Free running LFO's

[Batz Goodfortune]

Y-ellow Antti 'n' Paul 'n' Y'all.

At 01:18 PM 10/27/99 +0100, Paul Maddox wrote:
>Antti,
>
>  Good question... I reckon most have a common LFO, the
>newer ones seem to have an lfo for EACH voice (that is if
>you had aan 8 note poly synth, it would have 8).. One of the
>things I hope to impliment on my poly is 9 LFO's, one global
>and 1 for each voice... but time will tell.

In the Yamaha synths which are polytimbral as well as polyphonic, there is
an LFO on every voice. except where a polyphonic voice is selected. So if
you have 8 different patches you get 8 different LFOs. Or if you have a
single patch but 8 voices you get a single LFO since it is all the same
patch. One LFO per patch in what ever combination. Up until the late 80s,
Yamaha synths required the level of polyphony to be pre defined. IE:  you
had to specify how many voices of which patch you needed to use. If you
needed to play a triad then you would specify 3 voices on that patch.
Meaning that you were then left with a total of 5 voices out of 8 to do
anything else with.

Roland, on the other hand, always used dynamic assignment and this has
become the norm today. What this means is that if your synth is 8 voice
Polly then you can specify 8 voices across 8 MIDI channels. The CPU
automatically assigns priority depending on what's coming in. So if there
were 8 single notes required of 8 separate patches, you'd hear one of each
patch. But if a chord was required then it would gobble up enough notes out
of the pool of 8 to play that chord. If more notes were required than were
available in the pool then a priority system comes into effect.

If the Envelope is decaying and the note released then the voice channel
can be reused. But only if there are no other voice channels which have
been released for longer available. It's a voice of last resort system and
means  that in lower polyphony synths, some polyphonic voices tend to drop
off the end. A major reason why pianos sound crap. With larger orders of
polyphony however this system works wonderfully because it can appear to
have a larger number  of voices available than it can actually play at
once. Which is why Yamaha eventually adopted their own version of this
scheme. With the twist that they use cyclic redundancy rather than
selective redundancy.

With LFOs it is the same. An LFO gets assigned to each voice on an
as-needed basis but this has the disadvantage  that the LFOs cannot free
run. Because there is no tracking of their cycling time.

What is really required in a polytimbral polyphonic synth is that it have
one assignable LFO per voice and at least one out side for free running. Or
an elaborate means of tracking the cycle time on an LFO per voice. So that
the LFO on a newly assigned iteration of that voice picks up where it
should be given the time interval rather than starting again afresh.

If it were a digitally implemented synth then you'd probably tend toward
the latter. If it w ere an analogue implementation then you'd probably want
a separate free running LFO or 2. Something which no dynamically assigned
polyphonic analogue-esk synth implements to my knowledge. Which is the
advantage of Yamaha's pre-assigned polyphony. If you assign a number of
voices then they will share a free running LFO.

Hope this helps.

be absolutely Icebox.

 _ __        _                              
| "_ \      | |         batzman at all-electric.com
| |_)/  __ _| |_ ____       ALL ELECTRIC KITCHEN               
|  _ \ / _` | __|___ |  Geek music by geeks for geeks
| |_) | (_| | |_  / /   
|_,__/ \__,_|\__|/ /    
                / ,__   http://www.all-electric.com
Goodfortune    |_____|