I found this page http://www.geocities.com/tpe123/folkurban/synthstick/synthstick.html

And it gave me an idea i could use for a uni project next year(most of my projects are pretty open, so ill be able to make it fit)

Basically the idea i had was to set up one or more of these ribbon controls into pd, and by wiping across it being able to playback a file with [tabread4~]
(I'm actually considering building a few musical toys which can all be used together)

My real issue right now though, is input
Considering the fact that audio is sampled at 44100 samples a second, converting the output to midi seems kind of pointless to me because there will be no where near enough range.

But im thinking, this is basically just a giant variable resistor right?
So I should be able to run something through it (say and audio signal, or even just a stable current should work?) then get the amplitude readout using [fiddle~] or something similar.

The only other way I can think is including an oscillator in it like the one on that page, and using pitch as a control.

Another idea i had, involved using a small motor(run by a hand crank or something) to generate a signal which i could pick up in pd.

I'm thinking that by turning the motor, i should be able to generate a different pitch depending on speed, and this could be used.

Any sort of ideas you want to chuck at me in the way of how to get this input working would be appreciated, sometime this week i need to head down to an electronics store and get experimenting.

Hi God,

Seems to me that you'd be wanting to use a midi-board or some sort to route such controllers into Pd/your computer.

There are all sorts of them sold readymade or as kits, of which I think Arduinho seems to be pretty highly regarded by some folks in this forum.

Keep well ~ Shankar

ShankarBaba, the problem with that being, as I mentioned, I don't think a range of 128 will be enough for what I want, considering that i want smoother control over the file.

I will look into Arduinho though.

no h in arduino. and yes..that's what you want

The technique to understand and use here is interpolation.

If you have 128 input steps, and perhaps one second of audio giving you 44100 steps to map onto, then you need about 345 steps for each input increment.

There's many ways to do this that give you a smooth table index from a limited input resolution.

Maybe the easiest is [line~]

You can also use a combination of

[sig~]
|
[lop~]

to get a smoothed control line.

You will discover the limitations of these ways, then maybe want to look at more complex methods that take the differential of the input to set the slope of the line.

Isn't it easiest to modify the range-value of a controller input by just dividing or muliplying with an appropriate number? As in:

[ctlin 1]
|
[/ 345] or whatever

It certainly worked for me in something I did sometime ago,.. but yes, I'm a permanent newbie with this stuff.

Keep well all ~ Sssssssssss

What will happen that way Shankar is that even though the divided signal is mapped onto the right range it will still jump between discrete values that are the resolution of the control input. This happens because we have a limited number of integer values in the input, there is nothing between 0 and 1 or 126 and 127.

So when mapped we will see, for example
0 will be 0
1 will be 345
2 will be 690

and so on....

We would like the table index to move over all the numbers between 0 and 345 or 346 and 690 at the correct rate. That's where putting a [line] object in helps.

Thank you Obi.

That is fascinating, and makes a lot of sense, as with all things Pd.

However, if you'll pardon my limited knowledge in these matters, how come a control input divided by 127 seems to give a smooth transition through a slider valued from 0 to 1?

Also, as an unrelated general question here, has anyone been finding that "paged2.googlesyndication.com" has lately been really slowing down access to these pages?

Keep well all ~ Shankar

Hopefully this graph will help explain it Shankar,

Firstly, in the case of this graph, lets say that we know that our data will lead to a perfect curve.

The red nodes, would be the scaled down points you would get from something like scaling 127 into 1.
The black line follows the path of that data. You can see because of limited reference points, we dont end up with a smooth curve, we are skipping whats inbetween and connecting what we do know.

The blue nodes, would be something scaled up, like scaling 127 into 400.
Now, our reference points are further apart, so we have less data to make our curve with.
The green line follows this curve, which isn't too much of a curve because we have even less data.

Basically, what im saying is, that the reason scaling 127 into 1 works, is because you have a higher resolution of data, more points to reference, so you get a smoother transition.

I guess a better example to use in this case would be steps though.

sorry, just glanced thru the msgs here, but if you use arduino, none of that really matters because the raw data from the controller goes str8 into pd. there is no midi conversion to 128 steps.

Thank you God,

I see now that I should have understood that from Obi's reply ~ sorry for the bother.

By the way, in case you'd like to make yourself a simple midi-board, do pop in on http://tomscarff.tripod.com/. Whereas I do have a couple of sweet Faderfox microcontrollers, I made myself the one at http://tomscarff.tripod.com/anal2midi/analogue_inputs_to_midi_out.htm for my "performance" in CeC & CaC 2007, because I needed to discover for myself, and then demonstrate before all that one can actually do such a thing here in Delhi,... instead of whining about what is or is not available locally.

And, Hey!, Hardoff! Looks like I'm going to have to also start learning about Arduino sometime, on top of all of this mind-bending stuff~:o/

Keep well ~ Shankar

PS God ~ I eventually implemented Tom's circuit into a guitar-strap, and if you like, I'd be happy to share with you the stripped-down circuitboard-design that I made for it. Its less than half the size of Tom's multipurpose board