Hi all,
i've been trying to get my head around how to physically model a didgeridoo. I've found some websites which provide some detailed information about didg's and how they work but i'm kind of at a loss as to the next steps, how to convert those pieces of information into something which produces sounds like a didg and can have inputs controls to adjust elements of the didg such as length/width of the tube, flar at the open-end, kinks, roughness of the tubes internal surface and all of the adjustments that the players mouth can make.
Although i generally use synthmaker to produce synths or effects, it has a code primitive which allows users to write code which is similar to C++, so any ideas which are written in C++ can be converted to work in synthmakers code module.
Could anyone offer any advice on how to get myself started with this?
Cheers
physical modelling of a didgeridoo
-
- KVRAF
- 5175 posts since 29 Apr, 2006
Cool idea
-
- Banned
- 12367 posts since 30 Apr, 2002 from i might peeramid
there's LOTS of info (and recorded techniques) of lip modeling online (cook, JOS, hut)
there's also LOTS of ways to do it.. and no rule that you can't devise your own techniques.
you could strap that to a waveguide for the bore but i don't think you'd be satisfied without some model of the mouth, as it's ~crucial to the performance tones we associate with didg.
??? a short kelly-lochbaum tract (like my fauna instrument, 5 segments can be extremely versatile) or even a couple of bandpasses or something.
if you're stuck at this stage, i suggest researching mass-springs more than using the nonlinearity methods some people use. that should give you an idea of some basic building blocks you can use to model various items.
(jos, everyone has plenty of mass-spring info.. try this.. http://www.iro.umontreal.ca/~eckdoug/vi ... pring.html )
there's also LOTS of ways to do it.. and no rule that you can't devise your own techniques.
you could strap that to a waveguide for the bore but i don't think you'd be satisfied without some model of the mouth, as it's ~crucial to the performance tones we associate with didg.
??? a short kelly-lochbaum tract (like my fauna instrument, 5 segments can be extremely versatile) or even a couple of bandpasses or something.
if you're stuck at this stage, i suggest researching mass-springs more than using the nonlinearity methods some people use. that should give you an idea of some basic building blocks you can use to model various items.
(jos, everyone has plenty of mass-spring info.. try this.. http://www.iro.umontreal.ca/~eckdoug/vi ... pring.html )
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.
-
- KVRian
- 814 posts since 12 Sep, 2005 from Renton, WA
I know it's not phsyical modeling at all but I've heard a convincing didgeridoo (at least it would be in a mix, where it counts) from an ARP Avatar (Odyssey w/o keys)....use of the sample/hold was the key 
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
-
- KVRist
- 81 posts since 28 Nov, 2005
I don't understand the motive to model a didgeredoo. I could understand if it was from a science 'Let us learn how a didgeredoo works], but why not just record a person playing a didgeredoo for your music. Where do you live? I live in Australia, where there is [obviously] a wealth of under utilised Aboriginal musicians who can play a Didgeredoo correctly and fully. Why not do some field recording.
I am a musician not a programmer, you may have guessed
I am a musician not a programmer, you may have guessed
-
- KVRist
- 112 posts since 20 Mar, 2004 from UK
Well, that's all very well when live in Auzwax+wire wrote:I don't understand the motive to model a didgeredoo. I could understand if it was from a science 'Let us learn how a didgeredoo works], but why not just record a person playing a didgeredoo for your music. Where do you live? I live in Australia, where there is [obviously] a wealth of under utilised Aboriginal musicians who can play a Didgeredoo correctly and fully. Why not do some field recording.
I am a musician not a programmer, you may have guessed
I know many musicians here in the UK, but none of them are under utilised Aboriginals...Matt
Pi is exactly three
Pi is exactly three
-
- KVRist
- 81 posts since 28 Nov, 2005
I suppose I am curious [sorry to be going off on a tangent] but what is the motive for designing such a thing?
I follow Harry Partch/Sonic Youth in my approach, and design instruments to solve musical requirements. Necessity is the mother of invention.
If you have a song you want didg on, could you get someone to learn to play it? [note, there are cultural beliefs attached to playing the didgeredoo]
that said it seems like a fascinating project, that appears quite simple on the surface
I follow Harry Partch/Sonic Youth in my approach, and design instruments to solve musical requirements. Necessity is the mother of invention.
If you have a song you want didg on, could you get someone to learn to play it? [note, there are cultural beliefs attached to playing the didgeredoo]
that said it seems like a fascinating project, that appears quite simple on the surface
-
- KVRAF
- 8389 posts since 11 Apr, 2003 from back on the hillside again - but now with a garden!
I did a track a while back in which I used a default Synth1 patch (either brass string or lowstrings, something like that), and ran it through Classic Auto-Filter synced to semi-quavers.
Whilst not a real didg by any means, in the mix it made a reasonably exciting electric didg for someone with an extraordinary lung capacity.
http://www.last.fm/music/Duncan+Parsons/_/Gonville
DSP
Whilst not a real didg by any means, in the mix it made a reasonably exciting electric didg for someone with an extraordinary lung capacity.
http://www.last.fm/music/Duncan+Parsons/_/Gonville
DSP
-
- KVRian
- 1002 posts since 1 Dec, 2004
Hmm, I suggest the following model:
1) Resonant delay/pipe/waveguide/combfilter
Nothing complicated, just use a 1 sided one, with a filter inside the feedback loop (otherwise it sounds kinda metallic). This is the principal means with which you can influence the instrument's shape/makeup (by adjusting the delay length, feedback, and filter).
2) Brass lip simulation
The output from the delay (and probably an impulsion and a tiny bit of noise) gets fed into this. It's simply a resonant bandpass filter with about the same resonant frequency as the lips.
