Oxygen Tank Bells?

Hello All,

I'm new to this group & was referred by someone from the oddmusic
group. I am currently in my predicament due to a neighbor getting me
hooked on this idea of making bells out of old gas cylinders. ;->

So I've found a source of decomissioned tanks (the ones welders use)
and have started cutting the bottoms off of them.

They resonate nicely, but I'd like to make 5 (or so) into a big bell
xylophone and I'm having some trouble figuring out how to get the
right "pitch" from each bell to get the appropriate range. I'm
trying to figure out what length will give me what general note. I
can then try and tune from there (add or remove metal ;->)

I'm keeping track of this work in progress at
http://www.squidgey.com/bells/ if y'all are curious or want to see
(and hear) what I'm rambling on about.

Right now, I have 2 main questions (or 2 vexing problems impeding
progress):
1) Which tone in a bell is the one that our minds select to be the
bell's pitch? For example, on a larger bell, the hammer/strike tone
and the hum/ring tone sound different. The strike tone is usually
higher.

2) Does anyone have any good ideas on how to measure these tones? I
am trying a variety of methods but am not quite sure of the accuracy
of each one. I can make my multimeter read in Hz so I can
watch the numbers flash by, but there are so many tones and they vary
in amplitude and decay time such that it is tough to get a reliable,
repeatable number this way.

I also have a tone generator which I can use to try and match the
bell's sound "by ear". This really only works for the ring tone and
of course is quite subjective.

I then found this little program called wavanal from a church bell
website. This thing analyses sounds files and spits out a list of
tones. If your bell fits into the simpson tone subset, the program
will try to ID the tones (hum tone, partial, etc). Of the 4 or 5
experiments so far, only 1 happened to accidentally fit into that
simpson set.

This is turning out to be a fun project.

While I understand some general musical relationships from playing
the guitar and banjo, my knowledge of musical theory is rudamentary.
Reading through some of these posts are pretty far over my head so I
thought I would go ahead and ask the general question.

Thanks in advance,
Flint Weiss

--- In tuning@y..., flint@s... wrote:

> Right now, I have 2 main questions (or 2 vexing problems impeding
> progress):
> 1) Which tone in a bell is the one that our minds select to be the
> bell's pitch?

Generally, it's not one of the physically present ones at all!
Instead, the brain sorts through the various spectral components,
finds the strongest hint of a harmonic series, and hears the "pitch"
as the best-fit fundamental to this series. I posted an excellent
webpage on this some time ago but don't have the link handy.
>
> 2) Does anyone have any good ideas on how to measure these tones?

As the "pitch" is not physically present, it's not easy to measure!
With standard carillon bells, centuries of experience with that
particular inharmonic spectrum have made the process rather well-
defined, but for your project we may well be in some uncharted
territory. I think your best working hypothesis will be that the
tanks won't have a very clear "pitch" but rather there will be
several competing "pitches" involved in our perception of these
sounds.
>
> I also have a tone generator which I can use to try and match the
> bell's sound "by ear". This really only works for the ring tone
and
> of course is quite subjective.

Right . . . that's what I was trying to say . . . but it may be your
best bet. Just don't throw away any "extra" pitches you may be
hearing in these tones . . . they may be productive basis on which to
form tunings and scales later on . . .

>2) Does anyone have any good ideas on how to measure >these tones? I
>am trying a variety of methods but am not quite sure >of the accuracy
>of each one. I can make my multimeter read in Hz so I >can
>watch the numbers flash by, but there are so many >tones and they vary
>in amplitude and decay time such that it is tough to >get a reliable,
>repeatable number this way.

Just a thought. Have you looked at any of Jean-Claude Risset's work with Bell tones?

The way people usually try to figure out things about timbre is to use computers, as you obviously know. Theres a whole ton of work that has been done in this field, meaning timbre, and so check out the computer music literature.

A book which I have in front of me at the moment which discusses bell tones is "Computer Music" by Charles Dodge and Thomas A. Jerse.

Theres lots more than this around on this topic, but its one example.

Cheers,
Andy