Requires Sound Science
We opened up the Google home page today and noticed something quite odd. An article called Sound Science was there, which was relevant to the world of acoustics. As acoustics can be quite a niche topic, it was a very interesting piece.
Reading through the piece has given us some valuable insight into the murky world of ‘underwater acoustics’. You can check out the article for yourself on Googles Homepage or go right to the article by clicking the button below.
This got us thinking about how technology will impact acoustic consultants in the future.
Welcome to Noise Blog.
The Google article follows the story of Daniel DeLeon, a young man who has found his way into engineering through underwater acoustics.
Acoustics is such a broad topic in physics that it can be broken down into several smaller areas of expertise, namely:
- Environmental Acoustics
- Building Acoustics
- Noise and Vibration
- Musical Acoustics
- Physical Acoustics
- Underwater Acoustics
In the article, Daniel uses a hydrophone to measure underwater sounds over a large period of time. As you would imagine, working with such vast amounts of data can quickly become overwhelming.
Consequently Daniel uses a machine learning tool called TensorFlow (created by the Google Brain Team) to help him analyse the data.
Machine learning is described by Samula Arthur (1959) as:
“Machine learning is a field of computer science that uses statistical techniques to give computer systems the ability to “learn”
(e.g., progressively improve performance on a specific task) with data, without being explicitly programmed.” – Source: Wikipedia – Machine Learning.
Much like Daniel, Prism Acoustics collect a large amount of sound data (airborne sound rather than waterborne). We are still in the early stages of development, but are now using machine learning via neural networks in our work.
Therefore it is our hope that these algorithms can not only speed up the analysis of acoustic data, but also find anomalies that further our understanding of the subject.
For example an acoustician could use neural networks to:
- Find correlation in sound insulation test data, allowing for more efficient building design.
- Work out the most efficient way of placing acoustic treatments in a concert hall.
- Analyse the effect of atmospheric effects on long term environmental measurements.
In conclusion; machine learning is already streamlining the workloads of many data heavy industries and will most likely become a standard part of our working lives.
The data revolution has only just begun in the world of Acoustics, but it seems like we can expect it to continue and bring greater accuracy and efficiency in its wake.
With it bringing a greater understanding of ‘Sound science‘, we can expect to provide even more