Re: [Harp-L] Tremolo harmonicas

[Winslow Yerxa <winslowyerxa@xxxxxxxxx>]

RichardHunter writes:


===
The Detune effect in the Digitech RP shifts pitch by cents, not semitones, and so can produce a convincing tremolo effect similar to the one you'd get with a tremolo harp.  In other words, a detune effect (whether produced by a Digitech device or any other) is a pitch shifter that 1) produces a parallel signal to the original, and 2) shifts the pitch of the parallel signal by very small amounts.  So I don't understand Winslow's comment that "tremolo is based on differences between actual frequencies, not cents," given that a cent refers to a frequency difference--specifically, a "cent" = 1/100 of the difference in frequencies between a given pitch and another pitch 1/2 step higher or lower. 
===

I guess you don't hang around accordion forums very much, Richard.

Actually, your formula is incorrectly stated. Cents are not based on 1/100th of the frequency. They are based on 1/100th of the semitone, which itself does not have a direct linear relationship to frequency.

The beating in what we call tremolo in harmonicas and accordions is a direct result of the difference in frequency (vibrations per second) between two reeds (or, in accordions, sometimes three reeds).

For instance, if one reed vibrates 440 times per second and the other beats at 441, you'll hear one beat per second - a very mild tremolo. Likewise if one reed vibrates 880 times per second and the other beats at 881, again, you'll get 1 beat per second - the exact same tremolo rate.

Yet the difference between 44 and 441 is 3.93 cents, while the difference between 880 and 881 is only 1.97 cents.

Now let's try using cents to create tremolo ratios.

If we decide on a beating of 4 cents between reeds, then you'll get 440 and 441.02, or again approximately 1 beat. Apply that same formula to 880 and you'll get 882.04, or about 2 beats. So the tremolo reeds will be tuned in octaves relative to each other.

That result may look benign, but tremolo rates are usually much faster, and when you use cents to do a wholesale, uniform shift, the beating rates in the upper octaves get so fast that they become hurtful to the ears. As a result, accordion tuners have devised many different tuning schemes to temper the sound of tremolo. By blending tremolo rates, they achieve a complexity that can be quite beautiful when chords are played.

I've tried using pitch shifters and have compared them directly to an actual tremolo harmonica in recordings of the same passage. In comparison, the pitch shift version sounds kind of artificial and lifeless. Which is why, in my original remarks, I said that you can achieve a *sort-of* approximation.

Winslow