Dolphain Studio

Python Code

import dolphain
from dolphain.plotting import plot_overview
import matplotlib.pyplot as plt

# Set dark theme for plots to match UI
plt.style.use('dark_background')
plt.rcParams['figure.facecolor'] = '#050713' # Match output container bg
plt.rcParams['axes.facecolor'] = '#050713'
plt.rcParams['grid.color'] = '#333333'
plt.rcParams['text.color'] = '#e0e0e0'
plt.rcParams['axes.labelcolor'] = '#e0e0e0'
plt.rcParams['xtick.color'] = '#e0e0e0'
plt.rcParams['ytick.color'] = '#e0e0e0'

# Load the selected file
# Note: In this browser environment, we fetch directly
print(f"Loading {selected_filename}...")
ears = dolphain.EARS(filename=selected_filename)

print(f"Loaded {ears.duration}s of audio")
print(f"Sample rate: {ears.fs} Hz")

# Create visualization
plt.close('all') # Clear previous plots
plt.figure(figsize=(10, 7)) # Widescreen aspect ratio
# Use smaller FFT window and overlap to reduce memory usage in browser
# Also downsample the data for plotting if it's too large
import numpy as np
from scipy import signal

# Custom plotting to avoid memory issues with full resolution
def safe_plot_overview(ears_data):
    data = ears_data["data"]
    fs = ears_data["fs"]
    
    # Downsample for waveform plot (10x)
    step = 10
    t = np.arange(0, len(data), step) / fs
    
    plt.subplot(2, 1, 1)
    # Use cyan color for waveform
    plt.plot(t, data[::step], linewidth=0.5, color='#00d4ff')
    plt.title(f"Waveform (Downsampled {step}x)", color='#00d4ff')
    plt.xlabel("Time [s]")
    plt.ylabel("Amplitude")
    plt.grid(True, alpha=0.2)
    
    # Compute spectrogram with coarser resolution
    plt.subplot(2, 1, 2)
    
    # Downsample data for spectrogram to save memory
    # High sample rates create massive spectrograms that crash the browser
    target_fs = 24000 # Limit to 24kHz (12kHz Nyquist) for overview
    if fs > target_fs:
        q = int(fs / target_fs)
        data_spec = data[::q]
        fs_spec = fs / q
    else:
        data_spec = data
        fs_spec = fs
        
    f, t_spec, Sxx = signal.spectrogram(data_spec, fs=fs_spec, nperseg=256, noverlap=128)
    Sxx_dB = 10 * np.log10(np.abs(Sxx) + 1e-10)
    
    # Use imshow instead of pcolormesh for much better memory performance
    # Use 'inferno' or 'magma' for better contrast on dark bg, or stick with nipy_spectral
    plt.imshow(Sxx_dB, aspect='auto', origin='lower', cmap='inferno', 
               extent=[t_spec[0], t_spec[-1], f[0], f[-1]])
    plt.ylabel('Frequency [Hz]')
    plt.xlabel('Time [s]')
    plt.title('Spectrogram (Coarse)', color='#ff00ff') # Magenta title
    plt.colorbar(label='dB')
    
    plt.tight_layout()

safe_plot_overview(ears.metadata)
# plt.show() is not needed as we capture the figure automatically

Output Console

Dolphain Studio Help

Getting Started

Available Libraries

Key Dolphain Functions