Conference napalm::commusic_v1

RE < Note 1834.0 by AQUA::ROST "Marshall rules but Fender controls" >

	So where does the DX21 fit in, which generates more than one timbre
at a time, but does so on a single channel?

>    1. Mono-timbral
>    
>    The synth generates one timbre at a time.  If it can do splits and/or
>    layers, each timbre responds over the same MIDI channel.

	If the synth only generates one timbre at a time, how can it do 
splits/layers?  Are the timbres on either side of the split unable to sound
simultaneously?  Are the layers simply layers of oscilators/operators within
the same patch?

	Dan
    

    Oh boy, we get to do it again!!
    
    ...and what about the average drum machine? According to the definition
    set forth by db (henceforth known as "db's dictum" or "Blickstein's
    Folly" depending on which side of the fence you're on) drum machines
    could not be considered 'multi-timbral' as they only receive on
    1 channel.
    
    On page 143 of the January '89 Keyboard there is a blurb decribing
    the YS100, YS200 and B200 as 'multitimbral'. Can someone describe
    how these devices implement this feature? One channel? Or many?
    
    Edd

    
    OK, Dan, you're right, make it "The synth generates one timbre at
    a time *or* if it can do splits and/or layers, each timbre responds
    over the same MIDI channel."
    
    
    
    

    How about this behaviour:
    Four different instruments available at a time.
    Four total sounds available at a time. (Quartet polyphony).
    ca. 18 instruments on hand at any time, meaning you can switch from
    instrument to instrument from note to note.
    Decisions are made automatically on how to distribute more than
    4 lines among 4 available channels.
    
    What's this called?
    Tom

    re: .2
    
    > drum machines could not be considered 'multi-timbral' as they only
    > receive on 1 channel.
    
    Edd,
    
    You are gonna lead us in circles because you're making the same
    mistake I pointed out in the other note.
    
    I agree (mostly) with your definition of 'multi-timbral'.  Where
    "we" disagree is what constitutes a "multi-timbral synth".
    
    I would not disagree that an HR-16 is not a multi-timbral synth on
    the basis of it not being "multi-timbral".  I would disagree on the
    basis that it is not a synth.
    
    By the way, I thought we had concluded that most of the world was on
    "my side of the fence", it's just that we were all 'bastardizing'
    the word.
    
    	db
    
    p.s. For whatever it's worth, where I disagree with "your" definition
         of multi-timbral is the implication in one of your replies that
    	 an Orchestra can be considered multi-timbral even if it
    	 plays ALL notes the same way (as a layer would), even though
    	 the orchestra sound is composed of many instruments.
    
    	 Fortunately, I know of no orchestra with this limitation,
         thus we can never debate this on real terms.

    < Note 1834.4 by ANT::JANZEN "Tom 296-5421 LMO2/O23 ECL MSI Test" >
>    Decisions are made automatically on how to distribute more than
>        4 lines among 4 available channels.
>    What's this called?

    Dynamically-allocated voices in a multitimbral architecture

    Multitimbral has to mean capable of playing multiple timbres
    independently (i.e., different notes) at the same time.  How it's
    done (i.e., via multiple MIDI channels or distinct note ranges (i.e.,
    splits)) is irrelevant.  That's the only definition that makes sense.
    
    len.
    

    re: .7
    
    	Len,
    		You sure about that? I would think that Mulitimbral
    	would mean multiple *timbres* (sp?). The timbre of an instrument
    	is what allows the listener to distinguish between a voilin
    	and a piano. Multitimbrality has nothing to do with how many *notes*
    	can be played at once, but how many *different sounds* can be
    	generated at once. If you want to discuss how many notes can
    	be played at once talk about Polyphony.
    
