BRYX

SoundScanNX

This model helps streamline the process of reviewing audio for traffic noise studies by automatically distinguishing traffic noise from non-traffic related noise, eliminating time-consuming and error-prone manual processes.

400 credits (with a subscription)

800 credits (without a subscription)

Version 1.3
Free Trial available!
SoundScanNX
Machine Learning (Neural Network)
This machine learning model performs audio identification to detect and classify a variety of environmental noise from audio files.
Use Case
Instead of listening to hours and hours of audio for noise studies, traffic engineers and consultants can use SoundScanNX to automatically categorize and measure traffic noise from other noises in the environment.
Fast Analysis
This model is currently trained to detect:
  • Traffic Sounds:
    • Cars (tire squeals, sirens, modified mufflers)
    • Motorcycles
    • Trucks
  • Non-Traffic Sounds:
    • Airplanes
    • Birds
    • Dogs
    • Lawn Equipment
    • Music
    • Voices
Accuracy
Training Accuracy: 90.8%
Wild Data Accuracy: 44.2%
Download Sell Sheet

Required Inputs

  • MP3, OGG, and WAV files

    • Note: For low-volume signals, we recommend amplifying the audio files prior to running SoundScanNX.

    Expected Outputs

    XLSX summary report which includes the following attributes for each sound clip:

    • Class
    • Confidence Level
    • Peak Amplitude (dB)*
    • Average Amplitude (dB)*
    • LAeq (dBA)
    • Peak Duration (sec)

    *Calculations assume a microphone calibration of 114 dB

    BRYX

    SoundScanNX analyzes audio files taken for traffic studies and identifies/classifies a variety of traffic and non-traffic sounds. With high identification accuracy while significantly reducing the amount of manual effort required, this model provides an optimized solution for addressing noise complaints.