The Crosstown Parkway Extension in Port St. Lucie, Florida, was a long-awaited project that included 1.5 miles of corridor improvements and a new 4,032-foot-long bridge over the North Fork of the St. Lucie River serving as a critical hurricane evacuation route. At the end of the project, residents along the Crosstown Parkway corridor began reporting traffic-related disturbances that were interrupting sleep and limiting normal conversation outdoors. In response, the City contracted KCI to evaluate the issue. A two-phase traffic noise monitoring study was performed aimed to determine whether levels exceeded FHWA Noise Abatement Criteria. As part of the study, 4400 audio recordings were captured from 84 hours of monitoring at eight measurement sites, and each file had to be listened to and classified manually—a tedious process that consumed two team members for nearly three weeks.  

Recognizing the need for a much more efficient approach, our team developed SoundScanNX.   

Instead of manually analyzing a massive amount of data, spending days if not weeks to get results, and challenged with inconsistent labeling due to fatigue and distractions, SoundScanNX uses machine learning to automatically identify the dominant sound in each audio recording, distinguishing traffic noises such as cars, trucks and motorcycles, from non-traffic sounds, including dogs, birds, voices, and more. The end result? Those 4400 audio recordings that took 3 weeks to analyze were processed and classified by SoundScanNX in 30 minutes.  

How SoundScanNX Helps  

AI solutions like SoundScanNX bring AI efficiency to traffic noise analysis, which is especially important for long-term monitoring projects. Benefits include: 

• Optimizing Time & Costs: SoundScanNX shifts the classification effort from weeks of expensive human labor to minutes with automated audio file processing. 

• Delivering Consistency and Accuracy: Unlike humans, SoundScanNX won’t get tired, suffer burnout, or deliver inconsistent results. SoundScanNX is trained to provide consistent, objective, and actionable results.​ 

• Providing Results for Targeted Mitigation: SoundScanNX goes beyond the simple analysis of noise vs non-noise, providing granular detail about sounds (e.g., loud mufflers vs truck brakes) that is necessary for effective, targeted noise abatement solutions. 

• Generating Noise Characteristics. SoundScanNX delivers detailed characteristics about each detected sound, including amplitude, LAeq, and duration, which gives planners and engineers the data they need to customize solutions for their clients.  

Building Intelligence with BRYX 

SoundScanNX is one of several models we offer on BRYX, KCI’s secure, scalable Model-as-a-Service (MaaS) cloud platform for AEC users​. The BRYX catalog of pretrained models​ are tailored to use cases for transportation, water, power, environmental, and construction​ industries, and can run on demand or through integration with third-party APIs. Start a free trial of SoundScan NX to see how our model delivers faster and more detailed results for your traffic noise impact studies.