New Forest cicada or Roesel's bush-cricket?

Comments: 2

Posted by Davide, 30 July 2014

Today I wanted to share with you something a bit more technical about the call of the cicada. You may have seen on twitter that a few days ago we got excited about a very promising recording that triggered our cicada detection. It turned out to be, most likely, the very common Roesel's bush-cricket, in the picture here.

This bush-cricket is actually an insect that our algorithm is capable of classifying, normally at a good level of accuracy. It starts singing as the cicada season is coming to an end, around mid-July. But the most tricky aspect is that its call is actually very similar to that of the New Forest cicada, at least when recorded at 44.100 Hz (that's the sampling rate rate we use, and it's very common. Most commercial music, CDs and MP3 are distributed at this rate). They both have a very strong high-frequency component, and while the Roesel's bush-cricket's frequency range is a bit wider, this is not always visible if the recording is not perfect. Below you can see a spectrogram and a waveform of the two calls, where the Roesel's one is the one that got us hoping. If you look closely you will notice another thing. The call of the bush-cricket shows some very rapid chirps, which are actually even more noticeable when you listen to the sound. Don't be fooled by the pattern of the waveform (the blue line, which represents the amplitude, or the volume, of the call over time). The spikes you see are actually other noises around. In the case of the cicada, it's a car passing by.

Why am I telling you all this? Well, you've been searching cicadas for two years now, and we told you that listening to its call is the best way to find it, whether with a phone or with a good pair of ears. Now, we wanted to share something a bit more detailed about what you were actually looking for. There is plenty more to say, but I'm sure I'm starting to bore you already. So that's all for now, but as usual stay tuned for more news.


Name Comment
Clara Bressi May be a stupid question, but what's the green bit at the top of the frequency plot?
Davide Not a stupid question at all actually. That's the bit when the microphone in our phone starts to lose sensitivity. Like us, it has a frequency range at which it can "hear" well. Passed that, it can't register the sound any more. Green means really low values (because it can't hear them, not necessarily because there's nothing there), red means high values, so a strong sound at that frequency.

Leave a comment

You need to be logged in to comment. Login here.
©2012 University of Southampton -