touch & tone

[Stephen Birkett]

Some comments that appeared on caut list about the Askenvelt/Galembo 
article in the August PT Journal:
>I wouldn't read too much into the results of the shank vibration 
>study because the vibrations were generated in an experimental setup.

I think you mean "because the results were generated in an 
experimental setup that was not representative of the conditions 
pertaining in a real piano action in use". The hammers were clamped 
and driven by oscillators acting at the knuckle to determine the 
resonant frequencies.

>As far as I know the different modes have not been isolated in real 
>piano playing.  I also don't believe modes have been related to tone 
>production.  The systems we're dealing with are incredibly complex, 
>and generating meaningful data is unbelievably difficult.  Stay 
>tuned--there is more coming in future articles.

As for the first point, we have observed hammer shank vibration modes 
in high-speed imaging of actions played under normal conditions. They 
are not easily seen on the upward moving hammer, because the speed is 
fast compared with the natural frequency of the shank vibration, so 
you only see a small protion of a cycle going up. Also the amplitude 
is rather small (ca 1-2mm effect on the hammer head). On the downward 
journey, as well as during contact, you can see a distinct 
oscillation under certain kinds of touch. Post-processing of the 
videos is effective at extracting the data on these modes of 
vibration.

On the second point, that has yet to be determined. In fact, we don't 
really have a good understanding of the hammer-string interaction and 
how even hammer design affects it, so the effect of shank flexibility 
on tone would fall under that study. We are starting to look at the 
relationship between hammer properties (including shank) and the 
precursors of string vibration (those travelling pulses). This is 
actually a funded project that is just starting this week with a new 
PhD student.

One final comment. The effect of hammershank vibration can be 
observed by the effect it has on causing hammer head to move (the 
shank/head glue joint is effectively rigid). This can change the 
contact footprint (scuffing). A 4ms contact, for instance, 
corresponds to a 250Hz cycle, so the shank vibration frequencies are 
definitely in the range where about 1 cycle or more occurs during 
contact with the string. If the amplitude is large enough and 
friction with the hammerhead and string is low enough you can 
typically expect motion in the order of 1mm during contact for a loud 
bow with a flexing hammershank. This scuffing we have observed with 
high speed imaging. We have also confirmed the frequencies with 
strain gages mounted on the top of the shank. All during regular 
playing of the key, not artificially clamped and driven hammershanks.

(I've included this on the pianotech main list too, since it is a 
fairly general subject that may interest people not on caut as well.)

Stephen