Welcome to the

Hydroaeropropulsion Rocket Research Labs

homepage!

4 juli 2003

Yesterday evening, I discovered that my computer wants stereo input on the microphone connector. A quick and dirty fix is to unplug the microphone for a few millimeters (finding this out required half a day of fiddling of course); this worked but the signal was very faint, even when I put the microphone only a few centimeters away from the speaker. This is the signal after serious amplification:
The sampler program said that the frequency was 1256Hz, which is not bad given that it was calculated at 1240Hz!

Today, I made a mono-to-stereo converter plug which solved the problem. The signal is loud and clear.

You can see here that I held the microphone closer and closer to the speaker. But strange enough, the signal consists of bursts of sound during about 23.3 milliseconds with about even long silences in between them! This is what the first few milliseconds of such a sound burst look like:
The large peaks are about .83 milliseconds apart (corresponding to a frequency of about 1200Hz) and there are a bit smaller peaks about .44 milliseconds apart (the first overtone an octave higher).

However, further expermentation learns that the computer always seems to record about 23 milliseconds of sounds followed by 23 milliseconds of silence ... And maybe it's a coincidence, but at the sample rate we use (44100 samples/second), 1024 (=2^10, a number which computers are fond of) samples last 23.21 milliseconds ...

More experimenting learns that sound recoding in mono or at 8000Hz does not seem to work at all, but 44100Hz stereo (even though the microphone itself is mono) works satisfactorily (the strange 23-millisecond silences magically disappear). It's strange but who told you that computers are supposed be logical anyway? And to make things more interesting, the signal is now rather weak again.