    	                             Mark      

    Yes, precisely.  The stuff about notes was that the notes *might* be
    different, i.e., the timbres don't *always* play in unison.  I don't
    considered a "layered" synth to be "multitimbral".  Since the multiple
    timbres in a layer *always* play in unison, they might as well be a
    single (admittedly more complex) timbre.  I understand the difference
    between polyphonic and multitimbral.  Remember, the fact that it takes
    two players to play flute and piano at the same time doesn't mean a
    synth layering a flute and piano together is playing two timbres;
    otherwise I'd have to count each oscillator/filter as a timbre, and
    timbrality would depend on the synth's voice architecture.
                                                             
    A layered sound is not multitimbral; a split is.  The point about
    distinct timbres is that they can be playing different lines (each
    line may be monophonic or polyphonic).  How that's done (split or
    multiple channels) doesn't matter.
    
    len.
    

RE: len

Here Here! Hear! Hear! (...)

I'm on len's side.

Chad

RE < Note 1834.9 by DRUMS::FEHSKENS >

>    A layered sound is not multitimbral; a split is.  The point about
>    distinct timbres is that they can be playing different lines (each
>    line may be monophonic or polyphonic).  How that's done (split or
>    multiple channels) doesn't matter.

	This is what Edd and I were trying to say all along back in 1815 (it was
a good year, no?).  The point that the synth most in question (Yamaha DX-21) 
also layers was merely coincidental.
    
	Dan    

    And I agree that would make it "multi-timbral".  I only would say
    that it would be wrong to market it as a "multi-timbral synth"
    because that means something entirely different.
    
    By the way, according to your definition, anything that can produce
    different notes with different "timbres" at the same time is 
    multi-timbral.  Note that splits aren't the only way to do that
    you know.  You can do that with velocity, pressure, and variety
    of other controllers.
    
    There are an AWFUL lot of multi-timbral synths out there by that
    definition.
    
    On multi-mode.  It may not be in the MIDI spec, but as I've said,
    several different manufacturers refer to it in their literature and
    it seems to do the same thing for each manufacturer.   Hopefully
    your statements about it not being in the MIDI spec isn't akin to
    denying that it exists?
    
    	db

      
    
    > By the way, according to your definition, anything that can produce
    > different notes with different "timbres" at the same time is 
    > multi-timbral.  Note that splits aren't the only way to do that
    > you know.  You can do that with velocity, pressure, and variety
    > of other controllers.
    
    I wouldn't buy into that myself. You've still got the same controlling
    arguments/parameters associated with each key. In a typical
    multi-timbral application, the parameters are separate and independant
    for each timbre.
    
    I see a split keyboard (a la DX-21) as 2 independant synths, each
    on the same MIDI channel, but responding to 2 *exclusive* sets of
    note-numbers. (Transposition allows each half to produce the same 
    pitch, although said half may have received a 91nn message where nn did 
    not correlate 1 to 1 with the pitch produced.)
    
    Did I say that right?
    
    Edd   
    

>    I see a split keyboard (a la DX-21) as 2 independant synths, each
>    on the same MIDI channel, but responding to 2 *exclusive* sets of
>    note-numbers. 
    
    You do?
    
    > you've still got the same controller arguments associated with each
    > key
    
    Yes, and that is most unfortunate isn't it?:
    
    How do I do pitch bend on one "timbre" but not the other?
    
    How do I use the sustain pedal on one but not the other?
    
    How do I do wheel or pedal controlled modulation without the other?
    
    How do I do channel pressure on one but not the other.
    
    How do I change the volume of one but not the other while I'm
    playing a sequence (MIDI volume effects both).
    
    etc. etc.
    
    Not all controllers are associated with keys.
    
    Or you do consider pitch bend, sustain, modulation, pressure, and
    MIDI volume, etc.  to be unimportant as metrics to whether or not
    something constitutes an "independent synth"?
    
    See, part of what I'm saying is that you're idea of a "multi-timbral
    synth" is not SUITABLE for most "multi-timbral applications".
    
    	db

    Sometime ago, we talked about a timbre as corresponding to an
    "instrument".
    
    Dan won't find any exposition for this in the MIDI spec, so I suppose
    that you'll have to take my word for it (or not) that the MIDI
    construct for an "instrument" (perhaps "timbre") is a MIDI *channel*. 
    
    Therefore for a synth to be able to fully play multiple "independent"
    instruments (and perhaps "timbres") it must respond on multiple
    MIDI channels.
    
    	db

    Thanks, Edd.
    
