MOTM

So while waiting for my MOTM kits to arrive, I've been kicking about some
ideas.  And now that I'm on the list, I can finally punt and test the
"feasability" factor.  (Keep in mind that I've never built a modular before,
am not terribly hot with basic electronics, etc., etc.  In other words, have
mercy.  ;)

It seems to me that just about every analog synth out there is equiped with
oscillators that have sine, triangle, pulse, and sawtooth shapes.  Then
there are a few virtual-analogs which have wavetables.  Why not make a 100%
analog wavetable?  I can see two possible ways for this to happen:

1) The extremely expensive route.  Have between 2 and n (16?) oscillators
and cycle through them rapidly.  Essentially n oscilators + a sequencer
(arpegiator?) to cycle.  To trim down the price, you could fix the
individual oscillators to a single traditional shape (e.g., sawtooth).  To
be really obnoxious, each indiviual oscilator could have the features of the
MOTM-300 (ouch!!).  I believe this idea is total overkill.  Let's try option
#...

2) The somewhat-less expensive (more traditional wavetable?) route.  Have
between 2 and n (16?) voltages and cycle through them rapidly.  For each
"step", you could select the voltage level.  Globally you might select the
ramp between steps (linear, hard-step, exponential), and the time between
steps.  Perhaps a setting to select how long a voltage is held for a step--
between immediately go to next step and hold for some short number of
microseconds.  To get really fancy, the hold times and ramp selection might
be adjustable per descrete step.  Perhaps the number of steps could be
variable?

I don't know exactly how pitch is / would be controlled...  perhaps the CV
pitch in sets a level for the voltage steps to modulate around?

Does this sound remotely appealing to anyone?  I'm excited by the idea of
the various shapes that could be employed, and the ability to mutate the
shapes in real-time.

--PBr

Hello and welcome!

I agree that some sort of wavetable oscillator might be very interesting, I 
imagine that Paul has given this some though already. One thing I've found 
with MOTM is the development time can take a while, due to the fact that Paul 
will propose a perfectly good module, then three guys (let's call them Moe, 
Larry and Curly -- and occasionally Shemp) will start adding all sorts of 
frivolous crap: "Add some blinking LEDs!" or "Add a reversing attenuator!" or 
"How many ft/lbs of torque will those new switches take?" After which, Paul 
goes away in disgust for a while and redesigns the module. Then, after a few 
intelligent suggestions by TH, JS, or one of the hundred Daves on the list, 
the module is ready for production.

The point being, get comfortable. But in the meantime...

In a message dated 12/10/99 4:20:35 PM, PaulBr@... writes:

>2) The somewhat-less expensive (more traditional wavetable?) route.  Have
>between 2 and n (16?) voltages and cycle through them rapidly.  For each
>"step", you could select the voltage level.  Globally you might select
>the
>ramp between steps (linear, hard-step, exponential), and the time between
>steps.  Perhaps a setting to select how long a voltage is held for a step--
>between immediately go to next step and hold for some short number of
>microseconds.  To get really fancy, the hold times and ramp selection might
>be adjustable per descrete step.  Perhaps the number of steps could be
>variable?
>

If I'm not completely missing the point, I think you can accomplish much 
(though not all) of what you suggest by merely using a sequencer CV out as an 
audio output, and driving it by a 1V/oct VCO (like the MOTM 300). If you add 
the soon to be released VC Lag Processor (did I mention a reversing 
attenuator yet?) to the output you could get even closer.

John (very occasionally known as Curly either due to the 2' of curly hair I 
had removed from my head last summer, leaving much short curly hair, or due 
to my short curly temper) Barlow

PaulBr@... writes:

>>It seems to me that just about every analog synth out there is equiped with
 oscillators that have sine, triangle, pulse, and sawtooth shapes.  Then
 there are a few virtual-analogs which have wavetables.  Why not make a 100%
 analog wavetable?<<

Perhaps you just need to suppliment your MOTM system with either Doepfer, 
Modcan, or if I remember correctly, Wiard Wavetable Oscillators.  The 
suggestion of doing it totally in the analog domain doesn't sound practical.

