EMG Signal Processing Guide
Surface electromyography (sEMG) measures electrical activity produced by skeletal muscles. The signal is an interference pattern from many motor units firing beneath the skin, so it often appears as bursts of noise-like activity rather than a repeating waveform.
A professional EMG workflow should remove movement artifact, preserve muscle activation energy, make bursts easy to inspect and provide clear amplitude metrics such as RMS and ROI statistics.
What Makes EMG Different
ECG and EEG often focus on recognizable rhythms or waveforms. EMG focuses on changes in amplitude and frequency content during muscle contraction. A useful EMG page should therefore explain preprocessing, activation bursts, ROI metrics and frequency-domain quality checks in plain language.
EMG Frequency Range and Sampling Rate
Most usable energy in surface EMG lies roughly between 20 Hz and 500 Hz, with strong activity often appearing around 50 Hz to 150 Hz. Because the frequency range is broad, EMG recordings require a higher sampling rate than many ECG or slow EEG workflows.
If you want to inspect content up to 500 Hz, the sampling frequency should be at least 1000 Hz, and 2000 Hz is often preferred for cleaner analysis headroom.
Movement Artifacts in EMG
Movement artifact is the most common reason raw EMG looks unusable. Skin stretch, cable sway and electrode motion create large low-frequency swings, usually below 20 Hz, that can hide the actual muscle signal.
Removing that slow movement component before calculating amplitude metrics is essential. Otherwise, RMS, mean amplitude or ROI statistics can reflect electrode motion instead of muscle activation.
Recommended EMG Filter Settings
EMG preprocessing typically uses a stronger high-pass filter than ECG or EEG. The goal is to remove low-frequency motion while preserving the higher-frequency muscle activity that carries useful activation information.
Standard EMG filter preset
- High-pass filter: 20 Hz to 30 Hz, order 4.
Removes movement artifact and baseline drift while keeping most surface EMG power. - Notch filter: 50 Hz or 60 Hz.
Reduces powerline interference that can otherwise look like sustained muscle activity. - Low-pass filter: 400 Hz to 500 Hz.
Limits high-frequency noise when the recording was sampled fast enough to support that range.
How to Process EMG Data Online
BioSignal Workstation helps you move from raw EMG files to filtered bursts and exportable measurements quickly.
- Upload your EMG file: open CSV, TXT, XLS or XLSX data in the BioSignal Workstation analyzer.
- Apply the EMG preset: start with a high-pass filter around 20 Hz and enable the notch filter for local mains noise.
- Inspect bursts: use the chart to confirm that activation periods are centered and separated from quiet baseline segments.
- Select a region of interest: highlight a contraction burst with the ROI tool to calculate amplitude metrics.
- Export ROI statistics: save RMS, minimum, maximum, mean and standard deviation values for reporting or downstream analysis.
Validate EMG with Welch PSD
A healthy EMG contraction often shows a broad PSD curve rather than a single narrow peak. If the PSD has excessive power near 0 Hz, the high-pass filter may be too low or disabled. If there is a sharp 50/60 Hz spike, powerline noise is still present and the notch filter should be reviewed.
EMG Analyzer FAQ
Can I analyze EMG CSV files online?
Yes. The EMG analyzer accepts CSV, TXT, XLS and XLSX files and processes them locally in your browser.
What is the best starting filter for surface EMG?
Start with a 20 Hz to 30 Hz high-pass filter, a 50 Hz or 60 Hz notch filter and a low-pass filter around 400 Hz to 500 Hz if your sampling rate supports it.
How should I measure muscle activation?
After filtering, select a region around a contraction burst and review RMS plus ROI statistics. Compare similar task windows rather than mixing quiet baseline and active contraction periods.