
DVD-Audio
HD Audio format; Lossless Compression; 24-bit/96 kHz; 5.1 Surround or 24-bit / 192 kHz stereo sound
Square Wave
Have odd numbered harmonics
Sinusoidal
Having a repeated succession of waves or curves as in a sound waveform
Compression
High Pressure, Part of a longitudinal wave where the particles of the medium are close together
Multichannel Audio Digital Interface (MADI)
High channel count; 64 channels on one cable; Coaxial cable with BNC connector or fiber optic with ST1 connector
2 Dimensions of Sound
How Loud (Y-Axis) & How Fast (X-Axis)
Sampling Theorem
If a signal is sampled at a rate higher than twice the highest significant signal frequency and at evenly spaced intervals, then the samples contain all the information of the original signal
Acoustics
Pertaining to hearing or sound; Combination of the intensity of air pressure molecules with amplitude
Buffering Locations
Playback; I/O Connections; CPU (Streaming); Conversion from DAW or Software
Additive Synthesis
Process of building a complex tone by starting with the fundamental frequency and adding pure tone harmonics
Spectrum Multiplication
Process that begins with a fast FFT analysis of the spectra of two input signals, then the multiplication of like frequencies, and IFFT to finalize the process
Perceptual Coding
Psychoacoustic model of data reduction used for general audio compression that aims to transmit only features perceptible to the human ear; Useful for mastering streaming media
Buffering
RAM holds in memory audio data before it is transferred to the memory controller; Certain amount of data is processed before large amounts of data are streamed to prevent latency
Voltage
Rate at which energy is drawn from a source that produces a flow of electricity in a circuit; Expressed in volts
Signal-to-Noise Ratio
Ratio of magnitude of the analytical signal to the magnitude of the background noise signal
Decimation Filter
Reduces A/D sample rate from the oversampled rate to nominal rate by turning series of one-bit samples into a series of multi-bit PCM samples; (ex. => 2.8MHz sample rate converted to 44.1kHz and simultaneously converts 1-Bit samples to multi-bit); Low Pass Filter finishes the process
Dithering
Reduces the percieved distortion due to quantization error; Low level white noise source is introduced to make the profile of the quantization "noise" more irregular; Useful when reducing the number of bits per word in a signal (i.e. when converting 20 bit to 16 bit resolution); Introduced at 1/2 of quantization interval when bouncing at shorter word length
Aliasing
Improper (false) sampling of high frequencies that cause much lower frequencies to be produced when the audio is reconstructed; Foldover happens at higher frequencies as harmonics reach levels that exceed the Nyquist limit.
Playback Buffering
In order to play multiple channels at one time data is buffered as read to disk; Latency happens between play command & beginning of playback
Interpolation Filter
Increases D/A sample rate from nominal rate to oversampling rate by turning series multi-bit PCM samples into 1-bit samples; (ex. => 44.1kHz sample rate converted to 2.8MHz and simultaneously converts multi-bit samples to 1-bit); Low Pass Filter finishes the process
Normalizing
Increases or decreases the digital signal so that the loudest sample is brought up to 0dBfs; Uses all bits from dynamic range and makes it even from track to track
Pulse Density Modulation
Reference voltage determined by summing the voltage values of a predetermined number of previous samples; Numbers of 1 in row = waveform peak; Numbers of 0 in row = waveform trough; Basis of Sony's Direct Stream Digital (DSD)
Recording Levels
Industry Standards: -6 dB Peak = -20 RMS Meter
Stapedes Reflex
Inner ear component that attaches to the stapes and helps to decrease the amplitude of vibrations; Causes the masking phenomenon
Anti-Imaging Filter
Removes high frequency images and noise and smoothes the stair case output coming from of the sample and hold circuit; Also called a SMOOTHING FILTER
Intensity Stereo
Joint-Stereo Technique; Since the human brain is unable to localize sounds at high frequencies well sounds above 9 kHz threshold are encoded in mono
Pad Head & Tail
Leaving space at beginning and ending of song for data crunching during mastering; Last step in mastering process
Threshold of Pain
Level above which audible sounds are painful (125 - 130 db)
Blu-Ray
Lossless Format; Can hold up to 25GB on a single-layer disc and 50GB on a dual-layer disc
Rarefaction
Low Pressure; Part in a longitudinal wave where the particles are spread apart
Motion Pictures Experts Group
MPEG; Standardizing body of audio coding
Quantization
Represents the amplitude component of the digital sampling process; Technique of incrementing a continuous analog event into a discrete set of binary digits (bits)
RMS
Root Mean Square; Refers to taking the square root of all instantaneous amplitudes; Takes the average of those squares; (-6 Peak Level is approximately equal to -20 RMS)
Speed of Sound
Roughly around 1,130 ft/s
Equal Loudness Contour
Measure of sound pressure over the frequency spectrum, for which a listener perceives a constant loudness when presented with pure steady tones
