Use your phone for Advanced AI-powered Acoustic Classification
Detect faults, verify quality, and document acoustic events directly in the field. Record, label, train, and classify sounds — without installing sensors or hardware.
-
Faster troubleshooting and quality checks
-
Train your own models on real-world sounds
-
Audio classifications with your mobile phone
Bring acoustic intelligence into your daily work
Resonyx on mobile helps you diagnose issues, verify production quality, and document how your assets should sound. Build a database of sounds snippets.
Troubleshooting
Capture on-demand sound clips when something needs attention and identify abnormal behavior quickly.
Quality Control
Explore whether acoustic differences can be used to distinguish between known quality states in components or materials.
Field Inspections
Add acoustic intelligence to routine inspection tasks and classify events directly in the field.
How it works
Record a sound
Capture short sound clips exactly when something needs attention — no continuous monitoring required.
Label what you hear
Assign each clip to the right condition or quality state, using your own labels.
Train your model
Train an AI model tailored to your equipment, materials or process when you have enough examples. Add more audio samples for better predictions.
Classify on-demand
Record a new sound and get an instant classification prediction using your custom model.
Improve with feedback
Confirm correct predictions or correct wrong ones. The feedback strengthens the model over time.
About Model Performance
Model performance depends on the quality and consistency of the audio used for training and classification. Mobile recordings can vary across devices with various built-in raw audio filters, and monitoring environments, which may influence results. Automatic model training is included in all plans, while advanced optimisation can be provided as an additional service.
Demo projects
These demo projects illustrate how acoustic classification can reveal patterns, quality states, and material differences that are difficult to detect by ear. They highlight the types of real-world problems Resonyx on mobile can support.
Liquid Level Detection
Small changes in water level produce measurable differences in sound.
In this project, more than 2,000 recordings were used to train a model that distinguishes 20 different fill levels in a glass — each representing roughly 5% increments.
The result is a precise classifier that recognises subtle acoustic patterns that are difficult to detect by ear.
Composite Material Quality
Light tapping on carbon composite plates reveals whether the material is produced according to specification.
The model distinguishes between compliant parts and parts with hidden voids or structural inconsistencies linked to production defects.
This demonstrates how acoustic classification can uncover internal issues that are otherwise invisible in a visual inspection.
Features
Record On-Demand
Capture short sound clips during checks, tests, or unexpected events — no setup or continuous logging required.
Built-in or External Mic
Use the phone’s microphone for quick checks, or an external mic for higher-quality or direct equipment monitoring.
Feedback Loops
Confirm or correct predictions directly in the app. Each interaction strengthens the model and improves future classifications.
Train Your Model
Build AI models using real recordings from your environment. The system handles training and optimisation automatically.
Cloud Interface
Recordings are classified and safely stored in the Resonyx cloud for later access, with predictions returned in seconds
Custom Labels
Define your own labels — the pre-defined states you use to categorise recordings for training and classification.