oops, meant to send tis to the list a well as mike berry...

pete

|
|
|
|
v

pete moss wrote:

> i have had many problems in the past with the repeating sequence.  and i would love to
> have a ugen which generates a longer sequence.  a period of a billion years sounds
> acceptable.  problem is, as you have suggested, that i have sounds that take advantage
> of the short period of rand.  it might be nice to have another ugen (whats one more!) to
> generate this extra long period rand for special use.  perhaps we could call it elprand
> or something.  i also agree with tolve that we need a way to use the clock to seed the
> random number generator.  this would be an easy addition perhaps requiring one line of
> code or as a new optional param in the instrument.
>
> pete
>
> Mike Berry wrote:
>
> > A comment here on rand.  The period of rand in csound on most platforms is
> > 65536, or less than 2 seconds of white noise at 44.1 kHz.  The repeating is audible.
> >         I have code for a random number generator which will repeat for most seeds
> > about every 1 billion years at 44.1 kHz.  The low end is about 100000 years.
> > I would be happy to pass this on to jpff.  The question is: Do we replace rand
> > with the new code or make it a new ugen, to preserve the flawed character of
> > rand for old scores?
> > --
> > Mike Berry
> > mikeb@nmol.com
> > http://www.nmol.com/users/mikeb