    Look, maybe I can make it real clear by going back to the source.
    If I would notate the "line" on a different staff, then it's a
    different timbre.  If the synth can play, via whatever means, music
    notated on multiple staves, then I'd call it multitimbral.  Velocity
    or pressure timbral switching or mixing only confuse the issue;
    they are not intended to provide "multitimbral" capabilities in
    the sense I have described here.
    
    Note that a "system" (e.g., the piano bass and treble clefs)
    corresponds to a single staff.  Also note that if two (or more)
    instruments play in unison throughout an entire piece, they would
    probably be notated on a single staff.  For example, a group of
    soprano voices singing in unison is a single timbre, even though
    each voice is distinct; they are all singing in unison.  Note also
    that polyphonic instruments can play in unison in this sense, even
    though they are playing multiple notes, if they are all playing
    the *same* set of notes.
    
    len.
       

 
    
    > you've still got the same controller arguments associated with each
    > key
    
    >Yes, and that is most unfortunate isn't it?:
     
     You didn't quote me properly. What I said was "controll*ING* arguments/
    parameters" in an attempt to find a generic description for such
    things as VCO freq, algorhythm, VCA; in other words, the parameters
    that any given architecture uses to produce a sound.
    
    >How do I do pitch bend on one "timbre" but not the other?
     
     Set PB to 0 on one.
    
    > How do I use the sustain pedal on one but not the other?
      
    Set PED to off on one.
    
    > How do I do wheel or pedal controlled modulation without the other?
      
    Set AMS or PMS to 0 on one...
    
    > How do I do channel pressure on one but not the other.
      
    DX doesn't respond to this. Moot point.
    
    > How do I change the volume of one but not the other while I'm
    > playing a sequence (MIDI volume effects both).
    
      Move the balance slider with your left forefinger....;^)
    
    > Or you do consider pitch bend, sustain, modulation, pressure, and
    > MIDI volume, etc.  to be unimportant as metrics to whether or not
    > something constitutes an "independent synth"?
    
      No, I don't consider them unimportant. Or relevant to the timbrality
      discussion.
    
    > See, part of what I'm saying is that you're idea of a "multi-timbral
    > synth" is not SUITABLE for most "multi-timbral applications".
      
     ...but the limited implementation doesn't exclude it from, in fact,
    being mutitimbral.
    
    Edd

    Len, and others in his camp,
    
    	At the risk of receiving severe flames...
    
    Timbre refers to the quality of *sound* that makes it possible to
    distinguish it from other sounds. I agree with Len about layered
    sounds actually being only a new timbre/sound. However by the strictest
    definition of the word "multitimbral" an instrument need only be
    capable of producing multiple sounds, period! This means that Casio
    202 (pre-MIDI) is multitimbral. It can make 49 different sounds.
    It can only generate these sounds one at a time but it can generate
    them.
    
    Perhaps all the confusion is over the fact that what we really are
    trying to describe is something more akin to SIMULTANEOUS
    Multitimbrality. That would be an instrument capable of producing
    more than one unique sound at the same time *in addition* to being
    able to play different music with each voice (as opposed to being
    layered where this is not possible.) 
    
    Len I realize that this term is misused, and therefore is misleading.
    If, however you are willing to add the word "simultaneous" to your
    use of the word "multitimbrality" I will be happy to rally to your
    banner in the great word definition battle. :-)
    
    				Mark
                                      
    
    ps. Is it safe to take off my asbestos underware now?

    re: Mark
    
    I think you're safe in assuming that the 'simultaneous' qualifier
    is inferred by both camps.
    
    Edd

    re 18, 19 - yep, it ought to have been clear from my discussion
    about lines and staves etc. that simultaneity was implied.
    
    len.
    

    all of this indicates what a lousy job the MIDI spec designers
    did way back when....   
    
    dbii

RE < Note 1834.21 by DNEAST::BOTTOM_DAVID "slimy cold weather!" >

	I dunno...  I think situations like the DX21's multi-timbrality is
not necessary to be covered by the MIDI protocol.  That's more the individual
manufacturor's prerogative.

