asleep at the wheel..

[Batz Goodfortune]

Y-ellow SYSph:

At 12:36 PM 10/27/99 -0500, SYSph at ntOPus wrote:
>Could someone define a free running LFO for me just to make sure I'm on
>the same page as everyone else? Sorry about the lack of attention...

The term "Free running LFO" I guess is rather like "Sysex on the fly". Both
terms would never have come about other than for modern technology which
seeks to cut corners.

A free running LFO is like that on a modular. You simply select a speed and
send it to something. It is not restarted or effected by a key-on event or
any other event for that matter. It just keeps on pumping out very low
frequency wave forms.

It's one way of doing filter sweeps for example. A great many synths (way
too many) because  of their corner cutting nature -as mentioned in a
previous post about LFOs in poly synths- will restart the LFO on every
key-on event. Subsequently even if your LFO is running @ 0.01Hz, it will
behave more like an envelope than an LFO unless the key is held constantly.
Which means you can't do filter sweeps or even rippling timbres across a
sequence of events. Only during the transient of a single note. Does that
make sense to  you?

Some older mono synths had the curious feature of allowing an LFO delay.
The effect of the LFO would not be active upon the voice for a time period
set by the delay. After which the effects would either A) turn on suddenly
in the  case of a cheap-assed synth or B) slowly increase in magnitude in
the case of a properly designed synth. It is reasonable  to assume that you
could also have a setting which would allow you to delay the LFO then set
the time interval over which it would increase it's effect. I've never seen
such an LFO but it's possible. 

In a cheap synth  the delay setting and the LFO are intrinsically linked
such as the LFO will always Come on or re-trigger at the key-on instant and
increase in effectiveness over time. A better designed synth will allow the
 LFO to free-run when the delay  is set to zero. Such that the LFO is not
effected by an external event unless otherwise programed to do so. If you
wanted the LFO to be triggered by the Keyboard or what ever, but you wanted
it to be reasonably instantaneous,  you'd simply raise the LFO delay ever
so slightly. Otherwise it would free run.

Alas cheap-assed modern digital synths based on single chip sample playback
engines, such as those found on $10 sound cards and 9000 dollar samplers
(they're all pretty much the same) generally don't offer free running
synths for technical reasons. Even some well known CPU controlled analogue
synths have this deficiency and it's all to do with dynamic assignment. And
you can Thank Roland for that...

As previously mentioned, Dynamic assignment has the advantage of giving you
more apparent voices than you actually have. You might have a 16 voice
polytimbral synth. Pumping each channel theoretically allows you to play 16
notes at once on all 16 channels giving a theoretical number of voices of
256 voices. Just as long as you don't play more than 16 of them at any 1 time.

The way these things work is that they have 16 generic voice channels. The
actual voice parameters are assigned to which ever voice is free at any
given time. So while a given voice channel might be a poly pad one second,
it could be a kick drum the next. This means that every time a new note is
assigned to a voice channel, a new LFO is defined for that voice channel as
well. It cannot, by definition, be "Free-running".

As mentioned before, as a designer you can do one of 2 things. You can
capture the phase and rate of the LFO and re-instate it over time on the
next iteration of that voice. Or you can simply have 1 or more external
free running LFOs. The former is better because it allows you to have a
free running LFO on every voice, but as you can imagine, it adds a
magnitude of complexity to the system. Alright for your DSP heads with a
little CPU bandwidth to spare but technically difficult in hardware. Unless
you're designing a custom chip or 3.

Yamaha never had this problem because for years they refused to follow
Roland's lead and kept with their policy of pre-defined polytimbrality.
Since each voice and it's associated level of Polyphony had to be
pre-defined in advance, all 8 or 16 LFOs per voice could be allocated in
advance and therefore you could elect to have a free running LFO on every
voice if you wished. Later, when Yamaha succumbed to the dark side, they
instantly found that continuing this tradition was not easy. So naturally
they dropped it.

Which is a damn shame. Synths like the little TG55 would be ace wicked if
it weren't for that. It has kick-ass filters but you can't do a filter
sweep because it's LFO won't free run. And due to a slight misjudgment in
it's software, you can't actually sweep the filter from MIDI once the key
is pressed. It simply locks out the controller until another key is
pressed. This can actually be quite useful some times but if you're trying
to "WOW" a voice by hand to give it some formant (read vocal) qualities,
forget it. It can't be done.

And just to sum up. As I said in the beginning, This terminology would
never have come into usage if we had all stuck with modular synths. Since
you define the LFO's parameters and it free runs of it's own accord. It was
just a given that it worked like that. I don't recall any modular synth
I've ever used even having the ability to re-trigger the wave form. Why
would you want to? The only reason they do it now is due to the technology
of dynamic assignment they use. And, in the words of Monty Python. "It
get's me very irritated!"

be absolutely Icebox.

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