2014-12-19 5:26 GMT+01:00 i go bananas <hard.off@gmail.com>:
>I'm surprised about this "double bang", but I suppose this means "do compute one audio block and only one" (here 1024 samples).

actually, i have NO idea why i needed to do the double bang.  For some weird reason, a single bang was triggering the [switch~], but then the next bang did nothing.  Only every second bang was triggering.  So, i used [t b b] just to send 2 bangs and trigger every time.  

if you remove one cable from one outlet of the [t b b], you will see what i mean. 

not sure if there's a bug in the implementation, or what, but yeah...seems weird. 


yes, i tried to use your trick using only one bang, and behavior was *not always* the same (sometimes seemed not switching...
however, i use the following and it works for me :

[bang~]
 |
[del 0]
 |
[0 (    [1 (
 [     /
[switch~]


clicking the [1 ( computes one block and then DSP get switched off by [bang~]
 

>I wonder if switching the DSP off does mess up with overlapping (as i understand "overlapping with previous audio block").

 sorry, i don't actually know how the overlap is implemented, but I did some trial and error tests, sending various signals through the switched off subpatch, and through a subpatch with no [switch~] object.   The outputs of each [env~] in the different subpatches seems to be the same.  



yes i'm curious too...
let's say i want to get the RMS amplitude of a 2048-samples "grain" of audio.
if i use [env~ 2048] however i'm using a 2048-samples window length in argument, i get two RMS amplitude results spaced by 1024 samples. so i'm sitting there and wondering : am looking to the RMS amplitude of first and second half of my 2048 samples ?