Ok, I get the trick, it seems similar to  the one used to make delay line shorter than one block.
However, I still feel that an object "give-me-RMS enveloppe-on-bang" (for the last N samples or blocks) would appear to me an easier way to handle this case.
If it doesn't exist, I'll try to build something like this... one day !
Many thanks !
Raphaël

2014-12-15 20:40 GMT+01:00 IOhannes m zmölnig <zmoelnig@iem.at>:
On 12/13/2014 11:14 PM, Raphaël Ilias wrote:
> 3)
> Last but not least, the question I can't answer myself !
> When using multiple [env~] it isn't very clear for me which one will
> output first.
[...]


this sounds like the bog-standard¹ "how can i force the order of
execution in the signal domain" question.

the order of signal-object execution depends on the implicit
relation-ship between objects: if object A sends data to object B, then
B is guaranteed to be called after A.
you can use subpatches that pass through signals to do signal order forcing.

see attached patch.

gfmadsr
IOhannes

¹ many people do not know about it though

_______________________________________________
Pd-list@lists.iem.at mailing list
UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list