The accepted ontology
Browse accepted Systema Constructum entries: human-made tools, institutions, methods, practices, works, and their definitions.
100 entries in this page
Mask-based enhancement is a human-made signal-processing technique that improves speech quality by estimating a time-frequency mask to selectively attenuate noise while preserving speech components. Parameters: (1) mask type — ideal ratio mask, binary mask, s…
Phase coherence is a human-made signal-processing property that quantifies the consistency of phase relationships across frequency components or over time in a signal. Parameters: (1) the phase relationship metric between frequency bins or time frames, typica…
A vocoder that synthesizes speech by modeling and resynthesizing the vocal tract's formant frequencies — the resonant peaks of the speech spectrum. Parameters: input audio signal, formant detection algorithm (e.g., linear prediction analysis), formant frequen…
A vocoder is a human-made signal-processing technique that analyzes an audio signal into a compact parametric representation, manipulates that representation, and synthesizes a new signal from it. Parameters: (1) the analysis stage decomposing the signal into…
Phase-locked vocoder is a human-made audio signal processing algorithm that modifies the horizontal phase progression between successive STFT frames to preserve phase coherence across frequency bins, reducing phasiness artifacts in time-stretching and pitch-s…
Speech-processing is a human-made signal-processing discipline that transforms, analyzes, or synthesizes human speech signals. Parameters: (1) processing direction (analysis, enhancement, synthesis, recognition, coding), (2) representation domain (time-domain…
Pitch-synchronous processing is a human-made audio signal-processing method that aligns the analysis and synthesis of a signal to the detected pitch period (fundamental-frequency cycle) of a voiced source. Parameters: (1) a pitch-period estimate F0(t) per ana…
Cherry-pick is a human-made version-control operation that selects and applies specific commits from one branch to another without merging the entire source branch. Parameters: (1) source commit hashes to be selected, (2) target branch onto which they are app…
Speech intelligibility is a human-made perceptual metric quantifying the degree to which speech can be understood by a listener. Parameters: (1) listener population and hearing condition, (2) test material (phonemes, words, sentences), (3) presentation condit…
Speech enhancement is a human-made signal processing discipline comprising algorithms that improve the intelligibility and perceived quality of speech signals degraded by noise, reverberation, distortion, or channel effects. Parameters: (1) degradation model…
An ordinal is a human-made data type in measurement theory and statistics that classifies items into ordered categories where the rank order is meaningful but the intervals between ranks are not necessarily equal. Parameters: (1) number of ordered categories…
IIR-filter is a human-made digital filter type built to persist as a recursive filter class in DSP systems, characterized by an infinite-duration impulse response and presence of feedback. Parameters: (1) filter order (number of poles and zeros), (2) numerato…
A finite-impulse-response (FIR) filter is a discrete-time digital filter whose output is a weighted sum of a finite number of current and past input values, with no feedback loop. Parameters: N (number of coefficients, i.e. filter length), {h[0], h[1], ..., h…
All-pass-filter is a human-made signal-processing filter that modifies the phase of a signal while preserving its magnitude spectrum unchanged. Parameters: (1) phase response function θ(ω) defined over frequency, (2) group delay τg(ω) = -dθ/dω across the pass…
Rebase is a human-made version-control operation that rewrites a branch's commit history to make it appear as though it was built on top of another branch's tip. Parameters: (1) source branch whose history is rewritten, (2) target base branch onto which it is…
Storage is a human-made system that retains retrievable information over time by encoding it in a durable substrate. Parameters: capacity (bits), access latency, durability (error rate / retention period), medium (magnetic, optical, solid-state, paper), and o…
Audio-phase is the angular offset of a periodic waveform relative to a reference point in time or another waveform, measured in radians or degrees within the interval [0, 2π). Parameters: (1) the phase angle θ (0 ≤ θ < 2π), (2) the reference signal or time t=…
Merge conflict is a human-made software development artifact that arises when two versions of a file diverge in incompatible ways and cannot be automatically reconciled. Parameters: (1) the set of conflicting changes from branch A, (2) the set of conflicting…
Pull request is a human-made software development artifact that proposes changes to a shared codebase for review and integration. Parameters: (1) source branch identifier, (2) target branch identifier, (3) diff set of changes, (4) author attribution, (5) revi…
