Digital audio sample rate does not limit time resolution — dithered audio has effectively unlimited time resolution

A common argument against standard sample rates (44.1/48 kHz) claims the period between samples (~22 µs) is larger than human time resolution (~7 µs), so events cannot be placed in time precisely. This is false: audio events do not start on sample boundaries. A sine wave can be advanced in arbitrarily small time increments by changing the phase — the waveform shape encodes sub-sample timing in the amplitude of every sample. An equation by Mans Rullgård shows time resolution depends only on bit depth and frequency, not sample rate, and 16-bit audio far exceeds human perceptual limits. Furthermore, dither eliminates even bit-depth constraints: the 16-bit dithered signal is mathematically identical to the 24-bit original plus low-level white noise, so any timing information present in the original survives.

Examples

A 44.1 kHz sample period is 22.68 µs. But a 440 Hz sine wave can be shifted by 1 µs simply by adjusting its initial phase — nothing in the digital representation prevents this. Dither proof: 16-bit dithered − 24-bit original = white noise; if the 24-bit has no timing errors, neither does the 16-bit+noise sum.

Assessment

A podcast host claims 192 kHz audio improves timing accuracy over 44.1 kHz because 192 kHz has 4× more samples per second. Explain the flaw in this reasoning using the concept of phase-encoded sub-sample timing.