[PD] backwards ead~

[Stephan Elliot Perez]

Greetings,
     Well, I used two [vline~] objects and it works fine, but only when I
set the right input one of the one [expr~] to 0 and the other to 1, which
both lead to division through 0 and the corresponding error message.  I
believe you mentioned this problem, Mr. Farnell, regarding your own
patches.  To what complications does this problem lead?  Does anyone know
how I could modify the equation to be rid of it?  Why does the patch
nonetheless work?

Many thanks,
Stephan
On Sat, Jun 4, 2011 at 12:54 PM, Andy Farnell <padawan12 at obiwannabe.co.uk>wrote:

>
>
> [vline~] is versatile :) It can be used to solve many problems
> with envelopes. Also, don't be scared to use two vlines if
> it makes the problem easier to understand, their good time accuracy
> ensures they will do what you expect most times.
>
>
> On Sat, 4 Jun 2011 01:22:34 +0200
> Stephan Elliot Perez <dreamoftheshoreofanotherworld at gmail.com> wrote:
>
> > Thanks for your response.  I tried to apply the branching principal using
> > the equation from the [exact-ead~] by having the envelope go to 1 and
> then
> > to 2 instead of 0, using min 1 and max 1 to create a branch, using an
> > expression to convert the ascending numbers over 1 into descending
> numbers
> > under 1, and then jumping to 0.  I think the problem is that, during the
> > switches, two 1s are sent at the same time, leading to a 2 (as seen in
> the
> > graph), where I actually need a 0...
> >
> > -Stephan
> >
> > On Mon, May 30, 2011 at 8:34 PM, Andy Farnell <
> padawan12 at obiwannabe.co.uk>wrote:
> >
> > >
> > >
> > >
> > > You might be able to easily get that behaviour by
> > > quickly editing the example I gave you.
> > >
> > > The maths is really geometry.
> > >
> > > There are a few things that can be done as time domain
> > > transforms when thinking about envelopes and suchlike
> > > in this way.
> > >
> > > 1) Flip it around zero with [*~ -1]
> > > 2) Invert arithmetically wrt 1.0 using [sig~ 1][-~]
> > > 3) Get the [min~] or the [max~] wrt another value
> > > 4) Clamp at a value using [clip~] ... is special case of (3)
> > > 5) Shift by an amount using [-~] or [+~]
> > > 6) Scale by some factor with [*~]
> > >
> > > IIRC the idiom for a two stage envelope like that is
> > > to use [min~]/[max~] to create a split point and treat
> > > each of the two branches differently (you can do piecewise
> > > waveform construction the same way).
> > >
> > > If you want time symmetry then have the [vline~] go to 1.0
> > > and back to [0.0] and just use one of the branches.
> > >
> > >
> > >
> > > On Mon, 30 May 2011 18:45:55 +0200
> > > Stephan Elliot Perez <dreamoftheshoreofanotherworld at gmail.com> wrote:
> > >
> > > > Greetings,
> > > >      Thank you for your responses.  I tried your suggestion, hardoff,
> and
> > > > the result is the attached [ead-reverse~].  Unfortunately, the
> expression
> > > > behaves differently with this [vline~]-construction as with the
> > > [phasor~].
> > > > Here, if 1 is entered into the right input of [expr~], the result is
> an
> > > > envelope whose steepness becomes exponentially smaller while
> ascending
> > > and
> > > > exponentially greater while descending (if near 0, the opposite form
> is
> > > > produced and values between 1 and 0 produce a divided, confused
> form). I
> > > > however wish to produce an envelope that becomes exponentially
> steeper
> > > both
> > > > ascending and descending.
> > > >      My problem with this [expr~] as well as with Mr. Farnell's
> patches
> > > is
> > > > that I do not quite understand the math behind them.  Any ideas?
> > > >
> > > > Best regards,
> > > > Stephan Elliot Perez
> > > >
> > > >
> > > > On Fri, May 27, 2011 at 5:54 PM, hard off <hard.off at gmail.com>
> wrote:
> > > >
> > > > > instead of the [phaser~], send a [line~] (or [vline~]) signal.
> > > > >
> > > > > for example:
> > > > >
> > > > > [1, 0.5 500, 0 1000 500(
> > > > > |
> > > > > [vline~]
> > > > >
> > > > >
> > > > > like most of my patches, i don't think i originally made that one,
> > > someone
> > > > > else posted it, and i just copied.   ;)
> > > > > must have been a long time ago though, cos i barely remember it.
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > On Sat, May 28, 2011 at 12:24 AM, Stephan Elliot Perez <
> > > > > dreamoftheshoreofanotherworld at gmail.com> wrote:
> > > > >
> > > > >> Greetings,
> > > > >>        I wish to produce a wave form with the opposite form of the
> > > normal
> > > > >> ead~ wave (the curves become exponentially steeper instead of
> > > flatter),
> > > > >> which I can achieve by entering a negative number into the phasor
> in
> > > hard
> > > > >> off's exact-ead~ patch, which I downloaded from the archive.
>  However,
> > > I
> > > > >> still want to be able to independently change the length of the
> > > ascent,
> > > > >> decline, and distance between waves as is possible with ead~.
>  Does
> > > anyone
> > > > >> know how I could do this, or possibly have access to an
> > > abstraction-version
> > > > >> of ead~?
> > > > >>
> > > > >> Best regards,
> > > > >> Stephan
> > > > >>
> > > > >>
> > > > >>
> > > > >> _______________________________________________
> > > > >> Pd-list at iem.at mailing list
> > > > >> UNSUBSCRIBE and account-management ->
> > > > >> http://lists.puredata.info/listinfo/pd-list
> > > > >>
> > > > >>
> > > > >
> > >
> > >
> > > --
> > > Andy Farnell <padawan12 at obiwannabe.co.uk>
> > >
> > > _______________________________________________
> > > Pd-list at iem.at mailing list
> > > UNSUBSCRIBE and account-management ->
> > > http://lists.puredata.info/listinfo/pd-list
> > >
>
>
> --
> Andy Farnell <padawan12 at obiwannabe.co.uk>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20110607/9384ced4/attachment.htm>