Lesson 3: Windowing and Time Domain Analysis

Objective:

This lesson helps you analyze time domain data and extract meaningful information from complex impulse responses through proper windowing techniques.

Windowing Fundamentals:

Windowing selectively analyzes portions of the impulse response, crucial for isolating specific acoustic phenomena.

Window Types in REW:

  1. Left Window: Before t=0 peak (direct sound and early reflections)
  2. Right Window: After t=0 peak (room decay and late reflections)
  3. Frequency-Dependent Window: Automatically adjusts width based on frequency (optimal for most measurements)

Shorter windows provide better high-frequency resolution but limit low-frequency content. Longer windows capture more low frequencies but lose high-frequency detail.

Frequency-dependent windows offer a compromise, with width inversely proportional to frequency (narrower at high frequencies).

Time Domain Analysis Techniques:

REW provides several tools for analyzing time-domain characteristics:

1. Energy-Time Curve (ETC)

Envelope of the impulse response showing reflection patterns and decay characteristics.

2. Cumulative Spectral Decay (CSD)

Specialized waterfall mode anchoring the right window, showing how energy decays at each frequency.

3. Burst Decay

Compares resonance Q factors across frequencies using a period axis.

Advanced Windowing Applications:

Specialized windowing can reveal specific acoustic phenomena:

1. Room Mode Analysis

  1. Use a long right window to capture modal decay
  2. Analyze the resulting frequency response for sharp peaks
  3. Examine CSD waterfall to determine Q factors of each mode

2. Speaker Directivity Patterns

  1. Use a short left window to isolate direct sound
  2. Measure at multiple positions to map radiation pattern
  3. Analyze phase coherency across measurements

Heisenberg's Uncertainty Principle in Audio:

The fundamental trade-off between time and frequency resolution applies to all windowing operations:

Practical Implications:

A 10 ms window cannot show valid data below about 100 Hz (1/length in Hz). For sub-100 Hz analysis, use longer windows despite reduced high-frequency resolution.

Quiz: Windowing Concepts

  1. What is the primary purpose of windowing in impulse response analysis?
  2. What does a frequency-dependent window do?
  3. What is the Energy-Time Curve (ETC) useful for?

Return to Lesson Index