SV: [sdiy] sawtooth phase shifter

[Magnus Danielson]

From: karl dalen <dalenkarl at yahoo.se>
Subject: SV: [sdiy] sawtooth phase shifter
Date: Tue, 9 Aug 2005 03:58:48 +0200 (CEST)
Message-ID: <20050809015848.99399.qmail at web25509.mail.ukl.yahoo.com>

Harry och Masen,

> --- harrybissell <harrybissell at prodigy.net> skrev:
> 
> > My project for the week is a sawtooth phase shifter...
> > more like a waveform animator.  It has three stages,
> > each driven from a Schmitz (tm) three phase VCO.
> > 
> > It sums the shifted sawtooths, the pulses that form them.
> > You can pan between pulse and saw.
> > 
> > I'm driving it from the "muffy" guitar synthesizer.  It has
> > sawtooth outputs at F and 1/2 F ~or~ 2F.
> > 
> > Sounds... how can I put this delicately ... underwhelming.
> > Like the 'hex fuzz' the more stages you sum, the more
> > subtle the effect.
> > 
> > One neat thing... If you enable both F and 1/2F ~or~
> > F and 2F at the same time (now there is a sum going into
> > the phase shifter)... you get a pronounced "fifth" sound.
> > This is extremely cool... as this is generated from the
> > direct 'waves' output of the guitar synth (so there is no
> > VCO involved ie, no delay... and you get a tracking fifth).
> > 
> > I have yet to re-simulate this and see where the fifth is
> > comming from.
> 
> From summation of waves, (additive).
> 
> > 
> > One possible enhancement I'm thinking of is to make
> > a sawtooth wave at 1/2F, F, and 2F simultaneously...then
> > shift each one separatly. If you use a single wave, the
> > phase shift should approximate FM (vibrato).  With multiple
> > waves... I have no idea.  I sure hope its more 'whelming'.
> 
> Its not, i have tried it (if we are talking about
>  the same thing offcourse) FM, yes, you will get
> that but as F increases the FM effect weeakens out,
> at high F you dont hear it at all!
> 
> I had also hoped that it would have achived
> FM without changing F, but this is small
> phase change in a time domain! Ah, well, 
> lets say that you could rotate phase futher,
> several tousand degrees as you increase F,..... then!

Doing post-core wide phase-modulations (and frequency modulations, with an
integrator included) is expensive. It can be done, but I really wonder if it is
worth it. You can play neat tricks by waveshapings, but in order to maintain
record of the offset you better leave that to a digital counter and only handle
the local delta phase shift to pure analogue or you will end up having problems
with dynamics. The trouble now becomes neat handling of the edge cases, but
doing double sawtooth processing a step away and automatically pan between them
would make the jumping sawtooth being totally dampend while the other is at
full force. It is possible but comes at a price of complexity.

Cheap tricks like my sawtooth phase-modulator shows the general idea, but for
deep effects and good quality onces I think my el-cheapo design might need
quite abit of modifications. Some dreamers have drooled puddles by the dream
of having multiple of those in series for deeper effect, which could also be a
possibility but comes at a price of added jitter, which might be acceptable
depending on what you want, but it is not very "purist"-like.

I still consider that in-core exercises have not yeat reached its full prime,
at least not for musical purposes.

Cheers,
Magnus