Skip to content

Spectrum and Time Domain Analyzers

Jump to bottom
youtube-at-vach edited this page Jul 23, 2026 · 1 revision

Spectrum and Time-Domain Analyzers

Relevant source files

The following files were used as context for generating this wiki page:

This page covers the implementation and data flow of the real-time analysis widgets in MeasureLab, focusing on frequency-domain (Spectrum, Spectrogram) and time-domain (Oscilloscope, Transient Analyzer, Goniometer) visualization tools.

Spectrum Analyzer

The SpectrumAnalyzer module provides real-time frequency analysis using Fast Fourier Transforms (FFT). It supports various estimation modes including Power Spectral Density (PSD) and cross-spectrum analysis.

Technical Implementation

The analyzer utilizes a dual-mode buffering strategy based on the FFT size src/gui/widgets/spectrum_analyzer.py:34-36.

  • Rolling Mode: For standard buffer sizes, it uses a high-performance ring buffer to maintain a continuous display src/gui/widgets/spectrum_analyzer.py:154-173.
  • Snapshot Mode: When buffer_size exceeds LARGE_BUFFER_THRESHOLD (600,000 samples), the system switches to a linear capture mode to ensure phase coherency for ultra-high resolution FFTs src/gui/widgets/spectrum_analyzer.py:138-152.

Signal Processing Chain

  1. Windowing: Applies Hanning, Hamming, or Blackman windows via get_cached_window src/gui/widgets/spectrum_analyzer.py:20.
  2. FFT Execution: Offloads transforms to the FFTManager, which utilizes pyfftw for optimized execution src/core/fft_manager.py:22.
  3. Multitaper Estimation: Optionally uses Discrete Prolate Spheroidal Sequences (DPSS) to reduce variance in noise floor estimation src/gui/widgets/spectrum_analyzer.py:55.
  4. Smoothing: Implements fractional-octave smoothing (1/3, 1/6, etc.) using a cached frequency-bin mapping src/gui/widgets/spectrum_analyzer.py:66-69.

Data Flow Diagram

The following diagram illustrates the flow from the AudioEngine callback to the SpectrumAnalyzerWidget.

Spectrum Analysis Data Flow

graph TD
    subgraph "Audio Thread"
        AE["AudioEngine Callback"] -- "indata (Stereo)" --> Q["audio_queue"]
    end

    subgraph "GUI Thread (Timer)"
        Q -- "get_nowait()" --> SA["SpectrumAnalyzer.process_queue()"]
        SA -- "RingBuffer / Snapshot" --> IB["input_data buffer"]
        IB -- "FFT Processing" --> FFT["FFTManager.rfft()"]
        FFT -- "Magnitude / PSD" --> SM["Octave Smoothing / Averaging"]
        SM -- "Plot Data" --> SAW["SpectrumAnalyzerWidget.update_plot()"]
    end
Loading

Sources: src/gui/widgets/spectrum_analyzer.py:114-130, src/gui/widgets/spectrum_analyzer.py:131-174.

Spectrogram

The Spectrogram module visualizes frequency components over time (waterfall display). It uses a dedicated worker thread to prevent UI blocking during heavy FFT computations.

Implementation Details

  • Worker Pattern: Uses SpectrogramWorker (inheriting QRunnable) to perform FFT, magnitude calculation, and dB conversion in the background src/gui/widgets/spectrogram.py:33-41.
  • In-place Optimization: To minimize memory bandwidth, dB conversion and normalization are performed using in-place NumPy operations src/gui/widgets/spectrogram.py:64-73.
  • Variable Speed: Supports "Meteor" and "Slow" modes. In these modes, the system uses a "Max Hold" mechanism to ensure transient peaks are not lost during downsampling docs/widgets/spectrogram.en.md:71.

Spectrogram Class Structure

classDiagram
    class Spectrogram {
        +fft_size: int
        +spectrogram_buffer: ndarray
        +add_spectrum(mag_db)
        +process_audio()
    }
    class SpectrogramWorker {
        +raw_data: ndarray
        +run()
    }
    class SpectrogramWidget {
        +img_item: ImageItem
        +update_image()
    }
    Spectrogram --> SpectrogramWorker : dispatches
    SpectrogramWorker ..> SpectrogramWidget : signals result
Loading

Sources: src/gui/widgets/spectrogram.py:29-77, src/gui/widgets/spectrogram.py:78-108.

Oscilloscope and Time-Domain Tools

Oscilloscope

The Oscilloscope provides a real-time waveform monitor with advanced triggering and persistence modes.

  • Triggering: Implements "Auto", "Normal", and "Single" modes with adjustable slope and level src/gui/widgets/oscilloscope.py:69-71.
  • Persistence: Uses an optimized histogram accumulation method (_accumulate_heatmap) to render phosphor-like decay effects src/gui/widgets/oscilloscope.py:117-151.
  • Performance: Utilizes a RingBuffer instead of a standard queue to avoid memory allocation in the audio callback src/gui/widgets/oscilloscope.py:101-104.

Goniometer (Stereo Phase)

The Goniometer visualizes the phase relationship between two channels (Lissajous patterns).

  • Processing: Captures stereo pairs and plots them as (L-R, L+R) coordinates to represent the stereo field src/gui/widgets/goniometer.py.
  • Persistence: Similar to the Oscilloscope, it uses a decay buffer to maintain trace visibility over time.

Transient Analyzer

The TransientAnalyzer is specialized for observing impulse responses and ringing.

  • Ringing Analysis: Detects pre-ringing and post-ringing artifacts in digital filters tests/logic_verification/gui/test_transient_ringing_analysis.py.
  • Windowing: Allows gating of transients to isolate specific reflections or filter taps.

Technical Summary Table

Feature Entity Name Implementation Detail
FFT Engine FFTManager Uses pyfftw with plan caching
Buffering RingBuffer Lock-free thread-safe circular buffer
Spectrum Modes SpectrumAnalyzer Magnitude, PSD, Cross-Spectrum
Triggering Oscilloscope Software-based zero-crossing/level detection
Smoothing SpectrumAnalyzer Fractional-octave (1/1 to 1/96)
Persistence _accumulate_heatmap 2D histogram accumulation for heatmaps

Sources: src/gui/widgets/spectrum_analyzer.py:34-71, src/gui/widgets/oscilloscope.py:49-106, src/core/fft_manager.py:1-25.

MeasureLab Wiki

Home

Clone this wiki locally