Intensity
Measure of the amplitude of a longitudinal wave
Sampling (Samples)
Measurement at regular intervals of the amplitude of a varying waveform (in order to convert it to digital form); There must be a minimum of 2 samples for each cycle in a waveform
Peak Level
Measures the highest levels of a signal being recorded or mixed; Monitors for clipping, which occurs at 0dBFS); Does not always reflect perceived volume of signal
Quantizer
Measuring equipment in A/D conversion that processes voltage and provides a value for that voltage
Digital
Method of representing an acoustic quantity with a series of binary numbers; Can have only specific individually distinct values
Oversampling
Method of sampling data at a higher resolution (higher sample rate) as a means of reducing harmonic content during D/A conversion; (x2) oversampling gets rid of all odd harmonic content
Pulse Code Modulation
Method used in digital recording and reproduction in which a signal is sampled at various points and the resulting value is translated into binary numbers
Interleaved
Mixing data and control characters in a single operation
Amplitude Accuracy
More accuracy in low amplitudes and less in higher amplitudes
Joint Stereo
More aggressive lossy data reduction techniques that require further manipulation of the stereo field; Examples are "Intensity" & "M-S"
Entropy Coding
Most significant lossless coding technique in current use; Measure of disorder in which long strings of data are represented by short symbols and uses the shortest symbols to represent the most common repetitive audio data maximizing data reduction
PCM
Multi-Bit Words; (Pulse Code Modulation)
Bit Rate
Number of bits per second processed when sampling sound; (Sampling Rate x Bit Depth) = Resolution
Bit Depth
Number of bits used to represent the smallest unit of information in an audio file; Greater bit depth = better quality audio
Exponent
Number or variable that represents the number of times the base of a power is used as a factor; Defines magnitude
I/O Connection Buffering
Occurs as data is assembled into meaningful bits or information and as left & right channels are separated
Base 2 System
Only 2 digits used; The value of each place (ones, hundreds, etc.) are as follows from greatest to least: 128, 64, 32, 16, 8, 4, 2, 1
Harmonic Content
Overtones that contribute to the timbre of a sound and make up a complex waveform's physical characteristics
Masking Analysis
Perceptual coding technique that uses louder sounds of a similar frequency to decide what information is to be saved during data reduction
A/D Conversion
"Capturing" part of digital audio; Never captures a signal perfectly
D/A Conversion
"Reconstructing" part of digital audio
AES3
(AES/EBU); 110Ω, 2-channel balanced digital audio cable with an XLR connection; NOT a mic cable!!
Y-Axis Terminology
(Amplitude Based) Amplitude: Voltage; Quantization; Bit Depth; Quantization Intervals; Quantization Noise; [Signal:Quantization Noise Ratio]; Dither; Dynamic Range
X-Axis Terminology
(Time Based) Frequency: Aliasing; Anti-Aliasing Filter; Sample Rate; Nyquist Limit
Storage Conversion Steps
1.) Bit Rate x Sample Rate (you'll get b/sec) 2.) Multiply by 60 if converting seconds to minutes 3.) Divide by 8 to convert bits to Bytes and get B/min 4.) Divide by 1,024 to get KB/min and keep doing it until you get desired bit rate specification (result should be under 1000) 5.) Multiply number of stereo channels 6.) Multiply length of time
Requirements for A/D Conversion
1.) Taking a series of evenly-spaced measurements 2.) Signal contains no frequency components higher than half the sample rate
DVD-14
12.33 GB; DS/ML
Red Book
12cm plastic disc; 1.2mm thick; One-sided; Red Laser; 1.6 microns between tracks; 125 nanometer pits
DVD-18
15.9 GB; DS/DL
Requirements for CD Audio
16-Bit; 44.1 kHz; PCM; Stereo
CobraNet
1st commercially successful AoE format for the transmission of digital audio, video, and control signals over 64-channel 100Mbps Ethernet networks
DVD-5
4.38 GB; SS/SL
DVD-9
7.95 GB; SS/DL
Tascam Digital Interface Format (TDIF)
8-in/8-out on one cable; 25-pin D-sub connector
DVD-10
8.75 GB; DS/SL
Impulse Response
A digital filter's time domain output sequence when the input is a single sample is input
Blu-Ray
A drive that can read and write on optical media that hold up to 50 GB on two layers; 24-bit/96 kHz for 8-Channel; 24-bit/192 kHz for 6-Channel
Cutoff Frequency
A frequency specified for a filter (digital or electronic) the marks the point at which the frequency content of a signal is altered +/- 3dB
Floating Point
A method of representing real numbers using a mantissa and an exponent
Ethernet
A network communications protocol that specifies how machines will exchange data; Uses a broadcast system in which one machine transmits its message on the communication medium and the other machines listen for messages directed to them
Convolution
A sample-by-sample operation on two signals
Overflow
A situation where a calculated value cannot fit into the number of digits reserved for it
Word Clock
A time regulator that makes all samples and bits to align when working with interconnected digital devices; Basically a signal that all of the digital devices refer to when operating.
