Text this: Leveraging Synthetic Data to Develop a Machine Learning Model for Voiding Flow Rate Prediction From Audio Signals