Hi there,
this is my first post on this forum but I'm working with PD already.

I'd need to customize a synthesizer patch I designed.
I'd need to add a feature to route adsr as the same time as the pitch to each voices.

example:
I'm sending a long attack envelope & the pitch 64.. it produces the midi note 64 with a long attack
But if I send just after that a short attack envelope & the pitch 45, it will produce a short attack sound related to the midi note 45.

How can I do that ?

This isn't very clear. Can you post a simple patch that illustrates the problem?

I'm sorry.

more clear:
when I used to make polyphonic synth, I'm using the native poly object which handles the voices routing.
I'm often using adsr concept in my voice(s) with send/receive feature in order to control all adsr with only 4 parameters "outside" of the voices.

Of course it works fine, but in that case, I cannot choose an envelope for each note fired to the voices.

I need to be able to send adsr + pitch + velocity to my voices, the whole concept handled with the poly object of course.

usually (traditionally) you'd just send note on / off messages (ie, velocity >1 , velocity 0) and have individual envelope generators inside each voice triggered by those messages.

considering that the signal-based component of a digital envelope generator is nothing more than a [line~] object, i can't really see any way that you'd be better off by 'sending' the adsr signal from outside each voice abstraction.

I needed that.
I just did it.

indeed, my incoming machine won't be usual

unless, of course, you want to have DIFFERENT envelope shapes for the 4 voices? is that what you're saying>

if you want to do that, then you might be better off trashing the [poly] object and making your own version as an abstraction, to deal with the extra parameters.

OR, what you may be able to do is exploit the existing poly object to work in your favour. I just checked, and it is NOT limited just to 0-127 range

so, you could encode all your data within the single float that is sent for velocity. a basic example would be something like:

A = 3
D = 7
S= 2
R= 9

actual velocity = 100

then, you could join all those values together as:

1003729

and then use [mod] [div] [%] or whatever to break them all apart again once they are inside your voice.

obviously this only gives you very coarse resolution for your ADSR controls, but it may be enough for your purposes?

it could be enough.
I made that with some pack/unpack & trigger.
(posting it as soon as it is clean)

here it is
http://julienbayle.net/2012/04/26/polyphonic-synthesizer-with-poly-adsr

ah yeah, your solution is very elegant, nice one

thanks mod.
I could have written that for yours

I used this kind of packing (yours, I mean) for many hardware projects where I was limited by data flows.