julianksw@gmail.com

I've been reading the MAXMSP manual and deligently studying the tutorials for a few days now, and it occured to me that I may not be able to achieve what I had set out for. If you can, please help me answer this question: how feasible is my project?

The project is about tranquility in urban spaces. I will have a microphone that takes in urban sounds, which will be converted (in real time) into other soothing sounds found in nature. So an audience can listen to sounds that relate to his/her chaotic environment but still feel inner peace. It may work in the following way:

So what I imagine the work that my project will involve is :

1) to find out the sound frequencies of my input triggers (buses, sirens, etc)

2) to record/get samples of desired sounds

3) write a patch that says "when you hear [this], play [that]" (I wish only it was this simple)

I work full time and am studing for a masters degree (which has led me to this wonderful little project), I will be spending my nights and weekends on it. Seeing that I am a total beginner, and the project is due in 3-4 weeks time, is the scope of work much larger than I should want to take on? How feasible is it to do what I intend in MAXMSP?

Further to this I would also like to use the new sounds to generate visual overlay on top of urban images (if time permits). Is Jitter the software I would want to explore for this purpose?

Sounds of passing buses = water dripping
Sounds of sirens = river flowing
ETC.

So what I imagine the work that my project will involve is :
1) to find out the sound frequencies of my input triggers (buses, sirens, etc)
2) to record/get samples of desired sounds
3) write a patch that says "when you hear [this], play [that]" (I wish only it was this simple)

how-feasible-is-my-project

The short answer to your "how feasible" question is probably "-ish."

> Sounds of passing buses = water dripping

Assuming that this is some kind of site-specific installation, your 

easiest approach is probably to categorise by frequency, amplitude, 

envelope and direction. There are various crude techniques for 

frequency analysis (bandpass filtering being the obvious one; you 

could get into the FFT world for more sophistication).

If you split your input into a handful of frequency bands, and also 

program in a few level thresholds, you get a simple frequency/level 

matrix into which you can drop your replacement sounds.

Add some simple envelope analysis and you've got a third dimension of 

attack/decay time, and/or average level over a number of seconds, 

which would distinguish between (say) buses and gunshots. (How many 

gunshots you'll encounter depends where you live, I guess.)

Add another dimension by using more than one directional microphone, 

and categorising sounds by their relative levels in the microphones 

(or, to be fancy, calculate panning changes).

> 3) write a patch that says "when you hear [this], play [that]" (I 

The trick to installation works is to design some kind of abstract 

control engine with enough degrees of freedom to be interesting. Once 

you have that, you can implement and test the analysis and synthesis 

> Further to this I would also like to use the new sounds to generate 

> visual overlay on top of urban images (if time permits). Is Jitter 

> the software I would want to explore for this purpose?

I'd say not in 3-4 weeks if (i) you're a beginner, (ii) you're not 

working on this full time and (iii) you've got the audio side to do 

> Sounds of passing buses = water dripping
> Sounds of sirens = river flowing

Assuming that this is some kind of site-specific installation, your  
easiest approach is probably to categorise by frequency, amplitude,  
envelope and direction. There are various crude techniques for  
frequency analysis (bandpass filtering being the obvious one; you  
could get into the FFT world for more sophistication).

If you split your input into a handful of frequency bands, and also  
program in a few level thresholds, you get a simple frequency/level  
matrix into which you can drop your replacement sounds.

Add some simple envelope analysis and you've got a third dimension of  
attack/decay time, and/or average level over a number of seconds,  
which would distinguish between (say) buses and gunshots. (How many  
gunshots you'll encounter depends where you live, I guess.)

Add another dimension by using more than one directional microphone,  
and categorising sounds by their relative levels in the microphones  
(or, to be fancy, calculate panning changes).

> 3) write a patch that says "when you hear [this], play [that]" (I  
> wish only it was this simple)

The trick to installation works is to design some kind of abstract  
control engine with enough degrees of freedom to be interesting. Once  
you have that, you can implement and test the analysis and synthesis  
stages in isolation.

> Further to this I would also like to use the new sounds to generate  
> visual overlay on top of urban images (if time permits). Is Jitter  
> the software I would want to explore for this purpose?

I'd say not in 3-4 weeks if (i) you're a beginner, (ii) you're not  
working on this full time and (iii) you've got the audio side to do  
first.

Nick Rothwell / Cassiel.com Limited
www.cassiel.com
www.myspace.com/cassieldotcom
www.loadbang.net


Have you looked at Noah Vawter's 'ambient walkman' work at MIT?

http://web.media.mit.edu/~nvawter/thesis/

> I've been reading the MAXMSP manual and deligently studying the

> tutorials for a few days now, and it occured to me that I may not

> be able to achieve what I had set out for. If you can, please

> help me answer this question: how feasible is my project?

> The project is about tranquility in urban spaces. I will have a

> microphone that takes in urban sounds, which will be converted

> (in real time) into other soothing sounds found in nature. So an

> audience can listen to sounds that relate to his/her chaotic

> environment but still feel inner peace. It may work in the

> So what I imagine the work that my project will involve is :

> 1) to find out the sound frequencies of my input triggers (buses,

> 2) to record/get samples of desired sounds

> 3) write a patch that says "when you hear [this], play [that]" (I

> I work full time and am studing for a masters degree (which has

> led me to this wonderful little project), I will be spending my

> nights and weekends on it. Seeing that I am a total beginner, and

> the project is due in 3-4 weeks time, is the scope of work much

> larger than I should want to take on? How feasible is it to do

> Further to this I would also like to use the new sounds to

> generate visual overlay on top of urban images (if time permits).

> Is Jitter the software I would want to explore for this purpose?

   http://web.media.mit.edu/~nvawter/thesis/

>
> To all MSP maestri,
>
> I've been reading the MAXMSP manual and deligently studying the
> tutorials for a few days now, and it occured to me that I may not
> be able to achieve what I had set out for. If you can, please
> help me answer this question: how feasible is my project?
>
> The project is about tranquility in urban spaces. I will have a
> microphone that takes in urban sounds, which will be converted
> (in real time) into other soothing sounds found in nature. So an
> audience can listen to sounds that relate to his/her chaotic
> environment but still feel inner peace. It may work in the
> following way:
>
> Sounds of passing buses = water dripping
> Sounds of sirens = river flowing
> ETC.
>
> So what I imagine the work that my project will involve is :
> 1) to find out the sound frequencies of my input triggers (buses,
> sirens, etc)
> 2) to record/get samples of desired sounds
> 3) write a patch that says "when you hear [this], play [that]" (I
> wish only it was this simple)
>
> I work full time and am studing for a masters degree (which has
> led me to this wonderful little project), I will be spending my
> nights and weekends on it. Seeing that I am a total beginner, and
> the project is due in 3-4 weeks time, is the scope of work much
> larger than I should want to take on? How feasible is it to do
> what I intend in MAXMSP?
>
> Further to this I would also like to use the new sounds to
> generate visual overlay on top of urban images (if time permits).
> Is Jitter the software I would want to explore for this purpose?
>
> Thanks,
>
> Julian
>


How feasible is my project?