Thank you both for your replies. I might give William Brent's objects a shot, though the explanations on his website look very complex to me...

Cheers,

Pierre.

2015-05-27 2:49 GMT+02:00 Jaime E Oliver <jaime.oliver2@gmail.com>:
Hi Pierre,

If I remember correctly, William Brent in his timbreID library does onset detection on tables. I cannot remember now how he does it exactly, but I hope this might prompt him to reply…?

In any case, all of his timbre analysis objects have real-time signal and non-signal versions...

best,

J


On May 26, 2015, at 12:52 PM, Miller Puckette <msp@ucsd.edu> wrote:

> Unfortunately there's no way - the attack detection in bonk~ depends on
> a succession of frames and not just a single one.  (It might have made sense
> for bonk~ at least to be able to do its spectral analysis from an array but
> I think most uses of bonk~ are for the attack detection.)
>
> cheers
> Miller
>
> On Tue, May 26, 2015 at 10:08:54AM +0200, Pierre Massat wrote:
>> Dear list,
>>
>> Is there a way to analyze an array containing audio data with bonk~ ? I
>> believe sigmund~ is capable of this.
>> (By the way, is there any pun intended with bonk ?)
>>
>> I'm trying to build a sample slicer and right now I just read the array at
>> audio rate and feed the output to bonk~ . This is both slow (is the sample
>> is long) and not very reliable (for some reason the sets of attacks
>> detected in the same sample are sometimes different from one analysis to
>> the next).
>>
>> Cheers,
>>
>> Pierre.
>
>> _______________________________________________
>> Pd-list@lists.iem.at mailing list
>> UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
>
>
> _______________________________________________
> Pd-list@lists.iem.at mailing list
> UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list