Klarinet Archive - Posting 000534.txt from 1998/08

From: "Mark Charette" <charette@-----.org>
Subj: Re: Re: Re: [kl] A nasty question about Buffet
Date: Wed, 19 Aug 1998 12:38:41 -0400

>In my next life, I plan to do a thorough analysis and compilation of
these
>data...

Which would be difficult ...

I once did a statistical survey of Steinway pianos to try and answer the
question "are the old Steinways better than the new ones?" Most people
thought the older ones were much better than the newer ones "warmer,
better touch, etc., etc.) I went through Steinway's data on extant
Steinway pianos, and came up with noise. (I did this while looking for a
grand piano. Didn't end up buying a Steinway.)

The reason?

As the instruments got older, the ones which were marginal were not
repaired or serviced as often as the best ones, and were probably given
away to schools and such. Steinway then didn't follow the piano's
records anymore. Those instruments that were "the best" for those years
were carefully followed and all services recorded. The ones left (and in
use today from earlier years) _are_ better than the average Steinway of
today _or_ yesterday. However, because of the ones that got retired, the
data is skewed if corrections for retired instruments isn't made; it
will appear as if the older ones are better.

There were a number of very bad years in there - the number of pianos
left for those periods is very small - and there are of course some
anomalies, but in general it followed a pretty obvious pattern:
The longer the piano stays around, the better its original quality.

I can't legally extrapolate my Steinway data to clarinets; however, if a
clarinet cracked badly 20 or 30 years ago I will make an assumption that
the likelihood that it would be around today to be counted is small.
It's an educated guess, though, not based on any numbers.

One last thing. The number of clarinets made today (if we are to believe
that the serial numbers are consecutive) is enormous compared to earlier
years. My chart indicates about 2000 in 1936, about 3000 in 1946, about
3000 in 1956, about 4000 in 1966, 10300 in 1976, 14800 in 1986, and
15000 in 1996. If a steady 1% cracked, then we'd see about 20 cracked
clarinets in 1936 but 150 cracked clarinets in 1996! It would only
appear to be worse.

Since I don't have any data on cracked clarinets the percentage _may_
have gone up during particular years. This little note is just to remind
everyone of the danger of numbers without supporting documentation and
the skewing of numbers when compared over time.

Mark Charette
----
Mark Charette@-----.org
Webmaster, http://www.sneezy.org/clarinet
All-around good guy and devil-may-care flying fool.
"There can be no freedom without discipline." - Nadia Boulanger

-------------------------------------------------------------------------