I'm using co-ordinates from a grid to trigger a note. at the moment when a Button (co-ord) is pressed and held while a new Button press occurs, new button press on release will also cause the previous button press to stop ( despite it remains pushed).

Im trying to implement this block diagram where we receive a co-ord, check the condition. If true the co-ord is sent forward and stored as current co-ord. If false the co-ord is checked against current co-ord...

The thing is Im not sure how to store the co-ordinate values so as you would in a general-purpose program perhaps? Has anyone designed a keyboard application and had to do this?

Please find attached a flow chart and the pure data patch to implement to flow chart design.

many thanks

@Fauveboy You can assign an id to every button. If you have 8 rows you could do it with x * 8 + y. If there are also 8 columns you now have 64 datapoints that can have the value 0 or 1. From there you can go on. For storing the values you can use array for example.

@Fauveboy

Hi, this sends out the x y coordinates only when their state is 1. I suppose you could also do it with [poly].

store.variable.pd

I have a similar design implemented in c just for comparison to get the idea across

@Fauveboy

So, when the button is pressed (state == 1) you need the coordinates to be sent out and stored, but what do you need to happen when the state is 0 (= button release?) Do you just need that the coordinates are not sent through?

@weightless well what happens is that the new button press will need to, for example, trigger a note...and the button releases need to work independently. so that when a new button is pressed it doesn't cancel out the previously pressed buttons that my be sustained during.

@weightless maybe the coordinates in a 0 state arnt sent through but the that still needs to reach a button to tell it to turn off others buttons would continue sustaining?

@Fauveboy Well, I built your patch copying the flow chart...... never saved it..... and on testing the dreaded "stack overflow" put an end to all my work.
You can build a stack of notes like this........ that you can then access.....
stack.pd
although you only need a two level stack as I understand it?
David.
((Better "slow" example for your purposes....... stack2.pd ))

@Fauveboy

Ok I see, then I think it depends on whether each button (aka each set of coordinates) is assigned to a fixed note or not, because in that case rather than using this sort of logic you could use each unique pair of coordinates to trigger a specific note, and use the state value to trigger the velocity (0 = 0, 1 = 127, for example).

well the limitation there is that the grid is a monome and has 255 different buttons. the co ordinates are coming from serialosc. theres no velocity just on or off. not all the buttons will trigger notes some are sent to other areas of the patch to trigger an lfo signal for change a parameter.

for example in this screen shots we take values that meet a certain condition. if i press say 3 15 1 hold it ..press 5 15 1 thats fine 5 15 1 plays. But when i release 5 15 1 where i was holding 3 15 1 is lost.

what also happens is if i press any co-ord value that meet the conditions here but then press say 6 12 1 whatever job 5 15 1 was doing stopped?

@Fauveboy This cannot be solved by sending everything through one bottleneck because than there can only be one state on or off.

With Midi it would be very simple, let each button send a note-on when pressed and a note-off when released and you are finished.

If you want the sound processing in pd i guess you would need to build a polyphonic synth and send the on and offs to one of the synths. I haven't used the audio stuff in pd yet so i could be wrong on that.

I just realised I didnt link the other flow chart more specifically applied to the patch

@Fauveboy Sorry...... correction.......
Your osc messages come in as something like......?
button/xy 3 15 1 (down) and button/xy 3 15 0 (up)
button/trigger 1 (down) and button/trigger 0 (up)

If I understand your presses....... and (the result).....
What you posted looks like.....
press and hold 3 15 1 (note) ....... press 5 (new note)........ release 5 (nothing)

Is that true?
David.

@Fauveboy

if all button presses are going through the same logic gate, there is no way of keeping track of which coordinate is associated with which state, and as soon as a "state = 0" is received, 0 is sent to [s note1] regardless.

I think the best way is to somehow pair the state (on/off) with whatever value you want to pass through (which in the last picture you attached is only the x value?). Then you can send the list to the sound generating stuff and scale them, or convert them to midi and send them outside pd.

@whale-av the osc messages are correct yes:
...press and hold 3 15 1 ( button down), continue to hold while 5 15 1 happens. ( i think already in the process 3 15 1 is now cancelled out...if we release 5 15 1 and have 5 15 0, if 3 15 1 was still held it is now inactive ...

actually a different version of the problem here, thats more relevant perhaps.....with this current patch

any co ordinate such as 12 14 1 makes the condition output false. If a current true value is being used and a false value occurs it should't be stopping a currently active true value ...

@Fauveboy If you write the values to an array you can afterwords ask [array sum] if it is 0, which would mean that all buttons are 0.

a bit messy, but this seems to work: expr-trigger-test.pd

this is interesting ....Im trying describe to myself the logic thats taking place, to understand how its working exactly. Would you be able to correct my logic, it gets hard to follow after a few steps....
many thanks
starting with :

get co ords

r x is sent to store in three float objects
and the condition equal to - 1?

co word are test with equal to 3

what do we test for next? I guess if equal to three is true or false.

but before we can send the result of == 3 forward….we do something with values from r x, what happens are the -1 condition?
r x (condition) is multiplied with r y (condition) ?

then we either get 0 or 1.

we can then send the result of all the co-ord added together forward?
we then bang the r x values stored in two of the floats and a zero to spigot 1?

but if spigot 1 is zero we bang a -1? to an earlier condition….

I wanted to send this in the mean time while i look at this again but I think I’m describing it in a convoluted manner?

or perhaps a flow diagram?

@Fauveboy Having slept on this problem, I thought you might like this solution.
It is how I would have approached it in Pd, without the logic........

logic.zip
Extended and Vanilla.

You can save the patch with the grid drawn...... but then must be careful not to create the grid again..........
The drawing can be automated with a [loadbang] of course.

It is probably possible to attach an age to the xytog abstractions as well, so that "updates" are sent in the correct reverse order when there are many....... as all current xyz values are stored!
David.