How sampling turns a changing signal into numbers

Slide 1

Sampling measures a signal at selected times

Sampling measures a changing signal at selected times. The resulting sequence can describe the original only if the sampling process preserves the needed information.

Slide 2

Missed fast changes can appear as slower patterns

Different continuous signals can produce the same sampled sequence. When unwanted high-frequency content appears as a lower-frequency pattern, the result is aliasing.

Slide 3

The sampling rate needs a known frequency limit

For ideal reconstruction of a band-limited signal, one with an upper frequency limit, the sample rate must exceed twice its highest frequency component.

Real systems also need practical margin.

Slide 4

Filtering must remove unwanted content before sampling

An anti-alias filter reduces unwanted content before sampling. Later filtering cannot generally separate an alias from a real signal at the same recorded frequency.

Slide 5

Extra bits cannot repair an inadequate sampling rate

More bits can reduce quantization steps, the gaps between recorded amplitude levels.

They cannot repair an insufficient sample rate or a signal clipped beyond the converter’s input range.

Sources

• Lecture 16: Sampling

  MIT OpenCourseWare

• Chapter 20: Analog to Digital Conversion

  Analog Devices

• Fundamentals of Sampled Data Systems

  Analog Devices