AERO61021 Chap.6 Sampling Rate and Aliasing
Sampling Rate and Aliasing
Sampling Rate and Aliasing covers the decision that most often ruins a dynamic dataset before analysis even begins: how fast to sample. A digital system records a continuous signal as discrete points, and if those points are too sparse they describe a signal with entirely different frequency content from the real one.
The chapter sets out the Nyquist and Shannon results in plain terms, leading to the working rule that a sample rate can correctly resolve frequencies only up to half its own value, so the rate must be at least twice the highest frequency present.
It then walks through the distinct regimes the notes describe, from samples slower than the signal, through one sample per cycle, to rates between one and two times the signal frequency and finally correct sampling, and shows why recovering amplitude reliably needs a rate around ten times the signal frequency.
Practical constraints follow: rates shared across channels on multifunction hardware, camera frame rates, and spatial aliasing that appears in images as Moiré patterns. The emphasis throughout is on estimating the highest frequency before testing, because an aliased record cannot be repaired afterwards.
What this chapter covers
- 01
What aliasing does to a sampled record
- 02
The Nyquist and Shannon sampling results
- 03
The criterion that sample rate must be at least twice the signal frequency
- 04
Positive, destructive and negative aliasing regimes
- 05
Oversampling to recover peak-to-peak amplitude
- 06
Per-channel rates on shared acquisition hardware
- 07
Moiré patterns as spatial aliasing in images
Worked example · free
Classify an undersampled plan and choose a safe sample rate
- 2Form the ratio f/N = 200/150 = 1.33, which lies between 1 and 2, the regime that produces a false slower wave in the record.
- 1Apply f ≥ 2N: the minimum sample rate is 2 × 150 = 300 Hz.
- 2For amplitude, sample around ten times the signal frequency: 10 × 150 = 1500 Hz, then confirm the hardware delivers this per channel.
Key terms
- Aliasing
- Aliasing is the effect by which under-sampled data appear to contain frequencies that differ from those of the real signal.
- Sampling criterion
- The sampling criterion is the rule that the sample rate must be at least twice the highest frequency in the signal.
- Destructive aliasing
- Destructive aliasing is the case where sampling at the signal frequency catches the same phase each time and shows a constant value.
- Moiré pattern
- A Moiré pattern is the false banding produced when a fine repeating structure is imaged with too few pixels.
Sampling Rate and Aliasing FAQ
Is sampling at exactly twice the frequency enough?
It is the boundary for identifying frequency, not a comfortable working point. At or just above twice the signal frequency the record can still misrepresent the waveform shape, and the peak-to-peak size is unreliable. The module notes illustrate a far higher rate, around ten samples per cycle, when amplitude must be read correctly from the data.
Can I filter or process an aliased record to fix it?
Once samples have been taken too slowly, the false frequency is indistinguishable from a real one in the data, so no later processing can separate them. The remedy has to come before the test: estimate the highest frequency the experiment can produce, then choose a sample rate and hardware that respect the criterion with margin.
Why can a camera alias as well as a data logger?
A camera samples in time through its frame rate and in space through its pixels. A rotating or shedding feature filmed below twice its frequency appears to move at a false speed, and a fine pattern recorded with too few pixels shows Moiré banding. Both are the same under-sampling problem in different dimensions.
Assessment move
Learn this chapter by drawing rather than memorising. Take a sine wave on squared paper and mark samples at ratios of 0.9, 1, 1.25, 2.5 and 10 samples per cycle; join the dots each time and name what you see. Doing this once by hand fixes the regimes far better than reading their names, and it shows why a rate that technically satisfies the criterion can still give a misleading shape.
Next, practise the planning sequence for any experiment: estimate the highest frequency the physics can produce, apply the factor of two as a floor, decide whether amplitude matters and raise the rate towards ten times the frequency if it does, then check the per-channel rate your hardware actually provides with every channel enabled.
Repeat the sequence for a vibration test, a pressure fluctuation and a camera recording so it becomes a habit rather than a formula. Use the LabVIEW aliasing example and the sampling demonstration offered in the module to watch the effect live, changing the rate while the signal stays fixed and predicting each result before you run it.
When you write up laboratory or group project work, state the expected maximum frequency, the chosen rate and the ratio between them in one sentence, because a reader cannot judge your data without it.
Finally, connect this chapter to sensor response: a slow sensor already filters fast content, so the frequency you must protect is the one that reaches the acquisition card, and explaining that interaction clearly is a good test of real understanding.
Close each study session by explaining one aliasing case aloud in under a minute: the signal frequency, the sample rate, the ratio, what the record would show and the rate you would choose instead. If you can do that for a destructive case, a negative case and a camera case without notes, the chapter is secure, and you will be able to defend sampling choices in any report or project discussion.