How to emulate the sound of a ringing bell?

[Rootzee]

Summary: how to apply the the changes in frequency one can hear when a church bell is swinging through the air onto other acoustic sounds in a computer?

Hello all, new to this forum. For an (art) sound installation I would like to apply the the changes in frequency one can hear when a church bell is swinging through the air onto other acoustic sounds in a computer. I understand there are many variables ( size, material and thickness of the bell, speed it is swinging with, force it is being hit with, rate of decay, surrounding acoustics, etc etc) but I am wondering if anybody knows of certain frequency curves that are at play in the changing sounds (Doppler effect among others?). My background is in music, not in physics. I am using Logic Pro as my computer software to record and edit, but I can get my hands on other software if necessary. Perhaps it would be useful to first establish the sound frequency characteristics of a generic bell (sounding in place), and then add the expected changes when swinging in the air?
Many thanks for your thoughts.
( my sincere apologies if this is not the right place to post this - please advise as to where to post this properly)

 

[sophiecentaur]

Hello and welcome to PF. This a fair enough place to be posting this question imo.
That 'phasing[ sound you get when a bell swings depends on distance moved and also speed of the swing. Hard stuff to analyse and synthesise but it involves multiple modes of vibration of the bell, all of which have slightly different frequencies and which emerge at different phases, adding together to get that characteristic sound. The motion of the bell around its pivot (yoke) will add doppler shift according to the natural period and amplitude of the swing.

Why not record some bell sounds and use the samples you like most? To sound most convincing, I imagine it would be good to use a tenor bell as source for the lowest notes and treble bell for the highest notes - plus other bells for the notes in between - because the bell sounds won't scale convincingly. Then use your software to shift pitches to get the scale you need.

I suggest that because I'd bet that bell design is as much an art as a Science and the technology goes way back before the use of advanced Maths. I reckon that pitch change would achieve most / all of what you want, especially if the sound is to be mixed with others.

 

[glappkaeft]

You should probably turn to the synthesizer community if you want to do this. Here is an article on how to synthesize a simple church-bell sound using modular synthesis, you can probably find a lot more online.

https://www.soundonsound.com/techniques/synthesizing-bells

 

[hilbert2]

It should be easier than simulating the sound of dripping water as in this YouTube video. There has to be some way to determine the vibration normal modes of a bell of given shape.

Another interesting thing would be to simulate the sound of an incoming cannonball or grenade...

 

[sophiecentaur]

Another interesting thing would be to simulate the sound of an incoming cannonball or grenade...

Or make a recording - perhaps not.

 

[jedishrfu]

I found this paper investigating church bell dynamics.

https://www.hindawi.com/journals/aav/2012/681787/

 

[Rootzee]

Wow, thank you all for your thoughts and quick reply! I will investigate all the links you suggested.
To further clarify: it's not so much that I am trying to create an ‘instrument ‘ capable of playing scales (although that certainly could be very interesting), it is the shifts in pitch (and any other characteristics) that happen when a church bell swings through the air that I would like to apply to other (prerecorded) acoustic samples. This all to be used in a sound installation titled Room for Reflection.
If anybody has any other suggestions I will be much obliged, and again thanks to everybody who posted a reply!

 

[anorlunda]

If you had a recording of the bell sound, and access to a spectrum analyzer, you could measure the time variations in pitch, and reproduce them in a simulation.

 

[sophiecentaur]

If you had a recording of the bell sound, and access to a spectrum analyzer, you could measure the time variations in pitch, and reproduce them in a simulation.

I think the OP has little more than his ear to assess the sounds (from his earlier statements). That is quite good enough to tune a recorded sound to something like what's required. I'm sure a subjective approach would give an accepted result. Sampling is a well accepted method for use in music.

 

[Rootzee]

To Anorlunda and Sophiecentauer:
Thank you for your thoughts ;-)
I actually used to have a rather sophisticated spectrum analyzer app on my previous smartphone and with great interested recorded and investigated the prolonged decay (20+ seconds) of a (once struck) high quality singing bowl. It was a feast to see the multiple peak frequencies dance around. It was almost possible to follow the deformations of the bell over time and see how it influenced the pitch fluctuations.
I even recorded a singing bowl swinging through the air... close... but not quite like a swinging church bell...
Unfortunately the company went out of business. I have looked into acquiring a pro-level spectrum analyzer but they are fairly expensive.
However, even with a sophisticated spectrum analyzer I imagine one can get a good visual representation of the pitch fluctuations but not necessary a model ( or pattern) that one can apply to other sounds.
I was hoping there would exist a theory of “bell sound”. I did look into the British paper on English bell ringing - as suggested by Jedishrfu ( thank you!) - but it’s not immediately apparent how I could apply the formulas posted there.
Perhaps indeed I need to look into the synthesizer community - possibly some creative soul developed a sound sculpting tool that indeed would mimic the characteristics ( not just the sound) of a swinging church bell.
Thanks to all for your input!

 

[Spinnor]

When you say swinging you mean like the following?