The Sound of Bells

Home Page Next Previous Up

Relative intensity of partials

As shown elsewhere, the attack and decay times of the different partials vary considerably. In particular, the 'strike' partials - nominal, superquint, octave nominal etc. - usually have a high initial amplitude when the clapper strikes and then decay rapidly. This means the ear can hear quite a different sound when a bell is rung in changes as compared to it being rung or chimed alone. In changes, the strike partials have a very strong effect for the short period of time before the next bell strikes, so the predominant sensation of the bell is it's virtual pitch or strike note. Rung alone, other longer-lasting partials in the bell can dominate the sound.

In recordings of different bells under different circumstances there can be considerable differences in the relative intensity of the various partials. To illustrate this point, the plot below shows the intensities of partials for three bells of different periods; a Taylor bell of the early 1980s (the 'Perrin and Charnley' bell), a Mears bell of 1859 (the tenor from Ranmore, Surrey) and a London bell of circa 1400 (the smaller bell at Great Bookham).

Intensity profile of three bells

The plot deserves some explanation. The partial intensity of all three bells is normalised, so that the loudest partial has intensity 1.0. The horizontal axis has units of cents from the nominal, so that it is logarithmic in frequency, and corresponding partials of each bell are roughly aligned.

The recordings from which these plots were derived (all taken on the same video camera) have different qualities: the Bookham bell has a very loud hum, a moderately loud prime, a quiet nominal and little above. The Taylor bell has partials which are loud below the nominal and quiet above, though the intensity is more balanced that for the Bookham bell. The Mears bell has loud partials above the nominal. The difference in quality is clear; from dull and soft to brassy and bright.

Experiments in December 2002 using an apparatus that delivers a clapper blow of known and variable strength to a bell showed that at least two different effects were at work:

Other effects yet to be investigated include:

The changes due to microphone position are an example of the importance of room acoustics to the sound of bells. My experiments on simulation of bell sounds from basic principles, which have now reached an advanced and successful stage, have shown that it is essential to include room acoustics in simulations if the result is to be at all realistic.

I have not yet tried to quantify the result of any of these factors on bell timbre. However, it is possible that there are residual differences in partial intensity that are due to differences in the design of the bell, in particular its thickness at various points.


Last updated May 1, 2004. Site created by Bill Hibbert, Great Bookham, Surrey