Spaces:
Sleeping
Sleeping
| import gradio as gr | |
| import librosa | |
| import numpy as np | |
| import matplotlib.pyplot as plt | |
| import tempfile | |
| def extract_audio_features(audio_file): | |
| y, sr = librosa.load(audio_file, sr=None) | |
| mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=13) | |
| pitches, magnitudes = librosa.piptrack(y=y, sr=sr) | |
| pitches = pitches[(magnitudes > np.median(magnitudes)) & (pitches > 0)] | |
| energy = librosa.feature.rms(y=y)[0] | |
| return mfccs, pitches, energy | |
| def analyze_voice_stress(audio_file): | |
| if not audio_file: | |
| return "No audio file provided.", None | |
| try: | |
| mfccs, pitches, energy = extract_audio_features(audio_file) | |
| stress_level = (np.var(mfccs) + (np.var(pitches) if len(pitches) > 0 else 0) + np.var(energy)) / 3 | |
| normalized_stress = min(100, (stress_level / 1000) * 100) | |
| fig, axs = plt.subplots(3, 1, figsize=(10, 12)) | |
| plots = [ | |
| (mfccs, 'MFCCs', 'MFCC Coefficient', 'imshow', {'aspect': 'auto', 'origin': 'lower'}), | |
| (pitches, 'Pitch', 'Frequency (Hz)', 'plot', {}), | |
| (energy, 'Energy', 'RMS Energy', 'plot', {}) | |
| ] | |
| for i, (data, title, ylabel, plot_type, plot_args) in enumerate(plots): | |
| getattr(axs[i], plot_type)(data, **plot_args) | |
| axs[i].set_title(title) | |
| axs[i].set_ylabel(ylabel) | |
| axs[2].set_xlabel('Time') | |
| plt.tight_layout() | |
| with tempfile.NamedTemporaryFile(delete=False, suffix='.png') as temp_file: | |
| plt.savefig(temp_file.name) | |
| plt.close() | |
| stress_interpretation = "Low" if normalized_stress < 33 else "Medium" if normalized_stress < 66 else "High" | |
| return f"{normalized_stress:.2f}% - {stress_interpretation} Stress", temp_file.name | |
| except Exception as e: | |
| return str(e), None | |
| def create_voice_stress_tab(): | |
| with gr.Row(): | |
| with gr.Column(scale=2): | |
| input_audio = gr.Audio(label="Input Audio", type="filepath") | |
| with gr.Row(): | |
| clear_btn = gr.Button("Clear", scale=1) | |
| submit_btn = gr.Button("Analyze", scale=1, elem_classes="submit") | |
| with gr.Column(scale=1): | |
| output_stress = gr.Label(label="Stress Level") | |
| output_plot = gr.Image(label="Stress Analysis Plot") | |
| submit_btn.click(analyze_voice_stress, inputs=[input_audio], outputs=[output_stress, output_plot]) | |
| clear_btn.click(lambda: (None, None, None), outputs=[input_audio, output_stress, output_plot]) | |
| gr.Examples(["./assets/audio/fitness.wav"], inputs=[input_audio]) | |
| with gr.Blocks() as demo: | |
| gr.Markdown("# Voice Stress Analysis") | |
| create_voice_stress_tab() |