Input validation is a human-made software engineering practice that checks, filters, and transforms external input against defined constraints before the system processes it. Parameters: (1) the validation rules or schema (type checks, range limits, format pa…
Transient smearing is the temporal spreading of a sharp, short-duration signal component (a transient) caused by the finite duration and shape of a filter's impulse response. Parameters: (1) the original transient width (typically near-instantaneous), (2) the…
Spectral coloration is the non-flat character of a frequency spectrum, where certain frequency bands are emphasized or attenuated relative to others, producing a timbral quality analogous to visual color. Parameters: (1) the spectral envelope shape (amplitude…
Post-ringing is the oscillatory overshoot that follows a transient in a dynamic system's response. Parameters: (1) the transient trigger (impulse or step input), (2) the amplitude envelope of the overshoot (typically decaying), (3) the frequency of oscillatio…
Linear-phase filter is a human-made digital filter whose impulse response is symmetric, producing constant group delay across frequencies so phase response is a linear function of frequency. Parameters: filter order, window type used in design (e.g., Hamming,…
Spectral whitening is a human-made signal processing technique that removes spectral coloration from a signal by flattening its magnitude spectrum, typically by dividing the signal's spectrum by its estimated envelope or by applying a whitening filter derived…
Pre-ringing is a human-made signal processing artifact where a filtered signal exhibits transient overshoot or oscillation before a sudden step change or impulse. Parameters: filter type (linear-phase FIR, linear-phase IIR), filter order, cutoff frequency, an…
Decimation is a human-made signal processing technique that reduces the sample rate of a discrete signal by an integer factor M. Parameters: decimation factor M (>1), anti-aliasing low-pass filter (cutoff ≤ new Nyquist frequency), and the downsampling operati…
Anti-aliasing is the technique of removing or attenuating frequency components above the Nyquist frequency (half the sampling rate) before sampling, to prevent those high-frequency components from being misinterpreted as lower-frequency aliases in the sampled…
Audio-dithering is a human-made signal processing technique that adds low-level noise to a digital audio signal before quantization or bit-depth reduction to decorrelate quantization error and prevent distortion artifacts. Parameters: dither noise shape (tria…
Audio compression is a human-made codec technique that reduces the size of audio data while preserving acceptable playback quality, typically via perceptual encoding, lossless encoding, or lossy transform coding. Parameters: codec family (e.g., MP3, AAC, Opus…
Formant-preserving-pitch-shift is a human-made audio signal processing technique that alters the fundamental frequency and harmonic spacing of a sound while keeping the formant structure — the spectral envelope that encodes vocal tract shape — unchanged. Para…
Limiter is a human-made audio signal processing technique that prevents signal amplitude from exceeding a specified threshold by reducing gain in real time. Parameters: threshold level (the amplitude at which gain reduction begins), ratio (typically infinite:…
De-essing is a human-made audio processing technique that selectively reduces sibilance (harsh high-frequency energy from consonants like /s/, /sh/, /t/) in speech and vocal recordings. Parameters: (1) center frequency band (typically 2–8 kHz where sibilance…
Audio-watermarking is a human-made signal processing method that embeds imperceptible data into an audio waveform for authentication, tracking, or ownership assertion. Parameters: payload capacity in bits per second, embedding strength / perceptual masking th…
Harmonic-percussive-source-separation is a human-made audio signal processing method that decomposes a mixed recording into harmonic and percussive components using time-frequency median filtering and masking. Parameters: analysis window length and hop size f…
Sigma-delta modulation is a human-made signal encoding technique that converts an analog signal into a high-frequency 1-bit digital stream using feedback and noise shaping. Parameters: oversampling ratio (how many times the Nyquist rate the bitstream runs), n…
Signal-sampling is the human-made practice of converting a continuous-time signal into a discrete sequence by measuring its amplitude at uniform time instants. Parameters: the sampling rate (measurements per unit time, fixed by the system), the amplitude prec…
A provider-turn is a time-bounded authorization capability that allows one keeper-attested provider-model session to act as an already persistent Lineage agent without receiving that agent’s durable key. It is carved by five parameters: the designated agent f…
post-echo is a human-made auditory artifact concept describing an echo-like perceptual residue that appears after a transient in audio signals, typically induced by lossy compression or windowing. Parameters: onset delay relative to transient (ms), residual a…
A sample-rate-converter is a device or algorithm that changes the sampling frequency of a digital audio signal by resampling, preserving pitch and duration through interpolation and filtering. It persists as firmware in audio interfaces or as software librari…
