Chiff and Fipple Forums

I was fiddling around with a shareware spectrum analyzer and decided to put some of my whistles to the test. Here are the results:



For each whistle I would play high G and then low G. The order I placed the whistles in was predetermined by what I thought was purity of tone. The yellow line across the top is the actual wave file. The main vertical lines are approx 800hz apart. And the lighter the color, the louder the particular frequency.

I'm sure that there are a lot of folks on the board that know way more about this than me. I just thought it was kinda fun to SEE the sounds.

Super cool! Thanks for sharing -- the last time I looked at plots like these was circa 15 years ago when I was designing radar. This is _far_ more interesting. And man, is that hoover pure!

Would love to see the Sweetone v Meg plot!

Oh yeah, I forgot...the program I used is available for download at:

http://www.visualizationsoftware.com/gram/gramdl.html

Neato! Can you give us a brief, layperson's description of what we are seeing, and what the differences mean? Different overtones? Thanks

Vapor, this is way cool!

Paul, the frequency is the vertical scale, so the lowest bar is the fundamental (about 494 for the low-G and 988 for the high-G, and as you speculated, the succeeding bars are harmonics, but it gets more interesting. The harmonics should only show up at multiples of the fundamental, as in the Hoover picture. Also, the narrowness of the bar gives an indication of the purity of the sound.

So, in the upper octave, the Hoover is the only whistle that doesn't have a 1.5 harmonic (one beside the third line of the low-D). The presence of this "forbidden" harmonics is a kind of harmonic mixing, possibly an indication that some of the lower octave is coming through. Also, as you go to the right (good ears, Vapor), the bars become wider, eventually winding up with the notorious Weltmeister, which has an upper octave that sounds horrible.

The presence of the forbidden harmonics and the widening of the frequencies is presumably what we call chiff, so now we have a hard-and-fast definition of it.

Are the widths of the bars affected by the volume of the notes played or are they strictly harmonic tones or overtones?

Thanks, vaporlock! Very interesting! Now, if you can rig up a controlled air source, you could also look at decibels vs. air volume (sensitivity) and how much air it takes for the whistle to jump octaves.

A controlled air source is easy to do. The next issue would be "focusing" the air source. Better whistlers tell me there are things you do with your mouth that affect response beyond simply how hard you blow.
- Joel

On 2002-09-12 08:18, chas wrote:
Paul, the frequency is the vertical scale, so the lowest bar is the fundamental (about 494 for the low-G and 988 for the high-G, and as you speculated, the succeeding bars are harmonics, but it gets more interesting. The harmonics should only show up at multiples of the fundamental, as in the Hoover picture.

As I understand it, you don't change the fundamental when you switch to the upper octave, you just blow hard enough to get the fundamental to drop out of the sound, so that it's all overtones.

On the bassoon, for instance, you can play chords by fingering one of the lower notes but adjusting the way you blow so the fundamental doesn't sound. With a little work you can get
two or three of the overtones sounding with about the same volume -- instant chord!

The presence of the forbidden harmonics and the widening of the frequencies is presumably what we call chiff, so now we have a hard-and-fast definition of it.

Isn't chiff the sound of the attack? None of that appears to be visible in the graphs -- they're all sustain.

This graph makes me want a Hoover!!!

I think I'll add that to my Christmas wish list for the years to come...since I'm getting a Hammy flute, I think my Christmas list is pretty much taken already for the next few years.

Best wishes,

--James
http://www.flutesite.com

On 2002-09-12 09:26, colomon wrote:
As I understand it, you don't change the fundamental when you switch to the upper octave, you just blow hard enough to get the fundamental to drop out of the sound, so that it's all overtones.

Oops! you're absolutely right!

Isn't chiff the sound of the attack? None of that appears to be visible in the graphs -- they're all sustain.

That's how it's generally defined for organs, but there really isn't a hard-and-fast definition for whistles, at least not among the posters here -- whence the smiley.

This is what I've always thought, and the view I've always proposed here...but others disagree. Even the Clarke's web page once described whistle chiff as that mysterious undefinable sound quality of the whistle tone (paraphrased from memory). Me, I'm a computer programmer. I like a more hard-and-fast definition of things, and so I like the pipe-organ definition: Chiff is the sound of the attack.

On 2002-09-12 09:26, colomon wrote:
Isn't chiff the sound of the attack?

What does "sound of the attack" mean?

I just plotted a Hoover G, Dixon G and Burke G. The "noise" between the harmonics seems to be the "airy" sound we so often describe. I will post that picture as soon as I prepare it.

As far as volume goes on the plot, dark blue->yellow = softer->louder.

I would really like to see some people with other whistles do some comparisons and post the results. How about a Sweet, Thin Weasel, Busman, Abel comparison? Chieftain, Overton? Sweetone, Meg. Better yet, how about a Generation, Generation, Generation, Generation comparison to see just how bad quality control is. I can host any of the pictures if you email them to me.

Once again, the software is available free at:
http://www.visualizationsoftware.com/gram/gramdl.html

I agree with the other posters about chiff. The physics behind open organ pipes is identical to whistles, so I feel that chiff is defined the same way - the sound of the attach. A theory (I'm still working on this) is that the edgetone (the sound of the wind hitting the sharp edge) in the fipple is at a different frequency than the pipe vibration (there is lots of agreement here). However, the pipe can't begin resonating until the sound has had an opportunity to go the length of the pipe (and maybe back again, I'm not sure) so for that brief instant, you hear pure edgetone. After that point, the edgetone becomes coupled to the pipe vibration and you hear the notes we're used to.

Anybody else got a theory here?