Hydrophone and identification of underwater sounds

[mgrouch]

https://www.instructables.com/Lets-Build-Some-World-Class-Hydrophones/

https://www.instructables.com/The-Gladys-Hydrophone/

[mgrouch]

Here what DeepSeek suggests

Several open-source tools and software solutions exist for listening to underwater sounds and detecting propeller sound signatures. Here’s a summary of the most relevant options:

1. FOSSA (Free and Open Source Software for Acoustics)

• Developed by NOAA’s Southwest Fisheries Science Center, FOSSA provides a streamlined approach to analyzing passive acoustic data, including propeller noise detection.
• Includes packages like PAMpal (for extracting acoustic metadata) and BANTER (for event-based classification of sound sources).
• Built on the open-source language R, making it adaptable for machine learning and species classification tasks .

2. PAMGuard

• A widely used open-source software for passive acoustic monitoring (PAM), compatible with hydrophone data.
• Supports real-time detection and classification of underwater sounds, including propeller cavitation and ship noise.
• Integrates with FOSSA’s PAMpal for downstream analysis .

3. DEMON (Detection of Envelope Modulation on Noise)

• A traditional but effective method for detecting propeller signatures by analyzing amplitude modulation in acoustic signals.
• Implementable in open-source signal processing tools like Python (SciPy/NumPy) or MATLAB alternatives (Octave).
• Used to identify shaft and blade rates of propellers from hydrophone recordings .

4. Fast Kurtogram and Cyclostationary Analysis Tools

• Open-source implementations (e.g., in Python or R) of cyclostationary analysis and spectral kurtosis can detect propeller noise in low signal-to-noise conditions.
• Useful for isolating cavitation noise and machinery signatures from ships .

5. Discovery of Sound in the Sea (DOSITS)

• While not software, DOSITS provides open-access acoustic datasets, including ship and propeller noise signatures, useful for training detection algorithms .

6. CORMA (Low-Cost Underwater Noise Monitoring)

• An open-source, real-time passive acoustic monitoring system using off-the-shelf hardware.
• Designed for scalable deployments and can capture anthropogenic noise, including ship propellers .

7. SanctSound (NOAA)

• Provides hydrophone data and tools for analyzing underwater noise, including vessel signatures.
• Includes machine-learning-ready datasets for propeller noise identification .

Key Features for Propeller Signature Detection:

• Frequency Analysis: Propeller noise typically peaks in low frequencies (20–500 Hz) but includes cavitation harmonics up to higher ranges .
• Modulation Detection: Tools like DEMON or cyclostationary analysis extract blade-pass frequencies .
• Integration with AIS: Combining acoustic data with ship tracking (e.g., via AIS) improves source attribution .

For implementation, Python libraries like Librosa or HydroMoth (for low-cost hydrophones) can complement these tools. If you need specific code examples or datasets, the Watkins Marine Mammal Sound Database and NOAA’s SanctSound are excellent starting points .