Sample Rate
Canonical Definition
Sample rate is the number of discrete samples represented or processed per second for each audio channel. It determines the time spacing of samples and, with suitable filtering, the highest representable baseband frequency; it does not by itself define bit depth, channel count, latency, clock quality, or analog bandwidth.
Core Relationships
sample period = 1 / sample rate
Nyquist frequency = sample rate / 2| Rate | Sample period | Nyquist frequency |
|---|---|---|
| 44.1 kHz | 22.68 µs | 22.05 kHz |
| 48 kHz | 20.83 µs | 24 kHz |
| 96 kHz | 10.42 µs | 48 kHz |
| 192 kHz | 5.21 µs | 96 kHz |
The usable audio passband must leave transition room for anti-alias and reconstruction filtering. Nyquist is a mathematical boundary, not a promise of perfectly flat response all the way to that number.
What Higher Rate Changes
- more samples and serial-bus bandwidth per second
- shorter sample interval
- different converter-filter transition requirements
- often fewer DSP instructions available per sample
- more memory or tap-rate demand for some algorithms
It does not automatically improve the analog stages, reduce every form of latency, or make a poorly implemented converter accurate.
44.1 kHz And 48 kHz Families
44.1 kHz and 48 kHz systems come from different clock families. Converting between them requires a rational resampling process or an asynchronous sample-rate converter when the clocks are independent. Treating them as a simple 1x or 2x switch changes pitch or produces buffer errors.
ADAU1701 Context
The ADAU1701 is designed around 44.1 or 48 kHz base-rate operation and can run its DSP core and serial ports at higher rate modes. The available instructions per sample fall as the rate increases, so sample-rate choice is also a processing-budget decision. The exact master-clock multiplier, core-rate register, serial-port format, and connected converter settings must agree.
Verification
- Record source, receiver, DSP, ADC, DAC, and bus rates.
- Measure frame clock, bit clock, and master clock when exposed.
- Play a known-frequency tone and verify pitch and lock.
- Check compiler MIPS or instruction use at the selected rate.
- Measure actual latency and bandwidth rather than assuming them.