Physical Modelling - Sitar

[MisterEko]

Hello,

I am trying to physically model the Sitar for my masters thesis. I intend making it available as a plugin using Pluggo as I am using MaxMSP. I was wondering if someone could help me with my patch. I am trying to develop the string as a bi-directional digital waveguide but I am unsure I have done so correctly.

Can anyone help?

Cheers,

David

[Winstontaneous]

I am far from an expert but I'm very interested in PM--just got Tassman 4 last week. Subscribed!

I've been reading the PM chapter in Roads' "Computer Music Tutorial" and slowly working through Benade's "Horns, Strings, and Harmony".
Also reading Smith's "Physical Audio Signal Processing" at https://ccrma.stanford.edu/~jos/pasp/pasp.html
I grasp the concepts but need to brush up on my math skills.

Out of curiosity where are you studying?

edit: you might have better luck posting this in the DSP forum here...that's where the real smartypants hang out!

[madronalabs]

Hello,

I am trying to physically model the Sitar for my masters thesis. I intend making it available as a plugin using Pluggo as I am using MaxMSP. I was wondering if someone could help me with my patch. I am trying to develop the string as a bi-directional digital waveguide but I am unsure I have done so correctly.

Can anyone help?

Cheers,

David

Hi David,

I think you may want to write your own MSP external to do this flexibly. One problem is that the 1-D waveguide will require two delay lines to be set up in a ring configuration, which is not really possible in MSP's directed graph model.

As you've probably discovered, you can use tapin~ and tapout~ to make a ring setup, but with a signal vector of delay in between the input and output. So as far as making a 1-D waveguide this should work if you allow for the fixed delay of one signal vector in each direction. Obviously this will restrict where on the waveguide you can place the pickups, and you will need to use a small signal vector if you want high frequencies.

I think string tension nonlinearity is important to the sitar sound. This may be harder to incorporate if you don't write your own object.

If you get one basic linear string working reasonably, here is some research on the kantele that might give you ideas how to improve your sitar sound:
http://www.acoustics.hut.fi/research/asp/kantele/

good luck,
Randy

[xoxos]

this doesn't really help you, i don't have any idea of max's facility hopefully you can put a multiorder allpass in your wg.

i'd like to create a vst sitar model.. what i don't have is a discrete form to model. i need photos, measurements, maybe even sounds, so i can calibrate waveguides

[herboriseria]

look at this:
https://co.native-instruments.com/index ... d=3399&L=1

[MisterEko]

I am far from an expert but I'm very interested in PM--just got Tassman 4 last week. Subscribed!

I've been reading the PM chapter in Roads' "Computer Music Tutorial" and slowly working through Benade's "Horns, Strings, and Harmony".
Also reading Smith's "Physical Audio Signal Processing" at https://ccrma.stanford.edu/~jos/pasp/pasp.html
I grasp the concepts but need to brush up on my math skills.

Out of curiosity where are you studying?

edit: you might have better luck posting this in the DSP forum here...that's where the real smartypants hang out!

I am studying for a masters in music technology in the University of Limerick, Ireland. Unfortunately there is no one on the faculty to help me with physical modelling though. So I am kind of struggling by myself.

[MisterEko]

Hello,

I am trying to physically model the Sitar for my masters thesis. I intend making it available as a plugin using Pluggo as I am using MaxMSP. I was wondering if someone could help me with my patch. I am trying to develop the string as a bi-directional digital waveguide but I am unsure I have done so correctly.

Can anyone help?

Cheers,

David

Hi David,

I think you may want to write your own MSP external to do this flexibly. One problem is that the 1-D waveguide will require two delay lines to be set up in a ring configuration, which is not really possible in MSP's directed graph model.

As you've probably discovered, you can use tapin~ and tapout~ to make a ring setup, but with a signal vector of delay in between the input and output. So as far as making a 1-D waveguide this should work if you allow for the fixed delay of one signal vector in each direction. Obviously this will restrict where on the waveguide you can place the pickups, and you will need to use a small signal vector if you want high frequencies.

I think string tension nonlinearity is important to the sitar sound. This may be harder to incorporate if you don't write your own object.

If you get one basic linear string working reasonably, here is some research on the kantele that might give you ideas how to improve your sitar sound:
http://www.acoustics.hut.fi/research/asp/kantele/

good luck,
Randy

Hi Randy,

Thanks for getting back to me. If I sent you my max patch would you be able to have a look at it for me? I have quite a bit of work done on it and I think I may be close.

Cheers,

Dave

[MisterEko]

this doesn't really help you, i don't have any idea of max's facility hopefully you can put a multiorder allpass in your wg.

i'd like to create a vst sitar model.. what i don't have is a discrete form to model. i need photos, measurements, maybe even sounds, so i can calibrate waveguides

Yip, I have one put in, I'm just not sure if I have the bi-directional wg implemented properly.

[SpicyBeni]

Hi,

I don't know anything about MaxMSP, but compared to other string instruments I would say that you should emphasize your work on the coupling between strings (via the bridge) to capture the sympathetic resonances which seem to be of top importance for the characteristic sound of the Sitar.

Cheers

[madronalabs]

Hello,

Hi Randy,

Thanks for getting back to me. If I sent you my max patch would you be able to have a look at it for me? I have quite a bit of work done on it and I think I may be close.

Cheers,

Dave

If it doesn't require any externals, I can take a quick look at it. Go ahead and email it to me at my name at madronalabs.com.

