ExtendedSource checked

Short-Time Fourier Transform

Analyzes frequency content over short moving windows of time.

Explore relationships Open constellationBrowse atlas

Quick Facts

Compact metadata for implementation and discovery.

Category: Signal Processing
Difficulty: Intermediate
Family: Time-Frequency Analysis
Tier: Extended
Status: Verified
Review: Source Checked
Sources: 1
Formulas: 1
Relationships: 4
Code example: Not ready
Visualization: Ready
Aliases: STFT

Overview

The Short-Time Fourier Transform applies a window to signal segments and computes a Fourier transform for each segment.

Core Concepts

signal samplestime-frequency windowfrequency-domain transformwindow weightingtime-frequency analysissample windowsignal transformnoise handling

Quick facts

Complexity

Runtime and memory behavior at a glance.

Best

O(w log w)

Average

O(m w log w)

Worst

O(m w log w)

Space

O(m w)

When to Use

Use Short-Time Fourier Transform when you need to split a signal into overlapping windows.
Use it for spectrograms with inputs like signal.
Use it when sampled data needs filtering, transformation, smoothing, or frequency analysis.
Use it when the Time-Frequency Analysis tradeoff is acceptable for the problem size.

When Not to Use

Avoid it when sample rate, window size, or boundary handling is not defined.
Avoid it when the required output is not time-frequency representation.

What It Does

Splits a signal into overlapping windows
Computes frequency content for each time slice

Use Cases

Spectrograms
Speech analysis
Music feature extraction

Inputs

Signal
Window function
Hop size

Outputs

Time-frequency representation

Key Formulas (1)

Windowed transform

X(window, k) = DFT(signal_segment · window_function)

X(window, k) = DFT(signal_segment * window_function)

Computes frequency content over short overlapping time windows.

Pseudocode

1.Choose a window length and hop size
2.Multiply each segment by a window function
3.Run an FFT on the segment
4.Store the spectrum for that time position
5.Advance by the hop size

Canonical Example

Input: samples = [1, 2, 3, 2, 1]
Output: transformed or filtered samples
Explanation: Short-Time Fourier Transform: The algorithm reads local or frequency structure from the sample sequence.

Implementation Notes

Keep the input contract explicit: Signal; Window function; Hop size.
Use this step as the implementation spine: Choose a window length and hop size.
Then preserve the next state change: Multiply each segment by a window function.
Test with simple signals such as constants, impulses, and single-frequency waves.

Common Pitfalls

Short windows improve time detail; long windows improve frequency detail.
Interpreting transformed values without accounting for scale, phase, or edge effects.
Returning time-frequency representation without handling empty, invalid, or boundary inputs.

Interactive Mechanism Walkthrough

Iterative Convergence & State Trajectory

Show a waveform becoming a time-frequency heat map.

Step 1 of 5: Choose a window length and hop size

spectrogram

Notes

Short-Time Fourier Transform operates in O(m w log w) time and O(m w) space within the Time-Frequency Analysis family. Key governing relation: X(window, k) = DFT(signal_segment * window_function). Often compared with Fast Fourier Transform and Windowing.