	Dan

    Edd,
    
    Forgive me for not making this clearer.
    
    What if I wanted to USE ptich bend, modulation, sustain, etc. on
    both instruments but at different times?  Can you do that on
    your multi-timbral DX?
    
    And folks, once again I find myself having to remind people that
    what I (and I believe Len) are saying is that "multi-timbral synth"
    when used to describe a synths capabiltiews implies Multi-mode or 
    its equivalent.
    
    I think there's any point in discussing the meaning of multi-timbral as
    a general adjective because, I don't think we disagree in any
    significant way with what "multi-timbral" means as a general adjective. 
    Only when applied to the word "synth" to describe a feature of that
    synth.
    
    Let me state some resolutions and see where the real disagreement
    is (if any).  My hope is to formalize the EXACT points we are
    discussing in the hopes that we won't waste much more time.
    
    	1) A split is a multi-timbral application and therefore it
    	   could be said that said synth is capable of a limited form
    	   of multi-timbrality.
    
    	2) What is generally referred to as "MIDI multi-mode"
    	   is ANOTHER multi-timbral application.  Basically it is the
    	   ability to respond on multiple MIDI channels with distinct
    	   MIDI programs, although some synths might not call it
    	   "multi-mode".
    
    	3) The phrase "multi-timbral synth" is well established by
    	   its usage in product literature to imply the "multi-timbral
    	   application" described in (2) and ONLY the application
    	   described in (2).
    
    	4) A synth that is capable of some multi-timbral applications
    	   (such as those mentioned in (1)), but NOT the application
    	   mentioned in (2) should NOT be referred to as a multi-timbral
    	   synth" as this differs from the established usage of that
    	   term and could be deemed "misleading", even though such usage
    	   is formally correct.  
    
    	   The problem is that "multi-timbral synth" has been stripped of 
    	   its ability to be automatically and clearly identified as a 
    	   "formal" description.
    
    	5) The meaning of the word "multi-timbral" as a GENERAL ADJECTIVE
    	   implies the ability to, for some case, independently AND/OR 
    	   simultaneously generate different "timbres".  (And God only
    	   knows how we define "timbre").
    
    For the record, "my" position agrees with all of these points.
    
    I think (5) seems to be a "given" and is irrelevant to (1)-(4).
    
    I think there is general agreement on (1) and (2).
    
    I think I have established (3).  I just don't think there can be much
    question of that definition as anyone can observe a very consistent
    usage of it in product literature. Consistent across product lines,
    consistent across brands, etc.
    
    And I think (1), (2), and (3) imply (4).
    
    	db

    
    Yes, but let's look at mode 4, omni off mono.  This is how a
    Casio CZ operates for multitimbral stuff.  It also operates
    monotimbral, or with splits and layers in mode 3, omni off poly.
    It can never operate in the other two (omni on) modes.
    
    OK, now take an ESQ or an MT32 doing dynamic allocation.  Is that
    omni off poly or omni off mono?  Do these boxes even respond to
    mode change messages if you send them via sys-ex??  
    
    Or for further confusion, the Casio CZ5000 implements omni off/mono to
    the letter, all channels respond monophonically *only*.  However, locally,
    either from the keyboard or via the sequencer you can have mixed
    mono/poly operation....you just can't have it via MIDI.   ?????
    
    I don't think that the designers envisioned the various "multitimbral"
    configurations that are available today.   In 1984, who would have
    thought you could buy a $150 Casiotone that can play three polyphonic
    parts simultaneously over MIDI that would be sold in toy stores?
    
    

    In all fairness I don't think the designers of MIDI anticipated
    anything like what it has become, however, the spec should encompass
    definitions of functions that would probably be universal, like
    multi-tibral etc. The spec currently has many holes in it,
    necessatating the production of mergers, splitters, filters etc., as
    well as the varying interpertations of those features it does spec. If
    DECNET was as poorly spec'd as MIDI, DEC would not have as large an
    installed base. MAP is a good example of how bad this can get when a
    spec is poorly defined. MIDI is a digital communications network for
    musical instruments, designed by analog engineers... 
    
    dbii
    
    

	This is getty pretty funny.  For fun, should we have
	another topic debating the correct applications of
	voice, patch, and instrument?

