portamento vs. glide (was Re: [music-dsp] step sequencers)

[Theo]

Not sure what you mean with causal and non-causal poratmento.
But in analogue both linear and expo portamento are possible.
In case of the 303 the "slide" is exponential (cap to GND) while the
"slide-rate" depends on the "new" pitch.

In the schematic of the main board in the service manual,
the pitch CV DAC and the slide circuit are on the top left hand side.
The 6 bit DAC is a hex D-flipflop with a R2R net work and a buffer amp
(IC11b).
A bit unusual is the value of the R2R resistors, a whopping 200k, but not
without reason...

The clock line that latches the data in the FF double functions as slide
on/off signal.
A short pulse on the line just loads the FF.
Keeping the line high, loads the FF and switches the "slide" function on.

To switch the slide function the line goes to a "alternating switch" build
around two 4066 cmos switches.
This "alternating switch" connects one side of a 220n capacitor (C35) to
either; the output of IC11b or to the non-inverting input of the same.
The other side of C35 goes to GND.

When connected to the output of IC11b, C35 does noting (except being kept
charged at the CV voltage).

However when C35 is connected to the + input of IC11b,
it forms a LP filter together with the resistors of the R2R network.
This is why the R2R resistors have the unusual high value.
Because the R2R network is part of the "slide-filter" the time constant
(rate) for the "slide" depends on the new selected pitch.

Also not a look ahead function,
but more complicated than the usual "RC with a switch to make R=0"
portamento implementation.
There are some more points in the 303 where extra components are spend to
get a more musical synth.
Build cheap as possible? yes sure, but cheap as hell??

HTH
Theo



From: David Lowenfels <dfl at alum.mit.edu>

> JeffyB wrote:
> > Wait... is portamento different from slide? It seems to me that slide
> > is non-causal and that portamento is causal... right?
>
> Jens Blomquist wrote:
> > I think it's faster, simpler and easier to understand if you do all
> > modulation in the note "domain" and then convert it to the freq/pitch
> > domain in the last step. Gliding becomes a simple linear interpolation
> > == One add per sample.
>
> I think it's a myth that the TB303 has "look-ahead" portamento.
> Sure you could easily add this feature in a digital synth though...
> and as synth designers I think we _should_ do more than just simulate
> old gear.
> We might as well take advantage of adding new features that the digital
> domain
> affords us. But as far as I know analog synths cannot perform linear
> interpolation
> for a non-causal portamento. I think it would be a pretty complicated
> circuit
> and the 303 surely was built to be cheap as hell. Can anyone confirm or
> deny this?
>
> -David
>
> dupswapdrop -- the music-dsp mailing list and website:
> subscription info, FAQ, source code archive, list archive, book reviews,
dsp links
> http://shoko.calarts.edu/musicdsp
> http://ceait.calarts.edu/mailman/listinfo/music-dsp