-Elhardt

>
> Perhaps you just need to suppliment your MOTM system with either Doepfer,
> Modcan, or if I remember correctly, Wiard Wavetable Oscillators.  The
> suggestion of doing it totally in the analog domain doesn't sound
practical.

Well, let's not get *too* carried away here, shall we? :(

The truth is, that most other waveshapes offer little sonically when used in
a
traditional subtractive synth (ie VCO followed by VCF). The reason is that
the "boring" standard wavefors offer 95% of the combinations of harmonic
content. Without re-inventing a sampler/ROMpler, you can get some
interesting
timbres by modulating ("morphing") between waveforms.

But mainly, stuff like the PPG and such with expensive hardware never really
add all that much.
For one thing, having like 64 points per waveform, then being able to set
the voltage for each
point and saving that will cause you to try 958,485 different settings with
the result being 99%
of a "mushy buzzy" sound.

I have seen the schematic of the W***d design, and without being
...errrr.....{nasty} my overall
feeling was not positive.

Sure, I can design one. Maybe in about 2 years there will be one. But there
are 13 designs
in R&D right now that I think are more immediately useful.

Paul S.

synth1@... writes:

>>The truth is, that most other waveshapes offer little sonically when used 
in a traditional subtractive synth (ie VCO followed by VCF).<<

That's true, which is why I think most people don't realize just how boring 
and basic sounding any static waveform sounds, no matter what it looks like.  
The Access Virus has a bunch of other waveforms besides the standard ones, 
which really don't give you any tones that you can't synthesize using 
standard oscillators and filters.  Morphing between waves gives you something 
more, but you can morph/crossfade between two different synth sounds on any 
modular anyway.

-Elhardt

> From: "Paul Schreiber" <synth1@...>
> The truth is, that most other waveshapes offer
> little sonically when used in a traditional subtractive
> synth (ie VCO followed by VCF). The reason is that
> the "boring" standard wavefors offer 95% of the
> combinations of harmonic content. Without
> re-inventing a sampler/ROMpler, you can get some
> interesting timbres by modulating ("morphing")
> between waveforms.

Well, I agree in part with this concept that the "waveform" is not near as
important as what follows it.  Once you cover the "big boys" (sine, tri,
saw, pulse) you have a large percentage as Paul pointed out.  However, I do
not think we can dismiss the idea that complex waveforms through great
filters and other processing already offered by MOTM will offer a different
sonic possibility.  Certainly, I think we saw how ordinary poor piano
programs jumped to life in the ESQ-1.  The partial "sample" data on the
front end of the oscillator output really did make a difference.

> But mainly, stuff like the PPG and such with expensive
> hardware never really add all that much.
> For one thing, having like 64 points per waveform, then
> being able to set the voltage for each point and saving 
> that will cause you to try 958,485 different settings with
> the result being 99% of a "mushy buzzy" sound.

I agree, I would not want ANY part of that.  However, there is a much more
user friendly platform that is often overlooked when discussing the subject
of wavetable (because it is really a sampler, or was sold as one).

Korg made a "sampling synthesizer" just prior to coming out with the M1
called a DSS-1.  This keyboard had two unique features that allowed users
to create their own waves to experiment with the sonic experience.  The
first feature was one that allowed to create a waveform, using the
fundamental, and all the harmonics (thru some very high number you would
never use).  This particular features was quite useful for those that
wanted to do additive waves to emulate a Hammond sound where the harmonics
were individually added.  Another option actually allowed you to "draw" a
waveform.  Basically, any digital wave could be used as the VCO output to
the remainder of this analog synth.  However, these tools that allowed
either adding harmonics by # and level or doing a "drawn" waveform made
easier work that trying to edit 958,485 different settings.

I will definitely make the argument that at least some of these complex
waves allowed more sonic variety with one oscillator that what basic
waveform oscillators did.  However, one has to accept the fact that a
static relationship will always exist between the varied harmonic contents
in this wave.  Whereas, that would vary over time with multiple oscillator
waveform creation.  So, if you ever get a change to play with one of these
old beasts, I suggest you give them a try.  I have two of them and will
NEVER part with them.  Yep, they are a hybrid machine and will sound
totally digital if you use them that way.  However, if you just load in
basic and semi-complex waves into the memory and start tweaking all those
great analog filters, you will come to the conclusion that complex waveform
does offer some exciting possibilities given a decent way to create / edit
them.

