Signal processing and analysis.

A part of life.

Signals are a part of everyday life. They are in your cars, cell phones, power grids, your mp3 player and, yes, everywhere. By understanding signals better, we can improve a wide range of products and further the capabilities of mankind. MuchDifferent has built a new tool to perform real time signal analysis and processing to do just that.

Sonic AWE

This tool is called Sonic AWE. It’s like a spectrogram on steroids. And we want to share this with you, so you can play around with your own live recordings for free.

1. Overview of Sonic AWE
2. Download Sonic AWE

Audio-based mechanical analysis.

Machines sings with harmonies.

Background.

ABB Traction Engines in Västerås develops train engines with extreme demands on size, power consumption and torque. Yet, they manage to keep the design modular and flexible for different configurations. The construction also has strict requirements on perceived decibel levels which has to be accommodated for.

Insight.

MuchDifferent analysed the recordings to gain insight into how harmonies resonate with cavities in the construction and how super-positioned harmonies can create unpleasant peaks that violate standard regulations. Sonic AWE proved very useful in finding patterns thanks to its versatile ability to study the signal interactively.

Technologies.

Sonic AWE   /   GPGPU   /   Matlab   /   Octave

High frequency signal data.

Bats sings in ultrasound.

Humans hear up to 20 000 Hz.

Bats often produces sound at 50 000 Hz, or more.

CD quality means a sample rate of 44 100 Hz.

This recording was made at 500 000 Hz.

Background.

Bats use echolocation with supersonic sound to navigate and to find prey. MuchDifferent was asked to analyze some recordings made by scientist at Swedish University of Agricultural Sciences to see if different individuals could be identified by the sounds they emit. The audio was recorded using special hardware developed by Petterson Elektronik in Uppsala.

Insight.

Visualizing high frequency signals, such as bat noise, is an easy task when using Sonic AWE. Even with such dense signals Sonic AWE can still compute the Continious Wavelet Transform for the entire spectra (20-250 000 Hz). This proved very useful in understanding how the sounds and echos are composed. The high detail level produced by Sonic AWE and ease of use made it possible to see differences in the signals that would otherwise have been much harder to discover.

Technologies.

Sonic AWE   /   GPGPU   /   Matlab   /   Octave