> > i'm not sure what you mean here. perhaps how many
> > different values can  you
> > express with a sample?
>
> I was worrying about the ability of 4 samples to represent
> a sine for  example, if my cps is 11000 and my audio rate
> 44000 ? So I thought I  might not catch well the concept
> of a sample.

Yes, four samples will represent the sine wave. When it is
played
back the ADC process lowpass filters the signal and the
squarish
wave becomes smooth. In fact two samples will represent a
sine
at the Nyquist as well.


>
> Victor Lazzarini 
> Date: Fri 12 May 2006  10:41:40 Europe/Zurich
> wrote:
>
> > This creates a wavetable that  does not 'join at the
> ends'.
>
> Can you explain a little more?

Well, if you look at its shape, a sine wave with 6.4 Hz will
not fully complete its
cycles in one second, there will be 0.4 of cycle that is
left 'hanging'.
This is the same thing that will happen in a wavetable,
there will be
 a portion of the cycle hanging at the end of the table.
That will not
join with the beginning. When the oscillator plays the wave
back it
replays a periodic wave with a funny shape that has an
abrupt
transition at its ends. Abrupt transitions imply very high
components,
so that is where the aliasing comes from. You have
inadvertedly
created a wave with high partials, even if you did not
specify
those as sine components in your f-statement.



> By the way, I listened to your "Timelines la" and found it
> beautiful.  Following my concern about partials, I was
> particularly interested by  the sort of echo that follows
> the first chord and is repeated  a few  times, also at the
> end of the piece. Did you realize it with the help  of a
> record of the guitar? It's a kind of sound I would like to
> produce  entirely with csound opcodes.

No it's all live processing. There is a delay and all sorts
of
spectral stuff in it. You can have a look at the CSD if
you'd
like.

Victor

>
> --
> Send bugs reports to this list.
> To unsubscribe, send email to
> csound-unsubscribe@lists.bath.ac.uk