Digital Audio Recording Terminology Explained
Digital Recording Concepts and Terms
This document defines key terms and concepts essential for understanding digital audio recording and computer hardware in audio production.
Sample Rate
The number of samples of an audio signal taken every second. Standard audio CD format has a sample rate of 44,100 samples per second, often notated as 44.1 kHz.
Nyquist Theorem
The principle stating that a sample rate must be at least twice as high as the highest frequency being sampled in order for that frequency to be accurately reproduced.
Bit
A single unit of computer data, representing either a 0 or a 1.
Bit-Depth
The number of bits of data used to record the position (amplitude) of an audio signal in one sample. The more bits used, the more accurate the recording. For example, 1 bit allows for 2 possible values, 2 bits for 4 values, and 3 bits for 8 values. It determines the precision with which the amplitude (fluctuation and voltage) of the audio signal is recorded.
Quantization
The process of moving the recorded value of a voltage to the nearest discrete value that can be recorded, based on the bit-depth of the recording process.
Dither
The process of adding a small amount of white noise to a signal just prior to it being digitized. This small amount of error can offset the effects of the inaccuracy of digital recording caused by quantization, improving perceived audio quality.
RAM (Random Access Memory)
The region in which the data that the computer needs for the processes it is currently running is stored. This is short-term, volatile memory that is purged as soon as the process is ended. Historically, RAM was purchased in megabyte (MB) chips, but modern systems require gigabytes (GB) for optimal performance. Digital Audio Workstations (DAWs) usually require a large amount of RAM as sound files take up a lot of space, and there can be no disruption of the flow of audio in or out of the sound card, or there will be audible errors during recording or playback.
HD (Hard Drive)
This is the device on which the computer’s operating system, software, and files are stored. Generally, it consists of a series of small magnetic platters (discs) that store data as a series of magnetic polarities. Optimal audio computer design often involves storing the operating system and software on a primary drive (e.g., C: drive) and all audio files on separate, dedicated drives.
Linear Editing/Recording
A recording and editing process, such as onto tape, that does not allow the user to move tracks around in time or relative to one another. Edits are permanent and sequential.
Non-Linear Editing
An editing system that allows recordings to be manipulated in time, either as a block or by individual tracks. Tracks can be copied, pasted, or dragged, offering flexibility and non-destructive workflows.
Destructive Editing
Involves actual data manipulation and alteration, such as cloning. It should usually be performed on a copy of the original RAW data, as the changes are permanent. This method often requires more storage space for copies.
Non-Destructive Editing
Includes the ability to perform edits on a single file (not a copied file) while retaining the option to revert to the original. Edits are stored on”sideca” files or equivalent data structures, but never affect the original bits of the RAW data. This method often benefits from a faster processor to apply edits in real-time.
Bit Error
A single piece of data is corrupted. Error correction often involves redundancy, which sends the same information more than once to avoid or correct the error.
Burst Error
Many samples or large chunks of music are corrupted. Error correction methods include Interpolation, which estimates missing data to form a continuous signal, and Interleaving, which rearranges data to reduce the impact of consecutive errors.
RAID
Is the acronym for Redundant Array of Independent Disks, a data storage virtualization technology that combines multiple physical disk drive components into one or more logical units for the purposes of data redundancy, performance improvement, or both.
RAID 0
A striping solution: provides no redundancy or data protection from the failure of any disk. RAID 0 achieves faster data access by splitting or”stripin” the recorded data across all drives in the array, improving read/write times and effectively spreading the load.
RAID 1
A mirroring solution: All data is recorded on not just one but two identical drives, so that two copies exist at all times. If one drive fails, the exact copy on the other continues to function and no data is ever lost. It effectively presents both drives as a single logical drive. While highly reliable for data integrity, it halves the usable storage capacity.
Hybrid System (RAID 0+1 or RAID 1+0)
Combines striping and mirroring. It offers high performance and data redundancy, ensuring data is not lost, but typically requires more drives, power, and cost compared to single RAID levels.