generate_amplitudes_hearning_threshold.py

import pickle
import os
import numpy as np

from auditory_model.src.modules.gammatone_filterbank import GammatoneFilterbank
from auditory_model.src.runner import Config
from auditory_model.src.modules.preprocess import PreprocessLoadAudioLibrosa
from auditory_model.src.logs import LogLevel
from auditory_model.src.utils.helpers import center_frequencies_glasberg

# Prerequisites run plot_iso_standard.py before
wavelength = 4000
plot_channels = [2, 6, 7]
plot_channel = 7

center_freqs = center_frequencies_glasberg(20000, 20, 40)
extended_freqs = []

# If more than center freqs should be used
for i in range(len(center_freqs) - 1):
    start = center_freqs[i]
    end = center_freqs[i + 1]
    freqs_between = np.linspace(start, end, 5)[:-1]
    extended_freqs.extend(freqs_between)
extended_freqs.append(center_freqs[-1])

for phon in range(0, 91, 10):
    max_amplitudes = []
    with open(f"ohc_gain_fitting/pchip_phon_{phon}.pkl", "rb") as f:
        pchip_db_thresholds = pickle.load(f)

    print("Phon Level", phon)
    for i in extended_freqs:
        wavelength = int(i)
        db_level = pchip_db_thresholds(wavelength)

        config = Config(
            run_name=f"Generate_amplitudes",
            pure_tone_frequency=wavelength,
            save_plots=False,
            log_level=LogLevel.WARNINGS,
            pipeline=[
                PreprocessLoadAudioLibrosa(
                    take_only_first_x_ms=300, use_db_spl_scaling=True, db_level=db_level
                ),
                GammatoneFilterbank(
                    output_path="basilar_membrane_oscillation.png",
                    plot_channels=plot_channels,
                    plot_time_window=((1 / wavelength) * 4),
                ),
            ],
        )

        max_amplitudes.append([wavelength, np.max(config.run())])

    print(f"Final Result: phon {phon}")
    print(max_amplitudes)

    output_dir = "ohc_gain_fitting"
    os.makedirs(output_dir, exist_ok=True)
    output_path = os.path.join(output_dir, f"amplitudes_phon_{phon}.npy")
    np.save(output_path, max_amplitudes)
    print(f"Saved amplitudes to {output_path}")