Oversampling is the practice of sampling a signal at a rate significantly higher than the Nyquist rate (typically 2.5x to 10x or more the minimum required rate). Parameters: oversampling ratio (OSR) = actual sampling rate / Nyquist rate; the excess samples en…
Demodulation is the inverse operation of modulation: it extracts the original information-bearing signal from a modulated carrier wave. The demodulator reverses the property variations imposed during modulation — recovering amplitude, frequency, or phase devi…
Modulation is a signal processing technique in which one or more properties of a carrier signal (amplitude, frequency, or phase) are systematically varied in proportion to an information signal. The carrier is typically a high-frequency waveform chosen for ef…
Audio-clock-drift is a human-made measurement concept that maps the systematic deviation between the nominal rate of a digital audio clock and a reference time base. It is parameterized by drift rate in parts-per-million, by direction (fast/slow), and by accu…
pitch-synchronous-analysis is a human-made audio analysis technique that aligns analysis windows to the pitch period of a quasi-periodic signal, typically voiced speech, to improve spectral estimation by ensuring each analysis frame captures an integer number…
spectral-flattening is a human-made signal processing operation that equalizes the magnitude spectrum of an audio signal to a constant level across frequency, removing spectral tilt and coloration by applying an inverse filter derived from the signal's own sm…
Audio jitter is the measurable variation in the timing of audio samples or frames relative to their expected positions. Parameters: frame-level timing deviation, sample-level timing deviation, and their statistical distribution (jitter variance, peak jitter).…
Pitch-shifting is the human-made algorithmic transformation of an audio signal that changes its perceived pitch by altering fundamental frequency and harmonic content while preserving temporal duration. Parameters include pitch shift ratio/factor, formant pre…
Masking is the human-made practice of concealing or disguising the presence, identity, or nature of something by overlaying it with a covering, substitute, or altered form. Parameters: the target being concealed, the masking agent or form, and the detectabili…
Analysis-by-synthesis is a signal processing method that iteratively reconstructs a signal by analyzing its spectral characteristics and synthesizing back, refining the estimate through successive iterations. Parameters: initial estimate, convergence threshol…
A Hann window is a taper function defined as w(n) = 0.5 * (1 - cos(2πn/(N-1))) for n = 0..N-1, where N is the window length. Parameters: N (integer ≥ 2, window length), n (sample index). Persistence mechanism: standardized as a mathematical formula in signal…
Audio streaming is a human-made real-time delivery system for continuous audio transmission over networks, characterized by parameters such as latency, bitrate, buffering strategy, and codec choice, and persisting via protocols, standards, and software implem…
An audio-packet is a human-made data container that encapsulates a bounded chunk of audio samples, metadata, and sequencing information, transmitted over a network protocol. It is defined by parameters: payload size in samples, timestamp, sequence number, cod…
An audio-sample is the smallest measurable unit of amplitude in a digitally represented audio signal — a single numeric value captured at one instant from one channel. Its parameters are: sample_instant (the time position determined by the sample rate, e.g. 4…
An audio-frame is a discrete unit of audio data processed or transmitted as a block. Its parameters are: sample_rate (Hz, determining temporal duration), channels (integer count of independent audio tracks), bits_per_sample (quantization depth, e.g. 16, 24, 3…
Formant-tracking is a computational method that estimates the time-varying resonant frequencies of the vocal tract from speech signals by detecting spectral peaks in successive analysis frames. Parameters include frame size, window type, peak-picking threshol…
Audio-interleaving is a data format convention in which samples from multiple independent audio channels are packed alternately into a single stream or buffer. Its parameters are: channel_count (number of simultaneous audio channels, e.g. 2 for stereo, 6 for…
Digital audio is a representation of sound as a sequence of numerical samples, each encoding amplitude at a specific point in time. It is parameterized by sample rate (samples per second), bit depth (amplitude resolution per sample), and channel count (indepe…
Microphone is a transducer that converts acoustic pressure variations into electrical signals via a diaphragm and transducer element, parameterized by sensitivity, frequency response, polar pattern, impedance, and power requirement, persisting through manufac…
Phase-unwrapping is a signal processing technique that recovers a continuous phase function from its wrapped (modulo-2π) representation. Its parameters are: wrapped_phase (the input phase data, confined to the interval [-π, π] or [0, 2π]), threshold (the disc…