3) Nonlinearity
The output from the lips gets fed in here. Here, you want to basically turn a sine wave (the output of the lip simulation) into a series of impulses. This is where you'll want to come up with your own algorithm. You also absolutely want to limit the signal amplitude here, or else the model will "explode" and be unstable.
4) Vocal tract/wah wah simulation
The output from the nonlinearity gets filtered again. You might want to have 2 or 3 bandpass filters in parallel here, since the voice has multiple resonances. Here, you'll want to look at a document about vowel formants, it's all in there.
5) Multiply by blowing pressure
To control the flow of air, you have to multiply it somewhere. The output from the vocal tract goes in here, and the output from here goes back into the resonant delay. This means the whole model has 2 feedback loops (the delay's internal feedback, and the loop that goes through the lip simulation and back into the delay).
6) Output filters
The output from the delay line goes in here. You want to at least have a lowpass and highpass filter at the output to shape the sound.
7) Reverb
Adds a lot of "realism" to the sound.
1) Resonant delay/pipe/waveguide/combfilter
Nothing complicated, just use a 1 sided one, with a filter inside the feedback loop (otherwise it sounds kinda metallic). This is the principal means with which you can influence the instrument's shape/makeup (by adjusting the delay length, feedback, and filter).
2) Brass lip simulation
The output from the delay (and probably an impulsion and a tiny bit of noise) gets fed into this. It's simply a resonant bandpass filter with about the same resonant frequency as the lips.
3) Nonlinearity
The output from the lips gets fed in here. Here, you want to basically turn a sine wave (the output of the lip simulation) into a series of impulses. This is where you'll want to come up with your own algorithm. You also absolutely want to limit the signal amplitude here, or else the model will "explode" and be unstable.
4) Vocal tract/wah wah simulation
The output from the nonlinearity gets filtered again. You might want to have 2 or 3 bandpass filters in parallel here, since the voice has multiple resonances. Here, you'll want to look at a document about vowel formants, it's all in there.
5) Multiply by blowing pressure
To control the flow of air, you have to multiply it somewhere. The output from the vocal tract goes in here, and the output from here goes back into the resonant delay. This means the whole model has 2 feedback loops (the delay's internal feedback, and the loop that goes through the lip simulation and back into the delay).
6) Output filters
The output from the delay line goes in here. You want to at least have a lowpass and highpass filter at the output to shape the sound.
7) Reverb
Adds a lot of "realism" to the sound.
-
- Banned
- 12367 posts since 30 Apr, 2002 from i might peeramid
that's nice, having tim brooke taylor sing on 'you set me free' with youduncanparsons wrote: http://www.last.fm/music/Duncan+Parsons/_/Gonville
ditch the nonlinearity and use mass springs and an offset, aperture*aperture for transmission coeff
you come and go, you come and go. amitabha neither a follower nor a leader be tagore "where roads are made i lose my way" where there is certainty, consideration is absent.
-
- KVRist
- 291 posts since 2 Aug, 2007
-
- KVRer
- 13 posts since 13 Jul, 2006 from traveling usa
what an interesting discussion. as both a didj player and recording guy it is fascinating on many levels. the reasons for simulating a didj are probably similar to the reaosns we would want to simulate a guitar or a cello or really anything, for that matter. everyone has their own geek, and i wish i could play guitar or cello, but i can't; i know tons of great musicians who can, but i still can rationalize a creative need to be able to control these sounds myself.
seems like the dije has but a few basic sounds, and a few ways to alter them, resulting in a large # of possible combinations.
since didjes drone, it seems easy enough to use a few oscillators to make the basic sounds. you could find a way to adjust the even/ odd harmonics to adjust "edgy vs sweet", then add pitched pulses with adjustable ADSR to simulate the changes of air speed for diaphragmatic play, then add a vocoder with some filters and oscillators to add the vocalization aspect of didj playing. add some loud quick rhythmic inhale noises and that's at least a start...
the didj seems like a very complicated instrument, but isolating the various aspects of the sound producing mechanism may make it simpler. embouchure, breath, vocal chords (to oversimplify)
i have seen a few wavefoms of didj and they do not seem as harmonically complex as a cello...
i would love to hear what you come up with!!
seems like the dije has but a few basic sounds, and a few ways to alter them, resulting in a large # of possible combinations.
since didjes drone, it seems easy enough to use a few oscillators to make the basic sounds. you could find a way to adjust the even/ odd harmonics to adjust "edgy vs sweet", then add pitched pulses with adjustable ADSR to simulate the changes of air speed for diaphragmatic play, then add a vocoder with some filters and oscillators to add the vocalization aspect of didj playing. add some loud quick rhythmic inhale noises and that's at least a start...
the didj seems like a very complicated instrument, but isolating the various aspects of the sound producing mechanism may make it simpler. embouchure, breath, vocal chords (to oversimplify)
i have seen a few wavefoms of didj and they do not seem as harmonically complex as a cello...
i would love to hear what you come up with!!
-
- KVRAF
- 8389 posts since 11 Apr, 2003 from back on the hillside again - but now with a garden!
[OT]
And I'd like to see you hit that top E when the harmonies hit the last time... I didn't
[/OT]
DSP
I wasn't aware that the Goodies had translated to the US - or have you been listening to "I'm Sorry I haven't a clue" on the BBC website. Oh hang on, there's an element of brit about you somewhere isn't there...xoxos wrote:that's nice, having tim brooke taylor sing on 'you set me free' with youduncanparsons wrote: http://www.last.fm/music/Duncan+Parsons/_/Gonville
And I'd like to see you hit that top E when the harmonies hit the last time... I didn't
[/OT]
DSP