Uroflowmetry is a well-established and commonly conducted diagnostic test for Lower Urinary Tract Symptoms (LUTS), which assesses the severity of any blockage or obstruction by measuring the volumetric speed of urination. The urine flow is measured by a specially designed clinical scale called "uroflowmeter" while a patient urinates in a urology clinic. Recently, for at-home self-check of urinary flow to monitor full daily picture of voiding, several alternative uroflowmetry schemes with various modalities have been proposed, and sound-based models are regarded as a promising solution with its non-invasive and ease-of-use nature. These alternative acoustic-based approaches were developed and validated with the limited number of healthy human subjects, which is known to have a bell-shaped curve. However, the urine flow of patients shows various irregular patterns with low sound level and low signal-to-noise ratio, it is strongly needed to validate the prediction methods with various patients prior to utilizing new sound-based approaches as reliable medical applications. This study aims to validate existing acoustic-based urine flow prediction approaches with a large clinical sample including both normal subjects and patients. Uroflowmetry data and sound recordings of 137 subjects were collected in the urology clinic at a university hospital, and the performance of the existing prediction models was assessed with three key clinical urodynamic parameters including the maximum flow rate, voided volume and voiding time.