Short-Time Fourier Transform

Graduate Depth 89 in the knowledge graph I know this Set as goal
Unlocks 1 downstream topic
time-frequency stft spectral-analysis windows

Core Idea

The Short-Time Fourier Transform (STFT) computes Fourier transform of overlapping windowed segments to provide time-frequency representation: STFT(t,ω) = ∫ x(τ)·w(τ–t)·e^(–jωτ) dτ. It trades time and frequency resolution: narrower windows improve time localization but worsen frequency resolution. Spectrograms visualize STFT magnitude showing frequency evolution over time.

Explainer

The standard Fourier transform is like asking, "what frequencies are present in this signal?" — and getting a complete answer, but with no information about *when* those frequencies occur. For a piece of music, the ordinary Fourier transform tells you every note ever played, but nothing about their order or timing. The Short-Time Fourier Transform (STFT) solves this by asking a more local question: what frequencies are present *right now*, in this short window of time?

The idea is simple: multiply the signal by a window function — a smooth, localized pulse like a Gaussian or Hann window — that is zero everywhere except near some moment t. Then take the Fourier transform of what remains. This gives the frequency content of the signal near time t. By sliding the window across the entire signal and repeating, you get a two-dimensional map of frequency vs. time. This map is the STFT, and its magnitude squared is the spectrogram — the colored time-frequency plots you see in audio analysis and speech processing.

The catch is the time-frequency uncertainty principle (analogous to the Heisenberg uncertainty principle in quantum mechanics): you cannot have arbitrarily sharp resolution in both time and frequency simultaneously. A narrow window gives excellent time localization — you know precisely *when* a frequency appears — but the short duration means the Fourier transform sees very few oscillations, leading to smeared frequency content. A wide window gives sharp frequency peaks (many oscillations to count) but blurs together events that happen at different times. Formally, the product of time resolution Δt and frequency resolution Δω is bounded below: Δt · Δω ≥ 1/2.

This resolution tradeoff is the fundamental limitation of the STFT and motivates its successor, the wavelet transform. Unlike the STFT — where every frequency is analyzed with the same fixed window width — wavelets use a window that automatically shrinks at high frequencies and widens at low ones. This provides constant *relative* resolution (high frequencies resolved in time, low frequencies resolved in pitch), which is why wavelets are preferred for signals like speech and ECG where low-frequency content evolves slowly and high-frequency transients are brief. Understanding the STFT's fixed-resolution limitation is the conceptual bridge to that more flexible framework.

Practice Questions 5 questions

Prerequisite Chain

Counting to 10Counting to 20Understanding ZeroThe Number ZeroCounting to FiveOne-to-One CorrespondenceCombining Small Groups Within 5Addition Within 10Addition Within 20Two-Digit Addition Without RegroupingTwo-Digit Addition with RegroupingAddition Within 100Repeated Addition as MultiplicationMultiplication Facts Within 100Division as Equal SharingDivision as Grouping (Measurement Division)Division: Grouping (Repeated Subtraction) ModelDivision: Fair Sharing ModelDivision as Equal SharingDivision as GroupingBasic Division FactsDivision Facts Within 100Two-Digit by One-Digit DivisionDivision with RemaindersRemainders and Quotients in DivisionDivision Word ProblemsIntroduction to Long DivisionFactors and MultiplesPrime and Composite NumbersEquivalent FractionsRelating Fractions and DecimalsDecimal Place ValueReading and Writing DecimalsComparing and Ordering DecimalsAdding and Subtracting DecimalsMultiplying DecimalsDividing DecimalsDividing FractionsMixed Number ArithmeticOrder of OperationsInteger Order of OperationsVariable ExpressionsCombining Like TermsOne-Step EquationsTwo-Step EquationsSolving Multi-Step EquationsEquations with Variables on Both SidesAngle Pairs: Complementary, Supplementary, and VerticalParallel Lines and TransversalsCorresponding AnglesAlternate Interior AnglesTriangle Angle Sum TheoremExterior Angle TheoremTriangle Inequality TheoremSimilar Triangles: AA SimilaritySimilar Triangles: SSS and SAS SimilarityProportions in Similar TrianglesRight Triangle Trigonometry IntroductionTrigonometric Ratios ReviewRadian MeasureConverting Between Degrees and RadiansThe Unit CircleGraphing Sine and CosineGraphing Tangent and Reciprocal Trigonometric FunctionsDerivatives of Trigonometric FunctionsAntiderivativesIndefinite IntegralsBasic Integration RulesRiemann SumsDefinite Integral DefinitionFundamental Theorem of Calculus Part 1Fundamental Theorem of Calculus Part 2U-SubstitutionIntegration by PartsSeparable Differential EquationsIntegrating Factor Method for First-Order Linear ODEsFirst-Order Linear Ordinary Differential EquationsSecond-Order Linear Homogeneous Differential EquationsCharacteristic Equation Method for Linear ODEsComplex Roots and Oscillatory SolutionsSpring-Mass Systems and Mechanical VibrationsResonance and Damping in Forced VibrationsRLC Circuit Applications of Differential EquationsIntroduction to Differential EquationsLaplace Transform: Fundamentals and PropertiesZ-Transform: Fundamentals for Discrete-Time SignalsDiscrete-Time Fourier Transform (DTFT)Discrete Fourier Transform (DFT) and Fast Fourier Transform (FFT) AlgorithmsWindow Functions and Spectral LeakageShort-Time Fourier Transform

Longest path: 90 steps · 396 total prerequisite topics

Prerequisites (2)

Leads To (1)