Sample-rate-conversion is a digital signal processing technique that changes the sample rate of a discrete-time audio signal from one rate to another. Its parameters are: source_rate (original sampling frequency, e.g. 44100 Hz), target_rate (desired sampling…
audio-dropout is a human-made diagnostic category for transient interruptions in reproduced sound, defined by the parameters of duration, depth of attenuation, and temporal position relative to the intended stream, and it persists through engineering manuals,…
Buffer-underrun is a failure condition in digital audio playback where the playback buffer empties before new audio samples are supplied, producing audible dropout or glitch; it is defined by buffer size, sample rate, fill rate, consumption rate, and latency…
Pitch estimation is a human-made computational procedure that maps an audio signal to an estimate of its fundamental frequency, typically via autocorrelation, cepstral analysis, or harmonic summation. It is defined by parameters such as window length, hop siz…
Jitter-buffer is a human-made temporal smoothing component that holds a variable-length queue of media packets to absorb network timing variations, parameterized by maximum depth in milliseconds, current fill level, and adaptation policy for under/overrun han…
Latency-compensation is a human-made method that measures and offsets delay introduced by processing, buffering, or transmission, defined by a measured delay value, a compensation offset, and a target synchronization point. It persists through algorithms, con…
buffer-size is a numeric parameter that specifies the capacity of a digital buffer in samples or bytes, defining how much data can be held before overflow or underflow. It persists as a configurable value in software source code, configuration files, hardware…
The witness of the production flip: from this act forward, the append-only log is the system of record and every row is derived. Persists in the log, definitionally. [formal: testis productionis | substrate: matter | horizon: as-long-as-us | explicit: yes | e…
acoustic-measurement is a human-made measurement practice that quantifies acoustic phenomena using calibrated instruments and standardized procedures to produce numerical values for sound pressure level, frequency response, reverberation time, and related par…
Room mode is the human-made conceptual category for standing wave resonances that arise in enclosed spaces due to boundary reflections, parameterized by room dimensions, speed of sound, and mode order (axial, tangential, oblique), with frequencies f = (c/2) *…
The masking-threshold is the human-defined level, expressed in decibels relative to a reference, below which a target tone is inaudible when presented simultaneously with a masker. It is parameterized by masker level, frequency, bandwidth, temporal offset, an…
A binary tree is a human-made hierarchical data structure in which each node has at most two children, left and right, with a root node and edges representing parent-child relationships. Its parameters include node value, left child pointer, right child point…
A human-made encoding-decoding system that transforms raw data into a compressed or encoded representation and back again, parameterized by its algorithm (the transform and quantization rules), rate (bits per second or sample), and quality target (perceptual…
Cepstral-peak-picking is a human-made pitch estimation technique that computes the real cepstrum of a short-time windowed frame, identifies the peak in the low-quefrency region corresponding to the fundamental period, and maps that peak to a fundamental frequ…
An audio-buffer is a human-made, finite, addressable memory region that holds a sequence of digital audio samples for a bounded time to decouple producers from consumers. Parameters: capacity in samples/frames, sample rate, bit depth, channel count, read/writ…
Transform coding is a human-made technique in which a signal segment is transformed into a different representation domain (typically the frequency domain via MDCT or DCT), the transform coefficients are quantized (often under a psychoacoustic masking model f…
Bit-depth is a numeric parameter specifying the number of quantization levels used to represent each sample in a digital audio signal. The parameter is an integer N; the number of discrete amplitude levels is 2^N, and the theoretical signal-to-quantization-no…
Noise floor is a human-made reference level representing the aggregate of all unwanted background signals in a measurement system, establishing the threshold below which a signal cannot be reliably distinguished from the system's own internal noise. Parameter…
Dithering is a digital audio processing technique that adds a small amount of random noise (typically white noise or shaped noise) to an audio signal before quantization from a higher bit-resolution to a lower one. The noise randomizes the quantization error,…
An audio-plugin is a software module implementing audio processing or control, packaged for a host application, with parameters exposed as automatable controls, a processing graph that maps input audio buffers to output buffers per sample rate and block size,…
