Noise Flashcards
What are the two types of noise?
Systematic and Random
Describe systematic noise
Noise or error is consistent
Correlated with measurement process
Easily Removed
Describe random noise
Noise or error is random
Not correlated with measurement process
Not as easy to remove
What is signal?
A signal is a time varying quantity that conveys information
Nyquist Sampling Theory
The sampling rate/ frequency should be at least twice the rate of the fastest data expected
What is noise?
Noise is the unwanted part of sampled data; it masks the true value in which the experimenter is interested.
Three sources of Kinematic noise?
- Skin/Marker Movement
- Marker visibility
- Electrical Interference
Four ways to remove noise?
- Curve Fitting
– Derive a mathematical function to represent the data - Digital Filtering
– A weighted moving average where certain signal frequencies are attenuated - Moving Average
– Averages signal across a 3 or 5 point average. Cannot distinguish noise and signal. - Frequency Domain Techniques
– Transforms data into frequency domain to eliminate unwanted frequency elements and then reconstructs the data
Butterworth Filter (Low pass)
Low frequencies are allowed to pass without change, high frequencies are reduced
Most Commonly used
Butterworth Filter (High pass)
High frequencies are allowed to pass without change, low frequencies are reduced.
Butterworth Filter (Band Pass)
Frequencies inside a defined range are allowed to pass without change. Everything else is reduced.
Often used in EMG
Butterworth Filter (Cut off frequency)
Defines what is signal and what is noise
Identifies the border for low/high pass
Identifies the range for band pass
Three sources of Kinetic noise?
- Electrical Interference
- Amplifier Noise
- Stability of platform
