Skip to content
Technical Documentation

Spectrogram

Real-time scrolling time-frequency display with multiple analysis backends, GPU acceleration, and configurable color mapping

How It Works

The spectrogram displays a scrolling history of spectral data as a 2D heatmap. Each vertical column represents one moment in time, and each horizontal row represents a frequency. Color intensity maps to signal level in dB.

History Depth

Circular buffer, scrolling display

150 time samples
Frequency Axis

20 Hz – 22 kHz

Logarithmic
Color Scale

Purple (low) → Red (high), adjustable range

HSV colormap
Rendering

Filament renderer with CPU fallback

GPU-accelerated
Compact Mode

Hidden labels, hover-reveal controls

Sidebar-ready

Analysis Backends

The spectrogram can use any of AudioMap's analysis engines. Each produces different time-frequency characteristics.

FFT

128 – 131,072

Traditional STFT. Fixed resolution, uniform frequency bins. Good for general use.

MRA

Stage: 128 – 2,048

Multi-Resolution Analysis. Fine LF resolution with fast HF response. Latency-compensated per display point.

CWT

Bins: 128 – 2,048

Continuous Wavelet Transform. Smooth, artifact-free display. No octave boundary artifacts.

CQT

BPO: 12 – 48

Constant-Q Transform. Natural log-frequency resolution. Configurable bins per octave.

Superlet

Order: 1 – 10

Super-resolution transform. Highest simultaneous time-frequency detail. CPU intensive.

Reassignment

Threshold: -40 – -10 dB

STFT sharpening overlay. Available for FFT, MRA, CQT modes. Blend control 0–100%.

Display Configuration

Color Scale

Two-value slider adjusts the dB range mapped to color. Narrower range increases contrast for specific level regions.

HSV colormap: purple (minimum) through green and yellow to red (maximum)

Temporal Smoothing

Available for CWT, CQT, and Superlet modes. Reduces frame-to-frame noise in the spectrogram display.

FIFO or IIR modes with configurable depth

MRA Latency Compensation

When using MRA, per-display-point FIFO delay lines align lower octave stages with higher ones for a coherent display.

Automatic — no user configuration required