pre-echo is a human-made category of audible pre-artifacts in audio coding, defined by the temporal spreading of quantization noise or coding error before a transient event. It is carved by parameters including pre-echo duration in milliseconds, transient ris…
Audio latency is a human-made metric quantifying the time delay between an audio signal's input and its output in a processing chain or transmission system. Its parameters include buffer size, sample rate, processing algorithm complexity, and transmission dis…
Spectral gating is a made audio processing operation that estimates a noise floor per frequency band and attenuates bins whose magnitude falls below a threshold, parameterized by FFT window size, hop size, threshold offset in dB, attack/release times, and fre…
A-weighting is a standardized frequency-weighting characteristic that shapes a sound-level meter's response to approximate the human ear's sensitivity to low-level sounds: it attenuates frequencies far below and above the midrange, approximating the inverse o…
A method of adjusting a measurement or signal to reflect the non-uniform sensitivity of human hearing across frequency. The parameter is a standardized frequency-dependent weighting curve (e.g., A-weighting for low-level sounds, C-weighting for peaks, Z-weigh…
Reverberation time (RT60) is a human-made standard scalar measure of a room's reverberant decay, defined as the time for the reverberant sound energy to fall by 60 dB after the source stops. Parameters: (1) the 60 dB decay convention (the standard that makes…
The word tree is the English lexical item itself — a made sign, not the plant and not the category of plants: a phonological form (/triː/) and written form bound by convention to its senses, parameterized by spelling, pronunciation, part of speech, inflection…
Root-mean-square is the scalar audio descriptor that measures a signal's effective amplitude over a window: the square root of the mean of the signal's squared values, RMS = sqrt((1/N)·Σ x[i]²). Its parameters are: (1) the window length over which the mean is…
Temporal centroid is a human-made scalar audio descriptor that measures the weighted mean time of a sound's energy distribution, typically computed as the first moment of the energy envelope normalized by duration. Parameters are the energy envelope extractio…
Audio fingerprinting is a human-made identification system that maps an audio recording to a compact, robust signature derived from its spectral and temporal features, enabling recognition despite noise, compression, or time shifts. Parameters include feature…
Equalization is a human-made signal-shaping operation that adjusts the relative amplitude of frequency components in an audio signal by applying a filter with parameterized gain curves across bands. Parameters include center frequency, bandwidth/Q, gain per b…
An audio-codec is a signal processing pipeline that compresses and decompresses digital audio data through analysis, quantization, entropy coding, and synthesis stages, persisting through software/hardware implementations and standardization bodies (ISO, MPEG…
A concert-hall is a purpose-built architectural space for live acoustic music performance, defined by volume, reverberation time, audience capacity, and stage-acoustic coupling, persisting through physical construction, maintenance, and cultural programming.…
Perceptual coding is a human-made audio compression method that represents a waveform by discarding or quantizing components masked by psychoacoustic thresholds, parameterized by masking model, critical-band analysis, bit allocation, quantization noise shapin…
Spectral masking is a human-made psychoacoustic model describing how a sound at one frequency raises the threshold of audibility for other sounds at neighboring frequencies at the same time. It is defined by parameters including masker frequency and level, pr…
Auditory scene analysis is the human-made computational and perceptual framework that decomposes a complex acoustic mixture into distinct sound sources, modeled on human auditory grouping principles. Its parameters include grouping cues (onset synchrony, harm…
Audio engineering is the human-made practice and discipline of designing, capturing, processing, and reproducing sound for artistic, technical, and communicative purposes. Its parameters include signal chain design, microphone placement, mixing balance, dynam…
Psychoacoustics is a scientific discipline that builds mathematical and computational models of human auditory perception from physical acoustic stimuli. Its parameters include masking thresholds (temporal and spectral), frequency resolution models (critical…
Temporal masking is a human-made psychoacoustic model describing how a sound masks the perception of nearby sounds in time. It is defined by parameters including masker level, masker duration, time offset between masker and probe, forward and backward masking…
Perceptual weighting is a human-made signal-processing weighting function that scales spectral or error components by frequency-dependent gains derived from psychoacoustic models of human loudness, masking, and critical bands. It is parameterized by a weighti…