Generate stimuli with gaps

This demonstrates how to generate various stimuli with gaps (e.g., for measuring temporal processing).

import matplotlib.pyplot as plt
import numpy as np

from psiaudio import calibration
from psiaudio import stim
from psiaudio import util


# Creating calibration
# --------------------
# This ensures that the resulting waveform is in units of Pascals.
flat_cal = calibration.FlatCalibration.from_spl(94)
fs = 100e3

Gap detection based on tone pips

Commonly used in human studies.

waveform = stim.gap(
    fs=fs,
    fc=4e3,
    octaves=0,
    gap=2e-3,
    durations=[4e-3, 4e-3],
    rise_time=2e-3,
    level=80,
    calibration=flat_cal,
)

t = np.arange(len(waveform)) / fs
figure, axes = plt.subplots(1, 2, figsize=(8, 4), constrained_layout=True)
axes[0].plot(t, waveform)
axes[0].set_xlabel('Time (s)')
axes[1].set_xlabel('Amplitude (Pa)')
axes[1].specgram(waveform, Fs=fs);
axes[1].axis(ymin=200, ymax=20e3)
axes[1].set_yscale('octave', octaves=2)
axes[0].set_xlabel('Time (s)')
axes[1].set_xlabel('Frequency (kHz)')
plt.show()

Gap detection based on bandlimited noise

Commonly used in animal studies.

waveform = stim.gap(
    fs=fs,
    fc=4e3,
    octaves=0.5,
    gap=8e-3,
    durations=[500e-3, 500e-3],
    rise_time=2e-3,
    level=80,
    calibration=flat_cal,
)

t = np.arange(len(waveform)) / fs
figure, axes = plt.subplots(1, 2, figsize=(8, 4), constrained_layout=True)
axes[0].plot(t, waveform)
axes[0].set_xlabel('Time (s)')
axes[1].set_xlabel('Amplitude (Pa)')
axes[1].specgram(waveform, Fs=fs);
axes[1].axis(ymin=200, ymax=20e3)
axes[1].set_yscale('octave', octaves=2)
axes[0].set_xlabel('Time (s)')
axes[1].set_xlabel('Frequency (kHz)')
plt.show()

/opt/hostedtoolcache/Python/3.13.13/x64/lib/python3.13/site-packages/matplotlib/axes/_axes.py:8283: RuntimeWarning: divide by zero encountered in log10
  Z = 10. * np.log10(spec)

Multiple gaps based on tone pips

Commonly used in animal studies to help train the animal before switching to a single gap. This embeds three gaps in the stimulus.

waveform = stim.gap(
    fs=fs,
    fc=4e3,
    octaves=0,
    gap=8e-3,
    durations=[32e-3, 32e-3, 32e-3, 32e-3],
    rise_time=2e-3,
    level=80,
    calibration=flat_cal,
)

t = np.arange(len(waveform)) / fs
figure, axes = plt.subplots(1, 2, figsize=(8, 4), constrained_layout=True)
axes[0].plot(t, waveform)
axes[0].set_xlabel('Time (s)')
axes[1].set_xlabel('Amplitude (Pa)')
axes[1].specgram(waveform, Fs=fs);
axes[1].axis(ymin=200, ymax=20e3)
axes[1].set_yscale('octave', octaves=2)
axes[0].set_xlabel('Time (s)')
axes[1].set_xlabel('Frequency (kHz)')
plt.show()