Lossy Formats
AAC (Advanced Audio Coding); MP3; RA; WMA; OGG Vorbis; Dolby Digital/AC-3; DTS; ADPCM
Average Bit Rate
ABR; Codecs that encode data by determining how dense or sparse areas of the audio are while also keeping bit rate within specified limits to avoid rebuffering
Lightpipe
ADAT Optcal; 8-in/8-out on two cables; Fiber-optic, TOSLINK connector
Adaptive Pulse Code Modulation
ADPCM; Pulse code modulation that produces a digital signal with a lower bit rate than standard PCM; Records only the difference between samples
Audio Engineering Society
AES
Fidelity
Accuracy with which an electronic system reproduces the sound or image of its input signal
M-S Stereo
Algorithm uses matrix of a mid/side microphone pair to determine a side signal & that signal is reduced then distributed as code in stereo
Footroom
Allowance of noise floor below that which is required for the final product
Oversampling
Allows for an internal sample rate at multiples of the input and output rates; Alleviates the need for steep "brickwall" filters; Often combined with internal "1-bit" processing; Increases smoothing effect
Fletcher-Munson Curve
Also known as equal loudness curves; Graph that indicates the average ear sensitivity to different frequencies at different SPL levels (as volume increases, these curves flatten out)
RMS Meter
Amplitude meter that takes the square root of all instantaneous amplitudes and averages them to find a mean and squares that value; Useful with particularly complex waveforms
Redither
Anytime bit depth is reduced the gap gets bigger so more dithering is required
Lossless Formats
Apple Lossless; Windows Media Lossless; DTS HD Master Audio; Dolby True HD; FLAC
Sample Rate Effect on Anti-Aliasing
As sample rate is increased more room is created for a smoother slope of the attenuation band because Nyquist limit extends well beyond range of hearing with each increase
Effective Bit Depth
Based on Full Scale (dB/FS); -6dB represents a loss of one bit so account for this when calculating
Subbands
Based on psychoacoustics, these are the basis of frequency analysis for a perceptual codec;
Psychoacoustics
Branch of psychology concerned with the subjective perception of sound
Headroom Bits
Built into DAWs; Bits are added when signals are mixed together to avoid clipping
Foldover
Same as "aliasing"
Non-Compressed Audio Data Rate Formula
Sample Rate x Bit Depth x # of Channels
Noise Shaping
Samples are duplicated and the playback sampling rate correspondingly increased; Significantly raises the Nyquist limit to a range well beyond human hearing; Processing "pushes" the distortion resulting from quantization error into these higher frequency ranges, rendering it less "visible" to our ears
SCMS
Serial Copy Management System; main difference between AES3 & S/PIDF
Morse Code
Series of dots and dashes representing the letters of the alphabet; Most common letters are represented by the shortest dots and dashes; Example of entropy coding
Direct Monitoring
Signal conversions are mixed with playback tracks resulting in near-zero latency
Analog
Signal that uses variable voltage to create continuous waves resulting in an inexact transmission
Successive Approximation
Signal voltage is relayed to a register from sample-and -hold circuit; Holds reference frequencies in binary form that decrease in value; Finds approximated value & assigns binary number accordingly
Sony-Philips Digital Interface Format (S/PDIF)
Single-pin RCA cable or fiber-optic TOSLINK connector used for digital transfer; 75Ω coaxial, 2-channel unbalanced; "Consumer" format of AES3
Dolby 5.1
Six channel (five speakers and one subwoofer for bass) digital surround sound system by Dolby
Resolution
Smallest interval measurable by a scientific instrument; Defined by bit rate (sample rate x bit depth)
Frames
Snippets of time in which frequency analysis takes place in a perceptual codec
Threshold of Hearing
Softest sound that can be heard by the average human ear (0 dB)
SACD
Sony and Philips optical disc format; Utilizes sigma delta DSD to offer higher resolution; 1-bit; 2.8224 MHz; 6-Channel
Direct Stream Digital
Sony sigma-delta modulation based technology that bypasses the decimation and interpolation steps found in PCM converters
Algorithm
Specific set of instructions for carrying out a data reduction technique that determines how to "save" binary data information efficiently
Zero-Latency Monitoring
Splits the input signal and mixes it with an analog copy so that no latency is present
Inter-Channel Redundancy
Stores only one copy of a stereo signal and assigns it to both channels in order to save 50% of original bandwidth
Decimal-to-Binary Conversion
Subtract place values from the decimal number and place ones or zeros in the correct places
Fourier Series
Sum of all harmonics; Sum of sine and cosine waves which have frequencies f, 2f, 3f, 4f...
