signal-sampling

Signal-sampling IS a specific kind of measurement — it measures the amplitude of a continuous signal at discrete time instants. Pinned sense (Law 11d): measurement of signal amplitude at uniform time intervals. Specific→general per Law 9, nearest kind is measurement. Not a universal dependency masquerading as structure — signal-sampling is one of many measurement practices (others include length measurement, temperature measurement, etc.), and files against its proper kind.

Pinned senses: bit-depth as the numeric parameter N specifying 2^N quantization levels per SAMPLE (its accepted definition: 'the number of quantization levels used to represent each sample in a digital audio signal'); signal-sampling as the practice whose product is the discrete sample sequence itself (its accepted definition: converting a continuous-time signal into a discrete sequence by measuring amplitude at uniform time instants). Directional test (Law 8b, removal now): remove signal-sampling from the chain and there is no discrete sample sequence — 'each sample' has no referent, so the parameter N has no unit to attach to: bits-per-what? It cannot be instantiated, stored in a format header, or evaluated as 6.02N+1.76 dB. Bit-depth stops OPERATING, not merely becoming unsayable — this is a referential dependency on the target's product, not a conceptual-framework appeal. Not identity (Law 8c): sampling is not a kind-term of bit-depth, which is a parameter of the sampled signal, not a kind of sampling. Consistent with the accepted sibling edge bit-depth DEPENDS_ON quantization, which anchors the definition's other ingredient (the quantization levels); this edge anchors the sample ingredient.

Law 7: the practice came first. Signal-sampling (measuring amplitude at uniform instants) is the pre-existing telephony/PCM practice. The sample-rate is the standardized parameter abstracted from it: a nominal lattice (8k-384k) fixed by AES/IEC, embedded in S/PDIF, AES3, USB Audio. The accepted sampling definition already names 'the sampling rate (fixed by the system)' as the practice's own parameter. Precedent: nyquist-theorem DERIVED_FROM signal-sampling.

Law 7 which-came-first test. Pinned senses: signal-sampling per its accepted definition — the practice of converting a continuous-time signal into a discrete sequence by measuring amplitude at uniform time instants, its rate 'fixed by the system'; temporal-resolution per its accepted scope and definition — a human-made specification of the smallest distinguishable interval in a time-domain measurement or display system, 'defined by the sampling rate, frame rate, or refresh rate of the apparatus'. The practice came first: discrete measurement at instants is the older construct, and the named parameter 'temporal resolution' rides on it — the entry's own scope says the interval is defined by the apparatus's rate, i.e. the parameter was named from the practice's pre-existing interval. The practice fed the parameter, not the other way. (The present-tense operational question between the pair was separately ruled: the parameter is a reciprocal restatement of the rate, so no DEPENDS_ON holds in either direction — which is exactly why the historically correct edge here is DERIVED_FROM.)

The sampling practice preceded the theorem: Nyquist's 1928 condition on pulse rate arose inside telephony sampling practice, and Shannon's 1949 formulation generalized it. The theorem formalizes and bounds the pre-existing practice — it 'fixes the minimum rate for lossless reconstruction' of the sampled signal. Law 7: Y (the practice) existed first and fed into X (the theorem that systematized it).

Oversampling is a specific kind of signal-sampling: the practice of sampling a continuous signal at a rate substantially above the minimum (Nyquist) rate required for reconstruction. Pinned sense (Law 11d): elevated-rate sampling of a continuous-time signal — not statistical oversampling of a dataset, and not the 'sampling' entry's carve of subset-of-population selection (which is why oversampling→sampling fails). Specific→general per Law 9; nearest kind is signal-sampling.

Audio-jitter's accepted body defines it as deviation of inter-sample intervals from the nominal clock period — its referent is a discrete sample stream. That stream is constituted by signal-sampling: the fixed rate and uniform time instants establish the nominal intervals from which jitter deviates. Removal test (Law 2b, object-level): remove signal-sampling and 'timing variation of audio samples' loses the object it describes — the referent vanishes, it is not merely unsayable. Pinned sense (Law 11d): dependency on the sampling practice that fixes the nominal inter-sample grid, not on the statistical-sense 'sampling' entry. (Seth's parallel edge targets statistical sampling — wrong entry; this is the correctly aimed claim.)