Pink Noise

Adam Kumpf, a really smart guy at MIT, built a piezo pickup much like mine, although he installed his permanently in a guitar. On his page, he posted frequency response curves for his pickup versus a commercial piezo and a Fender Stratocaster. I liked this idea and given the varying results with my three pickups, decided to do something similar.

I'm guessing that the vibration of a guitar soundboard, especially when contacting human bodies of varying density, is complex. So, to replicate such an environment when testing each pickup, I conducted the following highly-controlled experiment. I taped each element to a thin wooden board using painters' tape and hung the board by string from a wooden ruler in front of my right monitor. The ruler was held down by the plastic lawn dwarf who resides on top of the monitor. I then played pink noise though the monitor and simultaneously recorded through each pickup.


The results show some of what I've heard so far with the three pickups. First, here is the frequency response curve of my pink noise sample (when adjusted for a 3 db slope, it's flat):

Here is the curve for the small piezo disc pickup:

For the large piezo disc pickup:

And, for the piezo film pickup:

All of the pickups have jagged responses, even when averaged. This isn't entirely bad, since you can use different microphones and pickups for various purposes, to emphasize or deemphasize different frequencies. But, these curves don't compare well with an average microphone curve.

Nonetheless, I think the small disc sounds pretty good. Its curve, though jagged, has some high peaks throughout the range. The big disc sounds loud and tinny. I'm not sure if that's evident in its curve, especially when compared to the small disc. The film pickup sounds the best but produces very little signal (evident in the amplitude of its curve; I had to blast the pink noise just to get that). Otherwise, I'm not sure what these curves show, except that frequency response in these pickups is uneven.

By the way, if you compare these curves with Adam's, you'll see that his are much smoother, even though his pickup is almost identical to my disc pickups. He told me that he didn't use pink noise but instead "recorded audio data over a series of chords and notes that spanned the entire range of the guitar and then did post-processing in audio software." This is, in a way, a more accurate way to test these pickups since an acoustic instrument would never produce pink noise but instead a smaller range of frequencies. I may try it, but all this has made me tired.