MOTM

At 3:04 PM -0800 01/29/02, Tony Karavidas wrote:
>
>   No  it's not. Remember the local oscillator needs to span a range of 17
>octaves  ( not far off from Paul's ultra VCO, and that VCO alone costs
>$349  assembled. ) and so to get real calibrated CV, you'd need CV control
>of all  that.

I'm guessing that's .05Hz to 3.2kHz, or maybe my math is wrong??  Yes, that
is very difficult.  The trouble is, motm users expect oscillators to have
1V/Oct control.  It's one reason all of the other quality frequency
shifters out there cost over $1000.

>In reality, you could choose a small range, but since this
>has application  far different than an oscillator, calibrated CV isn't
>very important. What are  you calibrating anyway? You're taking an input
>frequency spectrum and shifting  it by X Hertz. As soon as you start to
>shift your entire harmonic  relationship starts to get mangled, so really
>what's the point to CV control  other than to allow modification of how
>"messed" up things are getting?

That isn't true.  In addition to what JH has already mentioned (getting the
amount of shift to track the keyboard), there are other applications that
depend on "tuning" the amount of shift.

>Once you  shift too far, things all sound basically like chipmunks.

In that chipmunks are very quiet creatures and extremely large shifts would
render the output inaudible, this is true.  If you are referring to Alvin,
Simon, and Theodore, that's pitch shifting using a tape machine -- an
entirely different effect :)

While there are frequency shifters that can shift by as much as 16kHz,
imho, that's far from practical, although I'm sure there are readers of the
Computer Music Journal who think it's a great idea :)

>The  beauty of really slow shifts is apparent
>in spacey music. It has a gorgeous lush  rising (or falling) sound.

Yes, low frequency shifts are required for "phasing effects" and feedback
control.  When the amount of shift is a frequency in the audible range,
carrier rejection becomes very important.

At 2:31 AM +0100 01/30/02, jhaible wrote:
>
>Why ? If you make the oscillators digital (which is a good idea),
>you might as well replace the multipliers with multiplying DACs.

While that certainly is a consideration, I like the idea of having
oscillator outputs.  That way you get both a quadrature oscillator and
frequency shifter for your ~$400.  They could be used as modulation or
audio sources with other modules.  It also makes it easier to tune the
amount of shift when in its in the audio range.  If the multipliers are
analogue, I'm thinking the oscillators could easily be tapped with dual
op-amp and two extra jacks.  If space is an issue, TRS jacks could be used.

Also, wouldn't there be problems with aliasing??  Notice I'm not
anti-digital, I'm merely raising the question.  A digital frequency shifter
might have its advantages.