MONO
Take up half as many bits (50%); Algorithm can keep same quality by lossing the stereo track
Internal Resolution
The ability of a digital system to perform complex DSP without running into problems with overflow or loss of resolution
Oscillation
The act of a frequency swinging back and forth with a steady, uninterrupted rhythm
Spectra
The amount of energy at each wavelength
Gain Staging
The art of deciding where to place a processor in signal flow based on how that processor will be influenced by the other processors in the path
Attenuation
The continuous loss of signal strengths as a signal travels through a medium
Quantization Error
The difference between the analog value and the approximated digital value due to the "rounding" that occurs while converting the analog signal to digital
Dynamic Range
The difference in volume between the loudest and quietest sounds of a source
Latency
The elapsed time it takes for a packet of data to arrive at its destination; Lagging or pause of an audio signal as digital processing occurs; Can be managed utilizing several forms of "audio monitoring"
Cutoff Frequency
The frequency above or below which attenuation begins in a filter circuit
Pass Band
The frequency range that is allowed through a filter
0 dB FS
The loudest point of a Full Scale system
Digital Signal Processing
The mathematics, algorithms, and the techniques used to manipulate signals after they have been converted to digital form
Bit Depth Effect on Dynamic Range
The more bits allocated during quantization, the more accurate the measurement
Frequency
The number of compressions or rarefactions in one second; The higher the frequency the more compressions & rarefactions per second; Measured in Hertz
Compression
The process of reducing the space required to store data by efficiently encoding the content.
Transfer Protocol
The set of rules that computers use to move files from one computer to another on an internet
TOSLINK
Toshiba developed digital audio interface utilizes fiber optics as a transmission medium.
Sampling Rule
Twice as many samples as the highest frequency at minimum; Sampling rate totally controls frequency response
EtherSound
Ultra low-latency, 512-channel (on a gigabit network), less flexible AoE format; Routed like audio cables...not network cables
Nanometer
Unit of measurement that is equal to one billionth of a meter
Micron
Unit of measurement that is equal to one millionth of a meter
Glass Master
Used as the main disc from which other discs are made; Composed of ground glass with a very fine photoresistor layer; An imaging laser burns pit and land patterns in preparation for duplication
dB/SPL
Used when the reference pressure of a sound is 20 microPa (0.00002); Sound Pressure Level; Measure of amplitude
Data Packing
Uses entropy coding as the basis; Computer data compression algorithm that packages files such as .ZIP & .RAR
Variable Bit Rate
VBR; Most common & best data reduction technique; Codecs that encode data by determining how dense or sparse areas of the audio are; Can result in buffering issues
Class-D Amplifier
Very quiet digital amplifier that produces a series of output pulses with the audio signal coded the same as the width of the output pulses; Pulses are used to represent wave forms and are either on or off; Intense signals have long pulses with short gaps; Cheaper, cleaner and used widely
Low-Latency Monitoring
Very selective method of lowering buffer levels by halting different levels of audio processing
Sonogram
Visual graph that shows how loud a sound is at different frequencies
A/D Conversion Signal Flow
Voltage -> Dither -> Anti-Aliasing (Low Pass Filter) -> Sample & Hold Circuit -> Successive Approximation/Quantizer --) 100111010 (PCM Audio File)
Sine Wave
Waveform of a pure tone showing simple harmonic motion
Buffer Size
When recording you want the smallest buffer available; When mixing you want the largest buffer available