[JoeDeF]

...compared to other string instruments I would say that you should emphasize your work on the coupling between strings (via the bridge) to capture the sympathetic resonances which seem to be of top importance for the characteristic sound of the Sitar.

Cheers

Hi David,

While I, agree with SpicyBeni that string resonances and coupling are likely to be important to a successful model, I feel that the most critical part of the sitar sound is the bone plate on the main bridge, or jawari, which is like a gradual ramp that ends at the final termination point of the bridge. This is what gives the sitar that buzzy twang (no, I don't mean a Nashville twang!), which can sound somewhat like a highly resonant filter being swept.

Here's a page that might illustrate the instrument parts and geometry a bit:
http://chandrakantha.com/articles/india ... aking.html

I only know a little about this subject, but I will share my understanding just in case you find it to be helpful. The transverse waves traveling along the string impinge upon the jawari before they reach the termination (or witness) point, which in a typical stringed instrument is the place where the waves usually flip over and travel the other way upside down. In a sitar, before that happens, the greater excursions of the string interfere with the jawari, altering the shape of the transverse wave bulge (in a way I don't really understand or know how to model, unfortunately), increasing the higher partial content greatly at larger wave amplitudes, and not so much at smaller wave amplitudes. The jawari/string interaction at greater amplitudes also reduces sustain significantly, as the energy is transferred to those louder higher partials.

The reason I feel that this is important to the sitar sound is that other instruments that also have a gradual and imprecise termination (which is basically what a jawari is) often are said to have a sitar-like sound. One that comes to mind is the fretless electric bass; many players desire and strive to attain that sound, which is often called "mwah." I am mentioning that instrument in particular because there might be fretless bass models out there somewhere that model the imprecise termination, and if so, it might provide an algorithmic basis for your necessary jawari modeling.

You may have already known all of this, but I wasn't sure so I'm putting it out there. Good luck, and let us know how this comes along; I'd like to hear it when it's up and running,

Joe

[MisterEko]

...compared to other string instruments I would say that you should emphasize your work on the coupling between strings (via the bridge) to capture the sympathetic resonances which seem to be of top importance for the characteristic sound of the Sitar.

Cheers

Hi David,

While I, agree with SpicyBeni that string resonances and coupling are likely to be important to a successful model, I feel that the most critical part of the sitar sound is the bone plate on the main bridge, or jawari, which is like a gradual ramp that ends at the final termination point of the bridge. This is what gives the sitar that buzzy twang (no, I don't mean a Nashville twang!), which can sound somewhat like a highly resonant filter being swept.

Here's a page that might illustrate the instrument parts and geometry a bit:
http://chandrakantha.com/articles/india ... aking.html

I only know a little about this subject, but I will share my understanding just in case you find it to be helpful. The transverse waves traveling along the string impinge upon the jawari before they reach the termination (or witness) point, which in a typical stringed instrument is the place where the waves usually flip over and travel the other way upside down. In a sitar, before that happens, the greater excursions of the string interfere with the jawari, altering the shape of the transverse wave bulge (in a way I don't really understand or know how to model, unfortunately), increasing the higher partial content greatly at larger wave amplitudes, and not so much at smaller wave amplitudes. The jawari/string interaction at greater amplitudes also reduces sustain significantly, as the energy is transferred to those louder higher partials.

The reason I feel that this is important to the sitar sound is that other instruments that also have a gradual and imprecise termination (which is basically what a jawari is) often are said to have a sitar-like sound. One that comes to mind is the fretless electric bass; many players desire and strive to attain that sound, which is often called "mwah." I am mentioning that instrument in particular because there might be fretless bass models out there somewhere that model the imprecise termination, and if so, it might provide an algorithmic basis for your necessary jawari modeling.

You may have already known all of this, but I wasn't sure so I'm putting it out there. Good luck, and let us know how this comes along; I'd like to hear it when it's up and running,

Joe

Hi Joe,

Thanks for getting back to me. Nice one for the info. I'm currently working on getting the dynamically changing string length going at the moment and I think I nearly have it cracked. What your saying about jawari is the next step I'm going to approach. Another bag of headaches no doubt. I'll post up my max patch when it's done.

[Winstontaneous]

I am studying for a masters in music technology in the University of Limerick, Ireland. Unfortunately there is no one on the faculty to help me with physical modelling though. So I am kind of struggling by myself.

Did somebody say limerick?

Code: Select all

There once was a student named Dave
Who strove to suss out how sitars behave
The better to model
Via MAX at full throttle
And guided bi-directional waves

[MisterEko]

I am studying for a masters in music technology in the University of Limerick, Ireland. Unfortunately there is no one on the faculty to help me with physical modelling though. So I am kind of struggling by myself.

Did somebody say limerick?

Code: Select all

There once was a student named Dave
Who strove to suss out how sitars behave
The better to model
Via MAX at full throttle
And guided bi-directional waves

HAHAHA!! You legend!!!

[xoxos]

heck i'm postin' (w/o thinking about it..)

i'm sure you've sussed this, to solidify it for you i'll parrot, bidirectional is only necessary if you've got anything happening at the terminii, otherwise you might as well treat it as a single delay loop.

throwing primitive, reactional thinking at it, the jawari (thanks for link!) could at least be modeled as a FIR, at a fixed sample rate as the jawari is of a fixed length??