MOTM

I uploaded the file "motm_strings.mp3" into the files section. They are early 
experiments processing the motm with violin/cello like body resonance, from a 
month or two ago. I didn't post them as I was waiting to do something better 
(less hollow, more real), but since JH seemed to like them, others may find 
them interesting, especially those who want to hear more than one note. The 1st 
example showing the raw MOTM tone and then the processed tone clearly shows the 
electronic equal of what I wanted to try using a transducer attached to a real 
instrument, for those who remember that thread. Other than foot pedal for 
volume, there are no real-time dynamics in these snippets. Read the following 
directions and notes first.

Notes:
This demo needs to be listened to through headphones only. Playing through
speakers can cause any of the following: * Can make some of the highly
resonant strings sound like they're playing through a hollow glass tube or
sound digital, scratchy and artificial in nature * Highly resonant sounds
will most likely cause distortion and clipping through speakers even at
relatively low volume * Can result in lack of sound detail or sound like
crap. I've experienced all of these, so if you ignore the headphone rule
don't come complaining to me about that the demo sounds lousy.

1) First example plays the raw synthesized sound used for input into the 
electronic wood resonator, then shows the dramatic change created by the cello 
wood resonator. The effect is almost too extreme here and sounds like a huge 
round cello on steroids.

2) Next shows the difference between a cello tone and a violin tone. The 
difference in size of the sound of the instruement is mostly do to the 
different placement of main wood and air resonances.

3) Plays a short rising scale emulating a 16 piece string orchestra. Each 
instrument is recorded one at a time and positioned differently in the stereo 
field. With intensionally played inaccuracies and differing bits of portamento 
per instrument, the effect is quite realistic.

4) Shows a piece of music using the above technique.

-Elhardt

:> I uploaded the file "motm_strings.mp3" into the files section.
:> They are early experiments processing the motm with violin/cello like
body
:> resonance, from a  month or two ago.


Do you mind expounding a little more on the technique used to generate the
body resonances?

Thanks,

Les

Les Mizzell writes:
>>Do you mind expounding a little more on the technique used to generate the 
body resonances?<<

I forgot to mention that part. In that MP3 I'm running the motm through a 
something like a 32 or 36 stage phaseshifter in a Roland SE70 to generate a 
complex series of peaks and notches. Then I'm using some parametric EQ to 
tailor the sound, such as boosting the main body and air resonance frequencies. 
Then it goes through a resonant 32 band filter bank patch in a Nord Modular. 
That filter bank was set up to sort of mimic the Moog String filter. However, 
if used by itself, it doesn't really give much of a wood-like tone. The phaser 
and EQ are probably contributing more in the MP3. I was trying to get away from 
using a bunch of delay lines in parallel as I did in my Nord Mod strings from a 
few years ago. But it is becoming clear to me that using devices that create a 
lot of unknown notches or peaks in the audio spectrum doesn't quite give me the 
exact sound I want (although for my acoustic guitar simulations, they work 
great). Some notes sound OK, some sound ugly, I get hollow tones, or metalic 
tones. I'm now trying to get rid of the phaser and EQ altogether and use a 
really flexible parametric filter bank patch and then painstakingly set it up 
to give me the exact harmonic structure I want. But that requires looking at 
violin body plots and the harmonics of different notes and trying to come up 
with filter bank settings that can get me close to what I see.

-Elhardt

> In that MP3 I'm running the motm through
> a something like a 32 or 36 stage phaseshifter
> in a Roland SE70 to  generate a  complex
> series of peaks and notches....and so forth...


Few other ideas and thoughts and ramblings....

*********************

Since Elhardt isn't afraid of digital (...and we both obviously love our
Nords....)...

I've been neglecting my Nord recently as I slowly get comfortable with
complex sound creation with ABSynth. ABSynth has, as one of the filter
choices, a tuneable comb filter that I find truly amazing. You can set up an
ABSynth patch that looks like:

OSC - FILTER  \
OSC - FILTER   --Waveshaper -- Master Filter -- Master Effects
OSC - FILTER  /

The Master Effects section has another multicomb filter with up to six taps,
or a multitap delay, or a "pipe".

