[sdiy] Bus for digital patching of synths

[cheater00 .]

On Fri, Nov 8, 2013 at 2:05 PM, Neil Johnson <neil.johnson71 at gmail.com> wrote:
> Hi,
>
> 1/ Converters need to be DC to, say, 40kHz.  DC specs need to exceed
> that needed for handling pitch CVs. Audio specs don't need to be that
> good as synth circuits generally not designed to the same quality as
> in high-end recording gear (how many bits do you need to capture the
> output of a TL072?).

True enough. I wonder if we'd need to specialize, i.e. audio vs CV, or
if such a project is better off finding a single good chip and
ordering in bulk.

However, both the bit depth and sampling frequency can likely be
optimized for in the specific circuit being built, if one so desired.

> 2/ Have you considered CopperLAN?

I've had to exclude Ethernet due to the above considerations. The most
important thing is that it wasn't TDM, and that it would have
introduced delays. This is not a problem for DAW users who are behind
a huge buffer anyways.

> 3/ Latency between modules, especially handling gate and trigger signals:
>
> http://www.innerclocksystems.com/New%20ICS%20Litmus.html

Yep, exactly why CopperLAN isn't good. With a better system you could
even have 1 sample delay.

> 4/ Cost - where is the breakeven point?


Good question.

Do you want a mono modular with patch storage? If yes, then likely
you're approaching the cost already.
Do you want a polyphonic modular? If yes, then you likely need digital
control of some parameters (because the knob needs to be replicated,
and analog distribution of that could be pretty troublesome - consider
the amount of wiring). Then there's the amount of not-so-inexpensive
cables you'd need to buy.

It also saves you on mults, CV mixers, LFOs, slow envelopes, CV logic,
and all that other stuff that digital can do.

> http://www.gearslutz.com/board/so-much-gear-so-little-time/867829-chat-about-future-digitally-controlled-analog-console.html
> http://www.muffwiggler.com/forum/viewtopic.php?t=68864
>
> Granted, you're only suggesting replacing the patch cables with
> ADC->cable->DAC so no need to worry about converting pots and switches
> into something digitally controllable.  But then you've split the UI
> into a panel of knobs, and an application running on a computer that
> manages the interconnections.  Or were you thinking of adding some
> sort of interactive button pressing/touchpads/etc scheme to set up
> virtual patch cords?

If the system is indeed a network then either version can work. Most
likely a tangible button for each patch source and destination is a
good idea. However, to manage those patch cords you're likely looking
at a screen of some sort. Whether it's a DAW, iPad, or a dedicated
touch screen with an embedded computer doesn't necessarily matter,
since the thing's a network, so you just need to be able to hook into
it with an interface.

> In the meantime write up the specs so that manufacturers can implement
> the functionality in their modules, arrange plugfests so that
> manufacturers can get together and debug the interactions between
> their modules, and so on.  Read up about CopperLAN and the approach
> they've taken (technically and businesswise).

Will do. It's likely a good idea to show prototypes to as many faces
as possible.

> And convince the (small) market that your solution, while not pure
> analogue end-to-end modular synth, is just as good if not better,
> definitely better than a fully software environment (e.g. Reaktor).

I see what you mean, however it's not really meant to convince the
users - if anything the developers, who are less likely to be... let's
say inexperienced. :-)

> Sounds like fun!
>
> Neil
> (PS: ADC on INPUT and DAC on OUTPUT)

Ah, what I meant was ADC sampling the output for use in a digital
patch, and DAC feeding into the input from that digital patch.