Samplingrate

Sampling rate, often abbreviated Fs, is the number of samples of a continuous-time signal taken per unit of time to create a discrete-time representation. It is measured in samples per second, or hertz. The sampling interval is 1/Fs. In digital signal processing and data acquisition, the sampling rate determines how faithfully the original waveform can be represented in the digital domain.

The sampling theorem states that if a signal is band-limited to B Hz, it can be perfectly reconstructed from its samples if Fs is greater than or equal to 2B, the Nyquist rate. If Fs is less than 2B, higher frequency components fold into lower frequencies, causing aliasing. To prevent this, anti-aliasing filters remove energy above B before sampling.

Common values and practices vary by application. Audio CD uses 44.1 kHz; professional audio often uses 48 kHz; high-resolution audio can use 96 kHz or 192 kHz. Video and multimedia workflows frequently synchronize to around 48 kHz for the audio track. Measurement instruments may employ much higher rates, depending on the signal bandwidth and accuracy requirements. In most cases, there is a trade-off between sampling rate, data size, and processing power. Higher sampling rates improve time-domain accuracy and the ability to capture fast transients, but increase data rate and computational load.

In practice, the choice of sampling rate depends on the signal bandwidth, the desired reconstruction quality, and system constraints. Proper filtering before sampling helps minimize artifacts and ensures a reliable digital representation.