Multisine Signals: Fundamentals and Practical Applications

Multisine vs. Single-Tone Testing: Benefits and Trade-offs

Overview

Testing signals are central to characterizing linear and nonlinear systems. Two common approaches are single-tone (pure sine) testing and multisine testing (superposition of multiple sinusoids). Each method has strengths and limitations depending on goals such as frequency resolution, measurement speed, distortion detection, and signal-to-noise considerations. This article compares both approaches and offers practical guidance for choosing and applying them.

What they are

• Single-tone testing: Excite the system with one sinusoidal frequency at a time, sweep across frequencies to build a frequency response.
• Multisine testing: Excite the system with many sinusoids simultaneously (possibly with carefully chosen amplitudes, phases, and frequency spacing).

Key benefits — Multisine

• Faster frequency coverage: Measures many frequencies in a single acquisition, reducing test time.
• Higher effective SNR: Energy concentrated on test frequencies allows longer integration and averaging, improving SNR for those bins.
• Nonlinearity detection: Harmonics and intermodulation products appear outside the excited bins and can be measured directly if not excited, enabling clear separation of linear response and nonlinear distortion.
• Flexible design: Can shape amplitude across bands, include random-phase or optimized-phase multisines to control crest factor and measurement dynamics.
• Concurrent system identification: Good for time-varying systems where sweeping would misrepresent dynamics.

Key benefits — Single-tone

• Simplicity and interpretability: Straightforward to implement and analyze; phase and amplitude at a single frequency are easy to interpret.
• Low crest factor: Single sine has minimal crest factor, reducing amplifier clipping and peak-power issues.
• Fine control at each frequency: Allows targeted, controlled excitation and settling at each frequency (useful for very narrowband nonlinear calibration).
• Classic techniques compatibility: Easily used with lock-in amplifiers and narrowband measurement setups that excel at extracting small signals in noise.

Trade-offs and limitations

• Test duration: Single-tone sweeps can be slow—especially at high resolution—while multisines shorten total test time.
• Leakage and resolution: Multisine requires windowing and careful frequency bin alignment to avoid spectral leakage; single-tone can use longer dwell per frequency to minimize uncertainty.
• Crest factor and dynamic range: Multisines typically have higher crest factor than single sine; this raises peak demands on signal sources and power stages and can provoke nonlinear behavior if peaks clip. Phase optimization can mitigate but not eliminate this.
• Complexity: Multisine design (frequency selection, phase optimization, amplitude shaping) and processing (bin alignment, harmonic analysis) are more complex than single-tone methods.
• Nonlinearity interaction: While multisines reveal nonlinear products efficiently, their simultaneous multiple excitations can produce complex intermodulation, making root-cause analysis harder than isolated single-tone harmonic testing.
• Transient and time-varying systems: Single-tone with steady-state per frequency can give clearer per-frequency steady-state measures; multisines are better when you must capture system behavior quickly or when the system changes over time.

Practical guidelines for choice

• Use multisine when:

  • You need rapid broadband measurements (e.g., vibration testing, acoustic rooms, electrical networks).
  • You want to quantify or separate nonlinear distortion from linear response efficiently.
  • The system is time-varying or testing time is limited.
  • You can manage crest factor and ensure your hardware handles peaks.

• Use single-tone when:

  • Highest peak-power efficiency and minimal crest factor are required.
  • You need very precise, narrowband characterization or stepwise tuning.
  • Simplicity is preferred or measurement hardware uses narrowband synchronous detection.
  • Root-cause isolation of a specific harmonic under controlled excitation is desired.

Design and measurement tips

• For multisines:
  • Align excited frequencies to FFT bins and use integer-multiple record lengths to avoid leakage.
  • Optimize phases to reduce crest factor (e.g., Schroeder phases) if hardware peak limits are tight.
  • Reserve unexcited bins to measure harmonics/intermodulation directly.
  • Use averaging across repeated multisine periods to improve SNR while preserving nonlinear signatures.
• For single-tone sweeps:
  • Use sufficient settling time at each frequency for steady-state measurement.
  • Employ lock-in or synchronous detection to maximize sensitivity.
  • Consider finer frequency steps in regions of rapid change (resonances).

Example comparison (short)

• Measurement time: multisine much faster (one-shot) vs. single-tone long sweeps.
• Nonlinear detection: multisine reveals intermodulation across bands; single-tone shows harmonics from controlled amplitude.
• Hardware stress: multisine higher peaks → more risk of clipping.

Conclusion

Multisine and single-tone testing each have clear roles. Multisine excels at fast, broadband, and nonlinear-aware testing but requires careful design to control crest factor and spectral leakage. Single-tone testing remains the preferred method for simple, low-peak, precise narrowband characterization and when analysis simplicity is paramount. Choose based on test duration limits, nonlinearity interest, hardware constraints, and desired analysis simplicity.