Dear Everyone,
I've been experimenting, trying to get a good recording for spectral analysis. For my thesis research, I'm going to use glissandi to help show the upper and lower boundaries of the singer's formant:
Trouble is, I am fairly inexperienced, and can't tell if maybe my setup needs help. There are some things I've noticed in my spectrogram, and I want to make sure they're normal. Here is the image:
http://people.mills.edu/toda/music/gliss.html
1) When the fundamental is between C4 and C5, bands appear above the singer's formant in a repeating pattern, defined by nodes where there is less acoustic energy. Is this normal, or is this some weird artifact?
2) When the fundamental is between C5 and A5, the usual singer's formant pattern degenerates - instead of bands, higher partials start being represented, in a classic dropoff pattern. The nodes seem to disappear.
Is this just the nature of my voice? Is this a fairly typical spectrogram? Or have I set up my mic/room poorly?
This glissando was recorded in a very dry room (but not completely), maybe 40 meters squared. I used an AKG 414 in omni mode, set 30cm away from my mouth.
Thanks for your feedback!
Tako
|