MOTM

In a message dated 00-03-16 12:06:24 EST, you write:

<< Maybe an adjustable quantizer module is a job for a PIC processor...?
  >>


ken,
beyond knowing it is a moderately powerful microcontroller type of cpu chip 
family, i wouldn`t have any idea about hacking the microcode to do the job. 
in theory, its just simple math, programmable scale for whatever note per 
octave  or just scale you wanted and feed a linear c.v. in one end and out 
comes an appropriately scaled, stepped c.v..  then, depending on cost / 
complexity, can it be an affordable module ? can it be done in a timely 
manner ?  aren`t a couple of folks on the list experienced pic programmers ?
best,
dave v.

On Thu, 16 Mar 2000 davevosh@... wrote:

> << Maybe an adjustable quantizer module is a job for a PIC processor...?
>   >>

> beyond knowing it is a moderately powerful microcontroller type of cpu
> chip family, i wouldn`t have any idea about hacking the microcode to
> do the job.  in theory, its just simple math, programmable scale for
> whatever note per octave or just scale you wanted and feed a linear
> c.v. in one end and out comes an appropriately scaled, stepped c.v..  
> then, depending on cost / complexity, can it be an affordable module ?
> can it be done in a timely manner ?  aren`t a couple of folks on the
> list experienced pic programmers ? best, dave v.

  PICs aren't particularly powerful--they have very limited memory
resources--but, they are *fast*.  For dedicated apps, this speed is
everything.  For something like a quantizer, a PIC should have no problem.
In fact, the tricky part of quantizing is in choosing really good A/D and
D/A subsytem parts so as to keep the error values as small as possible.  

Crow

/**/

In a message dated 3/16/2000 1:33:18 PM, oldcrow@... writes:

>> beyond knowing it is a moderately powerful microcontroller type of cpu
>> chip family, i wouldn`t have any idea about hacking the microcode to
>> do the job.  in theory, its just simple math, programmable scale for
>> whatever note per octave or just scale you wanted and feed a linear
>> c.v. in one end and out comes an appropriately scaled, stepped c.v..
> 
>> then, depending on cost / complexity, can it be an affordable module
>?
>> can it be done in a timely manner ?  aren`t a couple of folks on the
>> list experienced pic programmers ? best, dave v.
>
>  PICs aren't particularly powerful--they have very limited memory
>resources--but, they are *fast*.  For dedicated apps, this speed is
>everything.  For something like a quantizer, a PIC should have no problem.
>In fact, the tricky part of quantizing is in choosing really good A/D and
>D/A subsytem parts so as to keep the error values as small as possible.

Quantizer: another idea that we've beat the life out of here! When this first 
came up, maybe a year ago, there was a serendipity around Dave V.s suggestion 
of programmable modes and tonalities. The most interesting, for me, was 
Thomas H.s concept of "personalities" which could be programmed and later 
indexed through. This would probably be more than a PIC could do, and I'm not 
entirely sure what it would mean, but I'd like a quantizer like this as well!

JB

In a message dated 00-03-16 15:33:16 EST, you write:

<< n fact, the tricky part of quantizing is in choosing really good A/D and
 D/A subsytem parts so as to keep the error values as small as possible.   >>


crow,
unless d/a - a/d systems have fallen in price radically, that could be a 
limiting factor. it used to be ( is it still ? ) that increasing accuracy 
cost quite a bit more. plus, what is enough ? 10 bit ? 12 bit ? 16 bit ?  i`m 
not sure if those newer cheap d/a`s that are used in cd players, etc. are 
compatible with this sort of application. may need to fall back to the old 
brute force approach.
best,
dave v.

On Fri, 17 Mar 2000 davevosh@... wrote:

> crow,
> unless d/a - a/d systems have fallen in price radically, that could be a 
> limiting factor. it used to be ( is it still ? ) that increasing accuracy 
> cost quite a bit more. plus, what is enough ? 10 bit ? 12 bit ? 16 bit ?  i`m 
> not sure if those newer cheap d/a`s that are used in cd players, etc. are 
> compatible with this sort of application. may need to fall back to the old 
> brute force approach.

  Well, I wouldn't use linear PCM DACs as my first choice. :)  I will have
to look and see what the latest parts are that would 1) serve the purpose
and 2) not cost too much.  Maxim makes a bazillion A/D and D/As; I'm sure
something in there would work.  Isn't there some dude on the list that
works at Maxim? ;)

  SCI used a 16-bit DAC to do all their pitch (and parameter) setting;
I'll have to dig out the technical manual and review how they arrived at
pitches and controlled errors.  I think they also used the DAC as part of
a SAR circuit to digitize the knobs to something like 7 bits.  It is a
starting place, at least.

Crow

/**/

In a message dated 00-03-18 02:02:41 EST, you write:

<< SCI used a 16-bit DAC to do all their pitch (and parameter) setting; >>



crow,
maybe i`m too old and out of date.........16 bits seems like too much to 
"waste" as part of a control system. 12 bits would be adequate, i would have 
thought. then again, if the price difference is minimal, maybe it would be 
the better choice.
keep us posted !
best,
dave v.