Re: [LAD] CV data protocol in apps.

On Fri, Feb 19, 2010 at 04:05:02PM +0300, alex stone wrote:

> An obvious question i guess, but is 1/16 a fine enough resolution for
> a wide selection of use cases?

I don't know about any 'human interfaces' (faders, knobs,
mouses, touchscreens, acceleration sensors, whatever) that
can generate 1000Hz control signals.

Even a 'Mute' on a mixer would in the DSP code evaluate
to a controlled fade in/out of a few milliseconds.

Control signals may require exact timing (which this
form still provides), but they don't required high
bandwidth.

And as said, it's not meant to be a replacement for
generic control data, even if you could use it e.g.
for things like note on/off with infinte resolution.

The advantage of having 'analog' control signals
(by which I mean data interpreted as a sampled
waveform) is that they are easily patchable - the
receiver does not require knowledge of any specific
data encoding. It is this what makes control voltages
in real analog synths so versatile - there is no
discussion about what they mean - just a function of
time.

It would of course be possible to create an 'encoded'
data format that could represent arbitrary low-frequency
waveforms to be used for control. To be really useful
such a format should also permit to encode how fast a
value is moving, and maybe also the second derivative.

For example you could do this by specifying some form of
interpolation between arbitrarily spaced and timestamped
control points. Now imagine the complexity of just adding
two such streams. Or editing them.

Using just a stream of equally spaced samples is much
easier.

-- 
FA
O tu, che porte, correndo si ?
E guerra e morte !
_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-dev