transform_wave_with_score

transform_wave_with_score

Location: variations/transformations.py

---

Description

transform_wave_with_score replaces segments of a waveform with newly generated signal portions based on a score-controlled probability. This function is useful for simulating signal corruption, feature substitution, or change in activity through different measurements.

---

Notes

• The function generates a replacement signal using generate_signal().
• The number and size of replacement segments depend on the score value.
• Substitution is avoided within predefined interrupt regions, ensuring that important waveform segments remain intact.

---

Parameters

• original_wave (numpy.ndarray):
  The input waveform.

• score (float):
  A factor controlling how much of the waveform is transformed.

• Higher values → More segments replaced.

• Lower values → Less transformation.

• t (numpy.ndarray):
  Time vector for generating the replacement signal.

• n_sinusoids (int):
  Number of sinusoids in the generated replacement signal.

• amplitude_range (tuple of float):
  Amplitude range for the replacement signal.

• base_frequency_range (tuple of float):
  Frequency range for the replacement signal.

• interrupt_params (list of dict):
  List of dictionaries defining interrupt regions that should not be replaced.
  Example: [{'start_idx': 100, 'duration_idx': 50}].

---

Returns

• transformed_wave (numpy.ndarray):
  The modified waveform with score-based segment replacement.

---

Usage Example

transformed_wave = svg.transform_wave_with_score(
    wave, 0.5, t, n_sinusoids=10, 
    amplitude_range=(0,1), base_frequency_range=(70, 75), 
    interrupt_params=[{'start_idx': 200, 'duration_idx': 50}]
)