Between all these, it's pretty easy to set up all kinds of different
resonant structures. I've been working on a Cello patch that's coming along
very nicely.  Will post some audio files in the next few days, if I can kick
this freaking flu virus I've caught...


*********************

So, what to do with our analogue synths for this type stuff? Buy 15
multimode filters and a few delays and put up with the accumulated noise of
them all? Seems an expensive solution at best...

There are several electric guitar preamps out there (Roland comes to mind)
that uses modelling to give you a decent sounding "Acoustic Guitar" from
your electric. Doesn't sound bad either.  Now, a guitar body is a little
different that a Cello body, and too big for Viola or Violin, but it would
be interesting to see what it would do as far as helping create a more
"acoustic" timbre from a synth.

I've thought it would be interesting to try a Helicon VoicePrism VoiceCraft
on a modular synth as well, since it allows you to define vocal tract
resonance...

What I'd love to see, would be a digital module that would, probably with a
computer interface, allow you to define a resonant structure and then use it
as a filter in your system.  All the individual components to do this
already exist. To varying degrees the Yamaha VL series used software to do
this, and the Korg Z1 did this as well...

Now if somebody was smart enough to reverse engineer the process, and then
get just the resonant filter structures to run on a small processor on a
single module......jezz, would that be cool or what?

Since Paul is now adding MOTM modules using digital components, I don't
really see this as "cheating" in the analogue world, or am I in a minority
here...

I'm not particularly interesting in creating a dead-on violin or trumpet
myself - but what does interest me is the creation of something that would
be 1/2 way BETWEEN a violin and trumpet, and sound acoustic enough to sound
"real" and not electronic - even though it would be an instrument nobody has
ever heard before....

Just my random thoughts anyways...

In the same spirit, I nearly sold my Micro Modular when I heard that 
Clavia was moth-balling it (no more new revs).  But since I'm 
waiting on a new Ztar Z6-S (www.starrlabs.com) I demurred, and am 
glad I did.  Lots of cool sounds and very interesting to experiment 
with. The sound doesn't really come near the MOTM, but it is 
polyphonic and MIDI'd.  And the editor is cool.

ABSynth, I hear, is also cool.  The other incredibly interesting 
software synth is the Malstrom in Reason 2.0.  It's a combo of a 
wavetable synth and a granular synthesis with interesting routing 
and two comb filters.  Really unique timbres!

One more thing: Antares kantos is the total bomb.  Completely 
weird.  It will take a monophonic recording and resynthesize it in 
various strange ways.  What it does to vocals is, as my teenage 
daughter would say, just wrong.

Finally, has everyone seen the Oddity? Software ARP Odessy (sp?) 
with some extra tricks...

Mike 

--- In motm@yahoogroups.com, "Les Mizzell" <lesmizz@b...> wrote:
> 
> 
> > In that MP3 I'm running the motm through
> > a something like a 32 or 36 stage phaseshifter
> > in a Roland SE70 to  generate a  complex
> > series of peaks and notches....and so forth...
> 
> 
> Few other ideas and thoughts and ramblings....
> 
> *********************
> 
> Since Elhardt isn't afraid of digital (...and we both obviously 
love our
> Nords....)...
> 
> I've been neglecting my Nord recently as I slowly get comfortable 
with
> complex sound creation with ABSynth. ABSynth has, as one of the 
filter
> choices, a tuneable comb filter that I find truly amazing. You can 
set up an
> ABSynth patch that looks like:
> 
> OSC - FILTER  \
> OSC - FILTER   --Waveshaper -- Master Filter -- Master Effects
> OSC - FILTER  /
> 
> The Master Effects section has another multicomb filter with up to 
six taps,
> or a multitap delay, or a "pipe".
> 
> Between all these, it's pretty easy to set up all kinds of 
different
> resonant structures. I've been working on a Cello patch that's 
coming along
> very nicely.  Will post some audio files in the next few days, if 
I can kick
> this freaking flu virus I've caught...
> 
> 
> *********************
> 
> So, what to do with our analogue synths for this type stuff? Buy 15
> multimode filters and a few delays and put up with the accumulated 
noise of
> them all? Seems an expensive solution at best...
> 
> There are several electric guitar preamps out there (Roland comes 
to mind)
> that uses modelling to give you a decent sounding "Acoustic 
Guitar" from
> your electric. Doesn't sound bad either.  Now, a guitar body is a 
little
> different that a Cello body, and too big for Viola or Violin, but 
it would
> be interesting to see what it would do as far as helping create a 
more
> "acoustic" timbre from a synth.
> 
> I've thought it would be interesting to try a Helicon VoicePrism 
VoiceCraft
> on a modular synth as well, since it allows you to define vocal 
tract
> resonance...
> 
> What I'd love to see, would be a digital module that would, 
probably with a
> computer interface, allow you to define a resonant structure and 
then use it
> as a filter in your system.  All the individual components to do 
this
> already exist. To varying degrees the Yamaha VL series used 
software to do
> this, and the Korg Z1 did this as well...
> 
> Now if somebody was smart enough to reverse engineer the process, 
and then
> get just the resonant filter structures to run on a small 
processor on a
> single module......jezz, would that be cool or what?
> 
> Since Paul is now adding MOTM modules using digital components, I 
don't
> really see this as "cheating" in the analogue world, or am I in a 
minority
> here...
> 
> I'm not particularly interesting in creating a dead-on violin or 
trumpet
> myself - but what does interest me is the creation of something 
that would
> be 1/2 way BETWEEN a violin and trumpet, and sound acoustic enough 
to sound
> "real" and not electronic - even though it would be an instrument 
nobody has

> ever heard before....
> 
> Just my random thoughts anyways...

Les Mizzell writes:
>>ABSynth has, as one of the filter choices, a tuneable comb filter that I find 
truly amazing....The Master Effects section has another multicomb filter with 
up to six taps, or a multitap delay, or a "pipe". Between all these, it's 
pretty easy to set up all kinds of different resonant structures.<<

That's kind of like what I was doing a few years back on my Nord Modular 
strings. It took a lot of playing around to get rid of the effect of it 
sounding like it was playing through a tube, and instead, give a somewhat woody 
tone. They still sound pretty good. But depending on the note played, I'd 
sometimes get weird sounding resonances, or not so pretty timbres. That's why 
I'm trying to move on to other techniques that are more controllable and 
predictable and can get me closer to a real violin or cello.

>>There are several electric guitar preamps out there (Roland comes to mind)
that uses modelling to give you a decent sounding "Acoustic Guitar" from your 
electric. Doesn't sound bad either.  Now, a guitar body is a little different 
that a Cello body, and too big for Viola or Violin, but it would be interesting 
to see what it would do as far as helping create a more "acoustic" timbre from 
a synth.<<

The answer is no. I have a Boss "Acoustic Simulator" pedal. What it sounds like 
is EQ to give a more full and round tone, plus what sounds like an exciter that 
gives crispy sharp highs. Doesn't help for cello type stuff. The body shape of 
a cello is much more complex than an acoustic guitar and so are the resonances 
it creates.

>>I've thought it would be interesting to try a Helicon VoicePrism VoiceCraft
on a modular synth as well, since it allows you to define vocal tract
resonance...<<

You can also try the Nord vocal filter as I did. It doesn't provide the 
numerous notches and peakes for a wood-like tone though, and it probably leaves 
some big holes in the audio spectrum. But it could be fun to run ontop of an 
already good cello patch.

>>What I'd love to see, would be a digital module that would, probably with a
computer interface, allow you to define a resonant structure and then use it
as a filter in your system.  All the individual components to do this
already exist. To varying degrees the Yamaha VL series used software to do
this, and the Korg Z1 did this as well...<<

I don't think the VL or Z1 synths have much control over complex resonant 
structures. Have you heard the VL violin? Yikes. But there is a single piece of 
software currently available that can do the most complex resonant structures 
you'd ever want to do in real-time. It's "Spectral Delay" by Native 
Instruments. That's what I'm trying to get closer to on my Nord.

-Elhardt

:> >>There are several electric guitar preamps out there (Roland

:> The answer is no. I have a Boss "Acoustic Simulator" pedal...

:> >>I've thought it would be interesting to try a Helicon
:> VoicePrism VoiceCraft


Yea, I knew none of the above would give me "cello" or "viola", but was
mainly interested in the "acoustic" qualities they may or may not have added
to an electronic sound...

Even once you get the "sound" part nailed, you still have a problem playing
the thing convincingly from a keyboard. I know you use a Yamaha Breath
Controller for a lot of stuff, and that gets you part of the way there. For
brass and wind simulations, it's great, but there's just so so many little
nuances a string player has available - vibrato is never constant - how
close to the bridge is the bow - upstroke with the bow or down stoke? - is
it an open position or fingered? - the same note (depending on where in the
range it is) can be played, using different positions, on several different
strings, yielding a completely different sound...

A keyboard like the AS French Connection might help with some of
inflections - and combined with a breath controller you could get pretty
expressive. If you manage to pull this off, even part way, kudos to you!
You've got more patience than I!


Les

Les Mizzell writes:
>>Even once you get the "sound" part nailed, you still have a problem playing
the thing convincingly from a keyboard. I know you use a Yamaha Breath
Controller for a lot of stuff, and that gets you part of the way there. For
brass and wind simulations, it's great, but there's just so so many little
nuances a string player has available - vibrato is never constant - how
close to the bridge is the bow - upstroke with the bow or down stoke? - is
it an open position or fingered? - the same note (depending on where in the
range it is) can be played, using different positions, on several different
strings, yielding a completely different sound...<<

One doesn't need to imitate every possible obscure form of playing to get a 
realistic and expressive violin/cello emulation. Some of what you've mentioned 
I've already got implimented. Some of the other stuff isn't important 99% of 
the time and can be dropped. The playing dynamics will still end up sounding 
far more natural than using a sampled violin/cello. That's the goal.

BTW, interesting thing about upstrokes vs downstrokes. I'll have to investigate 
more, but I don't think there is any noticable audible tone difference between 
the two. The waveform is virtually the same except inverted and backwarks.

-Elhardt

:> BTW, interesting thing about upstrokes vs down strokes. I'll have
:> to investigate more,

Oh, I play violin by the way....

If you heard the two tones, one upstroke and one down stroke side by side,
you couldn't really tell the difference.

However:

Down stoke tends to be slightly stronger, and considered the "dominant"
stroke (my terminology). If you study some string scores with bowing
notated, you'll quickly see why certain strokes are called for in specific
places.

Where bowing really comes into play is repeated notes. You can obviously
repeat the same note alternating up/down strokes. This sounds different than
repeating a note during a single bow stroke by slowing/stopping the bow.
It's almost a single vs. multiple trigger thing, but not quite.

This is the same problem you run into when trying to imitate a snare drum
roll with only a single snare sample. Although they're pretty much the same,
you really need both left and right hits to pull it off and sound decent.

Looking through some of the Garritan String samples I have here, there's
many that are set up with the keyboard split, with down strokes on one side
and up strokes on the other. Individually they sound pretty much the same,
but playing a legato passage certainly sounds more authentic when you
carefully work out the "bowing" on the keyboard the same as a real player
would do.

Runs are probably the hardest to get right. Depending on the score, a run
could be done in just one stroke, or depending on how it's phrased, divided
up over multiple strokes in one direction or another. When played on a
keyboard, each note in a run is pretty much a discrete event. When real
players do it, there's a certain amount a "smearing" going on that makes
string runs sound the way they do. About the only way to get close to that
is to overdub a number of times to build up each section, using at least 4
or 5 tracks for 1st violin, another number for 2nd...and so forth.

I really look forward to hearing what you accomplish during your quest.
Getting a great string sound is one of the things that many keyboard players
fret over and never get close, because they have trouble thinking like a
string player. You and I know this, of course, but you can't simply call up
a string patch and plonk down a series of chords and expect it to sound
anything like the real thing, regardless of how great the
samples/programming being used is. You have to be really conscious of the
phrasing and dynamics of every single voice and where it's leading to even
begin to approximate a good string section.

Good luck! Keep posting sound samples as you work on this. I'm sure I'm not
the only one interested in how it all turns out.


Les

Les Mizzell writes:
>>Oh, I play violin by the way....<<

Ah ha! Now I understand why you seem more intested in a synthesized cello since 
you can just play a real violin if that's the sound you need.

>>Where bowing really comes into play is repeated notes. You can obviously
repeat the same note alternating up/down strokes. This sounds different than
repeating a note during a single bow stroke by slowing/stopping the bow.
It's almost a single vs. multiple trigger thing, but not quite.<<

The up/down strokes will be applicable in how I want to handle a more legato 
playing style.  Just as you imply, one doesn't want a completely new note 
attack but something that is shaped or slurred differently to sound more 
natural.  Just as on the Yamaha VL series you don't need to blow a new note 
into the BC-3 everytime you play one on the keyboard.

>>Looking through some of the Garritan String samples I have here, there's
many that are set up with the keyboard split, with down strokes on one side
and up strokes on the other. Individually they sound pretty much the same,
but playing a legato passage certainly sounds more authentic when you
carefully work out the "bowing" on the keyboard the same as a real player
would do.<<

In my case, I'm going to have to sense legato vs staccato touch and adjust for 
that. But that will only work monophonically.

>>When played on a
keyboard, each note in a run is pretty much a discrete event. When real
players do it, there's a certain amount a "smearing" going on that makes
string runs sound the way they do. About the only way to get close to that
is to overdub a number of times to build up each section, using at least 4
or 5 tracks for 1st violin, another number for 2nd...and so forth.<<

That's kind of what I did in the orchestral scale run in the MP3. I'm 
concentrating on getting a solo instrument sound, so the only way to get a full 
orchestra is to do just what you say anyway.

>>but you can't simply call up a string patch and plonk down a series of chords 
and expect it to sound anything like the real thing, regardless of how great 
the samples/programming being used is. You have to be really conscious of the 
phrasing and dynamics of every single voice and where it's leading to even 
begin to approximate a good string section.<<

That's one of the reasons for my quest. To do for strings what Yamaha did for 
brass and woodwinds. However, there will always have to be some compromises 
made.

>>Good luck! Keep posting sound samples as you work on this. I'm sure I'm not
the only one interested in how it all turns out.<<

I'll do that. My strings are split into two projects. One is for me to write a 
string synthesizer using the synth techniques I used to synthesize that single 
violin tone. The other is to come up with a accurate filterbank patch to 
process my current synths to sound more realistic. If the first one doesn't 
come to reality, I'll still have the second less accurate one.

-Elhardt

> playing style.  Just as you imply, one doesn't want a completely new note 
> attack but something that is shaped or slurred differently to sound more 
> natural.  Just as on the Yamaha VL series you don't need to blow a new note 
> into the BC-3 everytime you play one on the keyboard.

What I like about the VL7 is how it can emulate some of the BC work
just from keyboard playing (without a BC connected). There is a
mix of velocity and aftertouch, such that for a short time after
you hit a key, the velocity controls a certain parameter (like
pressue of the bow against the string), and after a certain time
(adjustable) the control for this same parameter is taken over by 
aftertouch. It's more or less a crossfade operation, which could
easily be done on a Modular with a dual VCA and a VCLAG.


> In my case, I'm going to have to sense legato vs staccato touch and adjust
> for 
> that. But that will only work monophonically.

Is there a clever way to handle it *duophonically* at least?
Like a bow working on two strings at the same time? (I'm
not a violin player, so I can only guess, but I think they
do this, don't they?)
Like, if you hold one note on the keyboard and then play a second
one, you'd have to emulate the bow being moved and touching a
second string? Does this make sense?

JH.

-------------------------------------------------
debitel.net Webmail

:> Ah ha! Now I understand why you seem more interested in a
:> synthesized cello since
:> you can just play a real violin if that's the sound you need.

Oh no - my violin playing can repel cats from miles away. If I ever need
real violins, I'm gonna hire somebody that really knows what they're doing!
My main instrument has always been piano, then organ (nothing like a good
Bach Fugue to clear the mind...), then bass and contrabass clarinet, which
is odd considering that the damn thing is bigger than I am....

I only picked up violin my last two years of college, and believe me, that
was a LONG time ago... because I got tired of the bass clarinet. I was in
the orchestra, but I was always last chair. I never got very good at it. For
fast passages, I'd just play every other note! It was the only way I could
keep up!

Why my interest in cello? I dunno! Maybe because my wife plays cello. I
really love the extreme high range of the instrument, although she hates
playing up there....

Heh...

:> Is there a clever way to handle it *duophonically* at least?
:> Like a bow working on two strings at the same time? (I'm
:> not a violin player, so I can only guess, but I think they
:> do this, don't they?)


Violins, and other bowed instruments are fully capable of double and triple
stops, though, due to tuning restrictions, only certain combinations of
notes are possible. Forget tight tone clusters, for example..

There's a reference here:

http://gigue.peabody.jhu.edu/~mathews/Orchestration/String%20Notes/multistop
s.html

or grab a book:

I've always like "Principles of Orchestration" by Rimsky-Korsakov, but some
folks don't as it's pretty specific to his own music....

--- In motm@yahoogroups.com, elhardt@a... wrote:
> Les Mizzell writes:
> >>Even once you get the "sound" part nailed, you still have a 
problem playing
> the thing convincingly from a keyboard. I know you use a Yamaha 
Breath
> Controller for a lot of stuff, and that gets you part of the way 
there. For
> brass and wind simulations, it's great, but there's just so so many 
little
> nuances a string player has available - vibrato is never constant - 
how
> close to the bridge is the bow - upstroke with the bow or down 
stoke? - is
> it an open position or fingered? - the same note (depending on 
where in the
> range it is) can be played, using different positions, on several 
different
> strings, yielding a completely different sound...<<
> 
> One doesn't need to imitate every possible obscure form of playing 
to get a 
> realistic and expressive violin/cello emulation. Some of what 
you've mentioned 
> I've already got implimented. Some of the other stuff isn't 
important 99% of 
> the time and can be dropped. The playing dynamics will still end up 
sounding 
> far more natural than using a sampled violin/cello. That's the goal.
> 
> BTW, interesting thing about upstrokes vs downstrokes. I'll have to 
investigate 
> more, but I don't think there is any noticable audible tone 
difference between 
> the two. The waveform is virtually the same except inverted and 
backwarks.
> 
> -Elhardt


Just to give you some other specific violin sound items to think 
about: try listening to (for instance) a g2 played on E string and 
then to the same g2 played on A string - quit a different sound. Or a 
G (1 octave below g1) on the D string compared to the same G on the G 
string. Both are sounding dramaticaly different. And what about 
spiccato, short spiccato, broad spiccato, staccato, detaché and
many 
more bowing techniques that will nut just slightly, but totally 
change the sound character? Les already indicated that a lot of 
parameters are needed to get a convincing string sound and I'm not 
talking about "every possible obscure form of playing" here! Your 
answer could be that only the good string players can manage these 
techniques, but then I guess in trying to synthesize a 'real' string 
sound you want to go further then playing 'Old Mac Donald' on two 
strings?
At the other hand, Ken, I know you will go to extreems to get it 
right so keep it up and amaze us all!

Michel Havenith, Netherlands