Characteristics of a ESP type of device...

[Harry Bissell]

OK...

Magnus Danielson wrote:

> Hi!
>
> I came to wounder what characteristics one should be able to measure on a
> Hz to CV converter like the MS-20 ESP etc. Thus, what diffrent figure of merit
> could we note that diffrent solutions have. It was just a little interesting
> exercise to figure on, and I would like to hear some opinions on this...
>
> Here are a few ideas that poped out of my head...
>
> Input Sensitivity
>         Input Dynamic range (noise floor and clip level)
>         Input scaling range
>         Input scaling method: Static, Manual, Auto

Good points... esp if the circuit is for static amplitude inputs (like a VCO or
theremin oscillator...) or dynamic like voice or guitar

>
>
> Selectivity
>         Method: BP, LP + HP..
>         Range: Fixes, Manual or Automatic
>         Tracking: Yes/No
>         Range & BW
>         Stop band rejection
>         Slope steepness

I think the test here would be Harmonic Rejection (or Fundamental Recovery). It
should be able to reject 2nd (and other) harmonics. Be assured that most guitars
will screw this one up !!! Maybe a certain IM mix of Fundamental to some Harmonic
would work. I use two sine waves (non harmonically related) to test these circuits
in SPICE.

>
>
> Envelope follower
>         Dynamic range with (+/- 3dB accuracy)
>         Rise time (static value or range)
>         Release time (static value or range)
>         Offset error
>         Signal leak through

Guess it depends on the expected nature of the input signal. For guitar I use a
modifed peak detector with a non-linear filter scheme to get the range of release
times required. But of course this wouldn't work for (lets say) a flute...

>
>
> Trigger/Gate
>         Trigger level (static or range)
>         Trigger type (fixed or peak relative)
>         Peak release time (static value or range if available)
>         Gate hold time (if available)
>
> Hz to CV converter
>         Scale type: linear, exponential
>         Tracking linearity (+/- cent for range)
>         Offset error
>         Rise time (for say +/- 25 cent estimate)
>         Dead input residue state: 0V, fixed value, hold
>         Max tracking speed (for sine sweep) in Hz/s and oct/s

There are TWO classes here (maybe more..) The Tachometer frequency converter that
takes many cycles to acquire a final value... and the slope method which usually
requires two cycles (could be one with more circuitry...).

The Tach circuit has very slow response and trades acquisition time for ripple in
the CV out. The Slope method has (comparatively) fast response but will step from
one value to another abruptly. Both have merit.  The Tach circuit is too slow for
tracking a musical instrument accuratly (unless a lot of ripple is a "feature"...)

Nice Idea, Magnus...

:^) Harry

>
>
> Should these be feasable to measure, use for judgement of relative
> applicability in real life, missing some important fault or having totally
> irrelevant information. Note that I had various forms of curcuits in mind and
> naturally diffrent limits applies.
>
> Cheers,
> Magnus