Re: more tuba spectra


[ Follow Ups ] [ Post Followup ] [ TubeNet BBS ] [ FAQ ]

Posted by Rick Denney on July 24, 2001 at 09:21:24:

In Reply to: more tuba spectra posted by Kenneth Sloan on July 23, 2001 at 22:19:53:

Yes, the fundamental is clipped, and pretty severely. The problem with this sort of clipping is that it will produce a whole raft of spurious frequencies that will appear in the spectra.

A few facts that were apparent: The note being played was an F at the bottom of the staff, but it was about a quarter tone sharp (or an F# that was a quarter tone flat). The fundamental frequency was 90 Hz, and the divisions on the time chart (which you helpfully cropped, heh, heh) are 0.005 seconds apart. I don't think you played the note really loudly, because of the lack of higher harmonics, unless my guess down below is correct.

The high recording level probably has the effect of compressing the differences between bottom few overtones, such that they are more similar than what appeared on my spectra. But there are all sorts of potential explanations for this. The strong fifth overtone compared to a weak fifth overtone in all my samples, and the abrupt change in value from the sixth overtone and up is unlike what I saw.

Of course, we don't know nearly enough about what these spectra mean to be able to identify the instrument solely from these samples, unless we can play them for our better spectrum analyzers on either sides of our brains. But I'll go out on a very thin limb and guess that this is not a tuba, just to liven up the discussion.

The clipping might be caused by a number of factors, not just the microphone. In fact, I would bet that it isn't the microphone. Note that the level you show in Cool Edit is much higher than in my own samples (by 15-20 dB). Yet I thought I was recording onto minidisc at fairly high levels. You might have overdriven the minidisc inputs, or the sound-card inputs when you recorded it onto the computer. If you went directly into the computer, you might have recorded it at too high a level. The microphone's peak input sound pressure level is 125 dB, which you'd only be able to exceed if the mike was down the bell. Of course, the microphone placement could easily affect the spectrum such that my previous guess is wrong.

Rick "who would have a hard time playing an F a quarter-tone sharp, but would have no trouble playing an F# a quarter-tone flat" Denney


Follow Ups: