Conference napalm::commusic_v1

    Some can, some can't...
    
    MANY keyboards are limited to one voice at a time. (The thing that
    sold me on my DX21 was it's ability to do 2. That's old hat now,
    but at the time I hadn't seen anything that could do that.)
    
    The Yamaha Fb01 can play 8. So can lot's of others.
    
    One thing to be concerned with is the number of notes that can be
    played at one time. If you have an 8 note polyphonic synth and are
    attempting to use 8 different timbres, you'll find yourself with
    the equivalent of 8 MONO synths and unable to play any chords.
    
    The number of notes assigned to a timbre is usually static and defined
    at the beginning. I've not seen any units which will borrow unused
    voices from other timbres. Has anyone?
    
    Edd

RE < Note 888.1 by JAWS::COTE >

	Here's a list from memory:

	Yamaha FB01 and TX81Z	-	Up to 8 simult. voices
	Casio CZ-101, 1000	-	4 simult mono voices
	Casio CZ-5000, CZ-1	-	8(?) simult. voices (mono)
	Most samplers (Mirage, s-10, s-50, EMAX, ...)
	Chroma Polaris    	-	2 simult. voices
	Yamaha DX-21		-	2 simult. voices
	Ensonique ESQ-1		-	Up to 8(?) simult. voices 

	I'm sure there are more, but memory fails.  This, of course is mostly
on the lower-end of electronic keyboards.

	To answer Edd's question, I believe the ESQ-1 will steal voices from
the other when needed and not being used.

	Dan

RE < Note 888.2 by AKOV75::EATOND "Deny thyself" >

	Yamaha FB01 and TX81Z	-	$300 and $450, respectively
	Casio CZ-101, 1000	-	$200 and $300, respectively
	Casio CZ-5000, CZ-1	-	??? and $600
	Most samplers (Mirage, s-10, s-50, EMAX, ...) around $1000 and going up
	Chroma Polaris    	-	?? (not made any more)
	Yamaha DX-21		-	$700
	Ensonique ESQ-1		-	$1300

	(BTW, these prices are new prices.  Amounts may vary from dealer to
dealer.  And, of course, there's always the used market.)

	From what you described, the Yamaha units (fb01 and Tx81z) would be the
perfect 'black boxes' for you.  They seem to be the most flexible in regards to
voice assignment; you can have any number of instruments with any number of 
voices until you use up the eight available.  The Casio CZ's will only allow
mono multi-timbrality (that word again!) and others usually split the voices
right in half.  I have the FB01 and have been more than satisfied with it.  
There are times when I wish I had gotten the TX81Z because it can be programmed.
But when it comes down to it, I really don't have any time to learn FM synthesis
anyway.  I want to spend more time making music than making sounds.

	Dan

    re: .2
    
    When all 8 voices on the Ensoniq ESQ-1 are being used and a new
    key is pushed, it "steals" the voice that is producing the "oldest"
    (least recently pushed) note.
    
    The ESQ-1 has an overflow mode that allows you to chain two or more
    ESQ-1 (and theoretically one non-ESQ-1) together to get more voices.
    
    In overflow mode, whenever a voice would normally be "stolen", the
    new note is sent out over MIDI instead.  That is, if the machine
    has all 8 voices in use, it sends a note on over midi so that some
    other machine can sound that note.
    
    Very nice feature, especially with the introduction of the ESQm
    (ESQ-1 in a rack mount).
    
    BTW, the ESQ-1 can sound 8 voices at once and they can all be different
    sounds.  You can also split the keyboard, and even have each half
    be layered sounds (i.e. top half is piano layered with organ, bottom
    half is bass layered with cello, etc.)
    
    The layering stuff effectively allows you to have 6 oscillators
    per voice which can give you some very complex sounds.
    
    	db

Wow!

This notesfile has the fastest response time of any I've seen!  It seems
as tho' half of Digital would rather be out playing synths. 

Thanks for all the info.  Now, of course, I have a follow-up question:

ok, so i buy a yamaha FB01 (is that what it's called?).  Now, can I use
Dr. T's ECS (el cheapo sequencer--actually called MRS) and it's 8 track
capability to PLAY these 8 different voices??  Do I need other software?
(he asked cringingly)  Or is it just a matter of sending "escape sequences"
to the synth to activate a given voice?       

otra vez, muchas gracias

rich murray

    The DX21 also does the "take the oldest note" algorythym, but I'm
    not sure that addresses the issue. (Guess where I work?)
    
    If I split my keyboard and assign 4 voices to each side of the split,
    pressing 5 keys on one side of the split won't "borrow" an unused
    voice from the other side. ("Dynamic Allocation"). It will however,
    use the "oldest note" scheme relative to *that side of the split
    only*.
    
    Dave's description of the ESQ-1 leaves me unclear as to whether
    that unit does anything different.
    
    Edd

    You're all set. Just think of the FB01 as lots of little synths
    in an even smaller box. Set 'em all to different MIDI channels
    if you are so inclined.
    
    ... and we *would* rather be playing synths!!
    
    Edd

RE < Note 888.5 by TIPPLE::MURRAY >

>ok, so i buy a yamaha FB01 (is that what it's called?).  Now, can I use
>Dr. T's ECS (el cheapo sequencer--actually called MRS) and it's 8 track
>capability to PLAY these 8 different voices??  Do I need other software?
>(he asked cringingly)  Or is it just a matter of sending "escape sequences"
>to the synth to activate a given voice?       

	O.K., given that you have the computer, and the MIDI interface, you
will have no need of further software.  Also, MIDI has made it a bit more
user friendly than 'escape sequences'.  Voices are assigned to MIDI 'channels',
of which there are sixteen available.  You can tell the FB01, via its front
panel, which voices to assign to which channel, and voila!  Any sequencer
that's worth its salt will make this clear in its documentation.

	Dan

RE < Note 888.6 by HAMSTR::COTE >

>  "Dynamic Allocation"
    
>    Dave's description of the ESQ-1 leaves me unclear as to whether
>    that unit does anything different.
    
	Dynamic allocation is the correct term.  You described it right.  I was
told that the ESQ will steal an unused oscillator, assign it the proper
parameters, and use it on the side that requires it.  I believe this is a new
concept in synthesizers, at least on the low-end of the market.  Almost all
keyboards made since the onset of the MIDI era will use some voice-stealing
algorythm like Dave described.  The only one that doesn't (that I know of) is
my Juno 106.  After six notes are played, any further key pressed will produce
nothing (nada, as KMII would say).

	Dan

    The Polaris will actually play up to six different voices
    simultaneously.  It can be set up to listen to 8 MIDI channels;
    you can use up to 6 of them at one time.  I don't know what its
    voice stealing strategy is.  You can only set up the patches for
    the 8 channels remotely, i.e., by sending a program change message
    on the appropriate channel.
    
    Two other bi-timbral (i.e., 2 voices at a time) synths are:
    
    	Roland JX-10  - 2 6-voice-polyphonic synths on any two channels
    	Roland MKS-80 - 2 4-voice-polyphonic synths on two adjacent
    			channels
                        
    len.
    

    Another nice feature of the Polaris is that in addition to having the
    two different sounds coming out of the keyboard by setting up a  split,
    you can also have a 3rd different voice (e.g. a bass line, etc.)
    coming out of the sequencer. Three different sounds simultaneously!!
    Not bad...
    
    Mike D.
    

The PAiA 1200 did this many years ago for under $2K.  It also had keys 
represented by byte numbers, with a top bit on for key down, off for up.
An interposed micro-computer could process key numbers lots of ways.
Kind of a pre-MIDI concept.  It's gone now, of course.
Tom

    I'm glad this note was started; I was not aware that the FB01 could
    do 8 different voices simultaneously. Malcolm over at E.U. was trying
    to sell me the Roland D-50 super hot-rodded "linear synthesis" axe,
    which can do eight voices simultaneously, but I just wasn't crazy
    about paying ~$2K.
    
    My question is: Assume I have decided to allocate a couple of
    voices for strings, a voice for a fat synth-bass, a voice or two
    for a mono lead, etc, and I MIDI the FB01 to my MC-500 so I can
    have several parts playing simultaneously. I wish to record
    these voices in a stereo field. How do I adjust the levels of
    each voice individually so that I can, say, fade the strings
    in and out? Does the FB01 have separate outs for each voice
    so that I can use an 8x2 mixer? Or do I have to edit the velocity
    data in my sequence to make the FB01 play at the level I need? For
    that matter, does the FB01 have stereo outs or do I have to *simulate*
    stereo using a reverb or delay?
    
    Thanks muchly...
    
    Brian Yablon
    
                

RE < Note 888.13 by ORACLE::YABLON >

	Ahh, now we get to the buy-off's of bargain priced MIDI...

	The FB01 uses stereo outputs.  Each voice has only three options
to play with; output to left, right or both.  In this way, you need to do
a bit of gymnastics to get the sound you want.  I have not used more than
three patches at a time from my unit, and usually only two; one to the
left and one to the right.  No stereo sound of any one instrument.  There
is one parameter you can set that helps; patch volume.  While it can only
be set for the overall volume for the voice and is not avilable for fading
in and out separate instruments in real time (unless your a whiz at
pressing buttons in the multi-page menu configuration on the unit while your
instruments are playing), it is better than nothing.

	Dan

    As has been mentioned elsewhere, the TX81Z is similar to the FB01,
    so I imagine that the tradeoff you make in panning is that to get
    individual pan on each voice you need to double up voices by putting
    one on the left, one on the right, and adjust the volumes.  So,
    at best you can get 4 pannable voices.  Most of the time you do
    this a little and can flip between patches to get lots of virtual
    combinations.  Only problem is that there is a fraction of a second
    delay between patche changes when all the voices are turned off (at least
    on my rev 1.2 of an 81Z).  So, I make up for it by filling with
    drums or another synth.  Adding stereo reverb makes banking voices
    only on the left or right sound acceptable so you don't have to
    give up too many voices for the sake of panning, but can reserve
    voices for effects (delay, controlled pan, chorus generated in the
    81Z).  BTW - I'm absolutely sold on getting good reverb (like the
    MIDIverb 2) because it frees up voices that I had to dedicate to
    panning and effects, besides just sounding an order better.
    
    Steve

    re: 13, 14
    
    I use an FB01 and MC-500 together.  The FB01 does respond to MIDI
    volume.  The MC-500 can record this control change (CC).  I use an
    MKB-200 to generate this info via a slider on this keyboard controller.
    
    Dave

    Re: ESQ-1's "overflow mode"
    	
    It really works quite well- you can chain as many 1's or m's together
    as you can grub up.  But watch out- there's a hidden "gotcha". 
    	
    The gotcha is this- an ESQ voice isn't considered free until ALL
    of the envelopes that make up that voice complete.  This can take
    80 seconds or so for long-decaying sounds.  It *includes* envelopes
    that aren't making any sound, but haven't timed out.  Now, the OS in normal
    mode (non-overflow) will just steal this ancient note and reuse
    it, and no one will be the wiser.
    
    But in overflow mode, the OS will say "I have no free voices- overflow
    this note-on".  And the note dutifully goes out the MIDI port.
    	
    All well and good- except that (1) if you forget and leave the ESQ
    in overflow mode, and then unplug the downwind units, your ESQ will
    appear to start missing notes.  What's happening is that those long-
    enveloped quiet sections are holding voices as "active" but not
    making any sound, and hence you are running out of voices.  Very
    scarey- you quickly become convinced that the CPU is losing interrupts
    or something even more expensive to fix.
    	
    Corrollary (2)  The LAST esq in a MIDI overflow chain must NOT be
    set to overflow mode.  Otherwise, it will toss excess notes down
    the line to the (nonexistent) next module.  Checkbook terror ensues.
    
    Hint: For overflow mode to be really useful, trim unused envelopes
    to the shortest feasible time, to free up voices quickly. Especially
    be careful of "release rate".  Even though you may not HEAR a long
    release, if it's in the patch, it *will* hold the voice as being
    active!

Mention was made earlier that the FB01 is not "programmable" while the
TX81Z is.  Do I understand this correctly?  I am a big fan of old
instruments (krumhorns, rebecs, psaltrys, etc) that are unlikely to
be built in, so I need a way to roll my own.

I gather that the CZ101 lets you do this, but with the tradeoff that
any kind of complex sound limits you to 4 voices at once, and each one
mono at that.  Well, mono is ok for most of what I have in mind, but
ability to play chords within a voice without tying up redundant voices
would be nice, so the FB01's ability to have multiple notes per voice
is nice.

Maybe I am confused.  Could someone prepare a short table comparing the
max notes per voice and voices per unit for the CZ101, FB01, and TX81Z?
We are talking over 2:1 price differential here, and I want to go with
the CZ101 if it will do what I want.

    Okay. Okay already.  You people have me totally confused.
    This seems to be a case of feeding the  pascal  compiler 
    some COBOL.  Please fill me in.
    
    What is YOUR definition of a VOICE ?  To me a voice is a
    single patch, sound, sample.   What I am hearing here as
    VOICE seems to be referring to a device's abilty to  sus-
    tain  multiple notes  simultaneously  (i.e.  POLYPHONIC),
    while at other times it seems as I have described above.

    I own a DX-7.   I can play one voice at a time  and  can
    sustain up to 16 notes simultaneously in polyphonic mode,
    using a FOFO (First on First Off) algorithm.
    
    What I want is multiple polyphonic voices.  Is this  the
    FB01 and TX81Z ?  I mean it sounds close  (i.e. multiple
    voices or sounds),  but if what I am reading is  correct
    you are saying that they only have 8 notes of  polyphony
    to share between the n voices that are be used.
    
    Zach.

    You are correct. You get 8 voices (notes) to share between
     n number of patches ('instruments').
    
    Edd
    
    BTW- The Fb01 *is* programmable, but not without an outside editor.
         You can't do it from the front panel...
    
    

RE < Note 888.18 by LYMPH::DICKSON "Network Design tools" >


>Mention was made earlier that the FB01 is not "programmable" while the
>TX81Z is.  Do I understand this correctly?  

	Yes.

>I gather that the CZ101 lets you do this, but with the tradeoff that
>any kind of complex sound limits you to 4 voices at once, and each one
>mono at that.  

	True, if you mean complex as being 'anything that mixes the two
sound-producing lines in the CZ'.  The CZ is capable of 8-notes at one
time, but it can only do this using a single sound-producing 'line' of
programming.  More complex sounds (i.e. layered and mixed sounds that
are programmed differently or detuned) are limited to 4-note polyphony.

>Could someone prepare a short table comparing the
>max notes per voice and voices per unit for the CZ101, FB01, and TX81Z?

	I'm not sure how good a table I can do, but I'll try.

		CZ-101		FB-01		TX-81Z

Max notes
available	   8              8               8

Max 
multi-timbrality   4              8               8

Mono/Poly in
Multi-timbral     Mono          Poly/Mono       Poly/Mono
Mode

Programmable	   Y              N               Y


	Does this help?  Wanna buy my CZ?

	Dan

I am getting two conflicting messages here about whether an FB01 is
programmable (meaning I can define new voices).  The ad in front of
me here says it can be done, but it doesn't say how.  Clearly the front
panel is not up to it.  Is it done thru the MIDI port?  Are there
Mac programs that can drive this?  (The programming, not the playing.
I already have the playing program.)

What are some examples of the kinds of sounds where you need "complex"
programming on a CZ?  I am guessing this is for sounds in which the tonal
colors change over the lifetime of a note.  Like a plucked string that is
brighter at the attack?  Is this "multi-timbrality"? 

Can any machines of this class handle notes with long decay times relative
to the inter-note time?  If I rapidly pluck three different notes on a harp
(16th notes, say) I will hear all three notes at the same time.  Do I have
to program this myself using three channels (really tedious, notation-wise)
or will the synth assign the next note automatically to the next generator
in a group in which all are set to the same voice? Or do only the machines
that do "note-stealing" do this? 

RE < Note 888.22 by LYMPH::DICKSON "Network Design tools" >

> panel is not up to it.  Is it done thru the MIDI port?  Are there
Mac programs that can drive this?  

	Yes, via system exclusive messages over the MIDI ports.  I can't
say conclusively what's available for the Mac, but I'd venture to say that
it is probably available through third-party software firms.

>What are some examples of the kinds of sounds where you need "complex"
>programming on a CZ?  

	Complex sounds (i.e. layered or detuned sounds) can be like when you
want to combine a long sustained tone that has a slow atack with a quick,
biting 'wooden' sounding tone to get the sound rolling.  It's this kind of thing
that requires you to use up two lines of programmed patches.  But, due to the
type of synthesis Casio uses (phase distortion), some complex sounds can be
created using only one line of sound-producing programming.  This is because
PD synthesis changes the timbre of a sound over time, using the flexible
eight-stage envelope (assignable to any or all of the pitch, timbre and 
amplitude generators).  So in answer to your question 'Like a plucked string 
that is brighter at the attack', this could possibly be accomplished using
one line of 'oscillators' and utilizing a complex envelope structure, and still
retaining 8 note polyphony.  Is this "multi-timbrality"? No, multi-timbrality
is the ability of the unit to create four different types of patches
simultaneously.  Each of the four voices are capable of creating this (or 
greater) complexity of sound.

>Can any machines of this class handle notes with long decay times relative
>to the inter-note time?  If I rapidly pluck three different notes on a harp
>(16th notes, say) I will hear all three notes at the same time.  

	I'm not exactly sure what you're asking, but I will say that if a 
machine has eight-note polyphony, it will have enough envelope generators and
the like to decay each of eight notes to the fullest time they are programmed to
decay (unless, of course, your looking at the Korg Poly-800, which only has one
filter for all eight oscillators).

	Do this help?

	Dan

I'm getting the picture.  Still not clear on the sustain bit.  Say I press
a key (NOTE-ON), then release (NOTE-OFF), but the voice is defined as 
having a long decay (say one second), then that channel will be busy
for one second.  But say only 250ms later I press another key, sending
a NOTE-ON on the same channel.  I could see one of the following happening:

	1)  The decay envelope from the first note is aborted and the
		envelope starts over again for the new attack, at the
		new pitch.

	2)  The second note-on is ignored.

	3)  A new envelope is started while the old envelope is still
		going, both pitches sounding simultaneously.

It sounds like some models use method 1, some 2, some 3.   3 is most
preferred, otherwise I have to program around it.  Is 3 "polyphonic"?
If so, is the FB01 or CZ101 polyphonic within one channel, or do you
use up channels doing this (like playing chords).  Can I dedicate say
4 channels to being a polyphonic harp, while keeping the other 4
channels as 4 monophonic instruments?

Clearly I can get any of these effects by programming all 8 channels
myself with each channel being one "finger" of the harpist's hand,
but it would be nice if the synth took care of this bookeeping.

RE < Note 888.24 by LYMPH::DICKSON "Network Design tools" >

>                               -< Almost there >-
					Yup.

>	1)  The decay envelope from the first note is aborted and the
>		envelope starts over again for the new attack, at the
>		new pitch.

	This is the scenario used almost universally.

>	2)  The second note-on is ignored.

	I know of no synth that does this to if the first key is realeased.

>	3)  A new envelope is started while the old envelope is still
>		going, both pitches sounding simultaneously.

	As you correctly assumed, this can only be done if an instrument is 
playing polyphonically.

>If so, is the FB01 or CZ101 polyphonic within one channel, or do you
>use up channels doing this (like playing chords).  

	The FB-01 (and the TX81Z) can assign both polyphonic patches and
monophonic patches to separate channels simultaneously.  They take 
multi-timbrality one step further than the CZ101.  The CZ can only play
monophonic voices accross separate channels in multi-timbral mode.

>Can I dedicate say
>4 channels to being a polyphonic harp, while keeping the other 4
>channels as 4 monophonic instruments?

	This could be done on the FB, but it would be much easier to assign it 
like this:

		channel a: 4 note polyphonic harp
		channel b: monophonic instrument
		channel c: monophonic instrument
		channel d: monophonic instrument
		channel e: monophonic instrument

			where a,b,c,d,e are any midi channel.

	Dan

>		channel a: 4 note polyphonic harp
>		channel b: monophonic instrument
>		channel c: monophonic instrument
>		channel d: monophonic instrument
>		channel e: monophonic instrument

This looks like what I want.  Channel a being "4-note polyphonic" means
I can have up to 4 notes sounding simultaneously and the synth takes care
of assigning generators?

If the FB01 can do this and not the CZ101, then my choice is pretty clear.

RE < Note 888.26 by LYMPH::DICKSON "Network Design tools" >

>This looks like what I want.  Channel a being "4-note polyphonic" means
>I can have up to 4 notes sounding simultaneously and the synth takes care
>of assigning generators?

	Yupper.

>If the FB01 can do this and not the CZ101, then my choice is pretty clear.

	Then you mean you'll buy my CZ?



	<8^)

	Dan

    This note is a little old but if anybody else comes here looking
    for info...
    
    IBM used the FB01 in their "Music Feature" - a card that sits in
    the PC and does music synth/recording/MIDI. I think IBM also has
    programmer's reference manual for the system. This might provide
    additional material on FB01 program capability.
    
    The Music Feature was reviewed in the October issue of Keyboard
    magazine.