However, they will always be just a waveform.  They will always start form
the same place and always interact with each other exactly the same way. 
However, MIXED with regular VCOs I think there is a LOT of value.  I
wouldn't be interested in MOTM wavetable because there is SO much hardware
out there already that one could use to emulate the same.  As much as I
hate samplers.  They do have a place in life, and complex waveforms is one
of them.  Their failure is that musicians become complacent with the
realism of the sample data and forget to do the remainder of the
programming that takes a dead, flat, dull, stale sample and transforms it
into an interesting sound.

Stooge Larry

Great thoughts on this one -- I agree with what's been said about static 
waves and a lot of what others have said in general.

In a message dated 12/10/99 9:23:04 PM, jlarryh@... writes:

>Korg made a "sampling synthesizer" just prior to coming out with the M1
>called a DSS-1.  This keyboard had two unique features that allowed users
>to create their own waves to experiment with the sonic experience.  The
>first feature was one that allowed to create a waveform, using the
>fundamental, and all the harmonics (thru some very high number you would
>never use).  This particular features was quite useful for those that
>wanted to do additive waves to emulate a Hammond sound where the harmonics
>were individually added. Another option actually allowed you to "draw"
>a
>waveform.  Basically, any digital wave could be used as the VCO output
>to
>the remainder of this analog synth.  However, these tools that allowed
>either adding harmonics by # and level or doing a "drawn" waveform made
>easier work that trying to edit 958,485 different settings.


OK, I have an old AlphaSyntauri (Mountain Music card and a keyboard that 
hooks up to an old Apple II -- don't laugh, it was free) which allows one to 
draw a waveform with a light pen (close to useless for reasons Paul and 
others have mentioned), and a way to build waves by adding harmonics. I used 
to do this on the Synclavier and remember finding it a very rewarding if 
somewhat time consuming way of working -- of course, I've hardly used this 
machine (Syntauri) so there could be a learning curve which I still may be on 
the wrong side of.

I can imagine a VCO which would allow one to select a waveform for each of 
the first 16 harmonics and its amplitude -- I think this could be an 
interesting module. But possibly Crow or others have tried something like 
this already(???).

Some sort of VC timbre modulation would of course be very interesting too!
JB

Specifing harmonics & their relative amplitudes is just a Hammond organ.

Now, what *is* interesting is having 16 VCAs and 16 EGs, 1 per harmonic!
Then,
you synthesize a ton of stuff....a la Synclavier!!

With a moderate $15 DSP this is pretty trivial.

Paul S.

On Mon, 13 Dec 1999 JWBarlow@... wrote:

> I can imagine a VCO which would allow one to select a waveform for each of 
> the first 16 harmonics and its amplitude -- I think this could be an 
> interesting module. But possibly Crow or others have tried something like 
> this already(???).

  Actually, I reprogrammed a DX7's OS ROM to allow me to operate the
phase-accumulator sine oscillators in this fashion.  This was a looooong
time ago, however, and there were a few bugs left in the code when I moved
on to other things.  Besides, I was just re-inventing the NED Synclavier
at that point.  (Or a cheap version of it).  Could probably do this with
an ADSP-2105 these days.  (40MIPS DSP..about $8 these days IIRC)

  --Crow

/**/

JWBarlow@... writes:

>>OK, I have an old AlphaSyntauri (Mountain Music card and a keyboard that 
 hooks up to an old Apple II -- don't laugh, it was free) which allows one to 
 draw a waveform with a light pen (close to useless for reasons Paul and 
 others have mentioned), and a way to build waves by adding harmonics.<<

I was highly influenced by the Fairlight's ability to draw separate free from 
envelopes for each harmonic.  So about 15 years I wrote a program in 
Applesoft Basic, that did the same thing (used trackball instead of 
lightpen).  It produced some really weird evolving sounds, including unwanted 
high harmonic aliasing that worked it's way down into the audio realm to mess 
things up.  Couldn't get more than about 2 seconds of sound with that 
whopping 48K of RAM.  But now the Kawai K5000 can do the same thing if you 
want to hear the effect.

