Saturday, April 10, 2010

How signal processors earn their paychecks

As we noted earlier, acoustic tomography relies on accurate measurements of the time it takes for a signal to travel between a source and a receiver. The speed of sound depends on the water temperature. Since sound travels faster in warmer water, we would expect the travel time between two points to get shorter as the water warms up.

Yesterday we showed you what the source signal looks and sounds like. That recording was made during one of source tests conducted a couple days ago. During these tests the source was lowered on a wire from the ship to a depth of about 1 kilometer. The receiver was located on the same wire, about 500 meters above the source. Since the receiver was so close to the source, it was quite easy to hear it.

In a typical acoustic tomography experiment the source and receiver are much further apart. In the experiment we are setting up right now, the shortest range between source and receiver is about 125 kilometers, and the longest range is about 640 kilometers! Since sound signals lose energy as they travel, hearing them at these long distances is difficult. The picture below shows the signal received from one of these sources at a range of 500 kilometers:

Can you tell where the signal starts and ends? Probably not from this picture! In order to see when the signal arrives, we have to use tools developed by signal processors. Applying these tools to the signal above, we get the following picture:

This picture is a lot easier to read since it contains only two big spikes. These spikes correspond to the arrival times of two signals.

Signal processors earn their paychecks by designing signals (like the one we're using) and processing tools to make those signals easier to detect at long distances. In addition to oceanographic applications, signal processing is also used in cellular phones, digital tv's, MP3 players, and lots of other electronic devices. Most signal processors are trained as electrical engineers (like Kathleen). If you're interested in learning more about signal processing, feel free to email Kathleen. You can find her email address on her website.