I had another look and using floats is still problematic in
single-precision; pulses won't take place at the exact BPM for low BPM
values.
The problem with the current ba.beat implementation is that it checks
against 0: if we decrease the second operand of the % operation, there is
no guarantee that the output will go through 0 when wrapping around. What
we could do is to check if the first difference is negative to cover the
more general case of wrapping around:
import("stdfaust.lib");
bpm = en.ar(20,10,1) * 2000;
beat(t) = (diff(ba.period(ba.tempo(t))) < 0) + impulse
with {
diff(x) = x <: _ - _';
impulse = 1 - 1';
};
process = beat(bpm) , beat(bpm);
Ciao,
Dr Dario Sanfilippo
http://dariosanfilippo.com
On Sun, 1 Aug 2021 at 16:12, Dario Sanfilippo <sanfilippo.da...@gmail.com>
wrote:
> Hi, Henrik.
>
> It could be due to some integer overflow in the ba.beat function; you
> could try that in double precision to see if it works better.
>
> Otherwise, here's another way that appears to be more robust:
>
> import("stdfaust.lib");
> imp = beat(en.ar(20,10,1) * 2000);
> beat(t) = diff(os.phasor(1, t / 60.0)) < 0
> with {
> diff(x) = x <: _ - _';
> };
> env = en.ar(0.00001, 0.05, imp);
> snare = no.noise * env * 0.1;
> process = snare , snare;
>
> Ciao,
> Dr Dario Sanfilippo
> http://dariosanfilippo.com
>
>
> On Tue, 22 Jun 2021 at 21:05, Henrik Frisk <fris...@gmail.com> wrote:
>
>> Hi,
>>
>> When I run the snippet below it's as if there are hickups in the stream
>> of pulses. These come at random but most commonly in the beginning of the
>> envelope. I have tried it compiled to SC code and to jack stand alone code
>> on both Linux and OSX as well as in the online compiler with the same
>> result.
>>
>> Is there a better way to achieve a smotth accelerando of pulses?
>>
>> imp = ba.beat(en.ar(20,10,1) * 2000);
>> env = en.ar(0.00001, 0.05, imp);
>> snare = no.multinoise(8) : par(i, 8, _ * env * 0.1);
>> process = snare;
>>
>> Thanks!
>> /Henrik
>> _______________________________________________
>> Faudiostream-users mailing list
>> Faudiostream-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/faudiostream-users
>>
>
_______________________________________________
Faudiostream-users mailing list
Faudiostream-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/faudiostream-users