-Elhardt

So when do we get one?

>From: "Paul Schreiber" <synth1@...>
>Reply-To: motm@onelist.com
>To: <motm@onelist.com>
>Subject: Re: [motm] Wavetable-esque oscillator
>Date: Mon, 13 Dec 1999 20:47:06 -0600
>
>Specifing harmonics & their relative amplitudes is just a Hammond organ.
>
>Now, what *is* interesting is having 16 VCAs and 16 EGs, 1 per harmonic!
>Then,
>you synthesize a ton of stuff....a la Synclavier!!
>
>With a moderate $15 DSP this is pretty trivial.
>
>Paul S.
>
>
><< text3.html >>

>Now, what *is* interesting is having 16 VCAs and 16 EGs, 1 per harmonic!
>Then,
>you synthesize a ton of stuff....a la Synclavier!!
>


----Now this kind of stuff is very interesting to me, but it seems like
you'd need a minimum of 50-some knobs (if you use AR envelopes).  Then your
$15 DSP gets pretty damn pricey (not to mention huge).  I think there's a
lot more to this than I'm capable of figuring out--e.g. what do you do about
filters?  Would vcf's not even enter into the picture?

Just wondering
tomr

There are no filters: the decaying harmonics completely characterize the
sound.

I would use an LCD screen to draw each envelope.

Paul S.

----- Original Message -----
From: <hodad1@...>
To: <motm@onelist.com>
Sent: Tuesday, December 14, 1999 6:12 PM
Subject: Re: [motm] Wavetable-esque oscillator


> From: hodad1@...
>
>
> >Now, what *is* interesting is having 16 VCAs and 16 EGs, 1 per harmonic!
> >Then,
> >you synthesize a ton of stuff....a la Synclavier!!
> >
>
>
> ----Now this kind of stuff is very interesting to me, but it seems like
> you'd need a minimum of 50-some knobs (if you use AR envelopes).  Then
your
> $15 DSP gets pretty damn pricey (not to mention huge).  I think there's a
> lot more to this than I'm capable of figuring out--e.g. what do you do
about

> filters?  Would vcf's not even enter into the picture?
>
> Just wondering
> tomr
>
> >

In a message dated 12/14/99 5:51:34 PM, synth1@... writes:

>There are no filters: the decaying harmonics completely characterize the
>sound.

>I would use an LCD screen to draw each envelope.

Put a few undedicated VC inputs in it (for us analog people -- to be 
determined with software upgrades) and I think this might be an interesting 
MOTM compatible module.

JB

In a message dated 99-12-13 21:41:09 EST, you write:

<< OK, I have an old AlphaSyntauri (Mountain Music card and a keyboard that 
 hooks up to an old Apple II -- don't laugh, it was free) >>



jb,
far be it from me to laugh......... i still have a c-64 hooked up thru a midi 
interface to a casio cz-101!  ( even though my main midi computers are a 
family of 3 amigas - 2 for midi control and one for sampling ) :^)   i`ve 
always found  "free" to be a very compelling argument. plus, my equipment 
"arsenal" ( now there`s a "don`t laugh" ) is modest at best. i sometimes feel 
a twinge of envy for the folks who have a ton of stuff. still, i have fun 
with what i have and i guess thats the main thing..........
best,
dave

In a message dated 99-12-13 21:45:40 EST, you write:

<< With a moderate $15 DSP this is pretty trivial. >>



paul,
great! when can we expect to see one added to the lineup?
best,
dave

>
>I would use an LCD screen to draw each envelope.
>
So would you have a light pen or a touch screen or an x knob & a Y knob ala
Etch-a-Sketch?
Another thought that popped into my head woulbe to have a plug for a
graphics tablet (isn't that what they call those things?) for remote
waveshaping.

tomr