	:^)   8*)  &*}

RE < Note 1834.23 by DREGS::BLICKSTEIN "Yo!" >

>    	1) A split is a multi-timbral application and therefore it
>    	   could be said that said synth is capable of a limited form
>    	   of multi-timbrality.

	I don't see any need to add the word 'limited' since, by your wording,
you haven't called it a 'multi-timbral SYNTH'.
    
	Roland seems at least to cooperate with points 3 an 4, but I doubt 
you'll find it formalized in any way.  Yamaha DEFINITELY disagrees with this in 
the example that dbii and I have given.
    
	Dan

    > the correct applications of voice, patch and instrument?
    
    That would be hopeless.   Multi-timbral synth is one of the precious
    few terms that has a consistent usage NOT specifically defined by
    the MIDI spec.
    
    Now on the "goodness" of the MIDI "spec".
    
    As a "standard" I think MIDI is almost unparalleled in terms of its
    success.  Yes there are glitches, but my experience has been that
    IT WORKS.  Stuff that says MIDI on it, can be used together.
    
    Think about that.
    
    Take a large program written in your favorite standardized computer
    language.  Now compile it with a different vendors compiler.  How many
    of you think it's going to both compile and RUN the same?
    
    All of you with your hands up go stand in the corner.
    
    In my experience, the MIDI connections we make between our instruments
    are more reliable, understandable and intuitive than the audio and
    power connections we make.   Think of how many standards there are for
    sending an audio signal, and how POORLY documented the inputs and
    outputs are on most devices.
    
    Also, how many of you knew about MIDI 5 years ago?  I mean, the final
    criteria of the goodness of a "standard" is how well it becomes
    accepted and adopted.
    
    On that "note", I rest my case.
    
    	db

    re .28: what he said!!!!!! 
    
    Yeah, yeahm, yeah, Dave, you got it.  Splits, multitimbrality, filters,
    etc. have nothing to do with interoperability.  A good standard
    is just as smart about what it leaves out as it is about whatit
    says.
    
    len.
    

>	Roland seems at least to cooperate with points 3 an 4, but I doubt 
>	you'll find it formalized in any way.
    
    Dan,
    
    It's not just Roland.  It's Roland, Kawai, Korg, Ensoniq, to name
    a few.  It may not be formalized but it's used very consistently.
    
    > Yamaha DEFINITELY disagrees with this in 
    > the example that dbii and I have given.
    
    I looked for the example that you gave but couldn't find it.  Could
    you provide me with a reference?
    
    Dave pointed out that Yamaha describes the DX7IID (or something like
    that) as being multi-timbral.  I asked what multi-timbral features
    it had (splits, vel-x-fades, or multi-mode-like stuff) and received
    no answer.
    
    I took the trouble to call my favorite music store and asked about the 
    DX7-IID.
    
    Guess what?  It's multi-timbral (with E!) by MIDI CHANNEL (i.e.
    like multi-mode).
    
    The new YS series is multi-timbral BY MIDI CHANNEL.
    
    Everything I seem to be able to find out says that Yamaha generally
    avoids using "multi-timbral synth", but when they do, it is consistent 
    with my camp/side/notion/etc.
    
    	db

RE < Note 1834.30 by DREGS::BLICKSTEIN "Yo!" >
    
	I only mentioned Roland 'cause I couldn't think of the other 
manufacturers you had mentioned.
    
	As far as the reference, it's in the catalog Yamaha put out called
"Introducing Yamaha Digital Music Systems".  Its a white glossy catalog.  I
think I got it in the mail (have a subscription to AFTERTOUCH, ya know).

	In at least two places it describes the DX-21 as being 2 voice 
multi-timbral.

	To be fair, as far as I know the DX-21 is the only synth in question.
Since I owned one and Edd still owns one I'm sure that's why we tend to harp
on the point.  All other multi-timbral units I know of definitely split their 
timbres across different channels.  So they do not need to be involved in the 
discussion.  It seems that the type of multi-timbrality Edd and I want to 
include in the definition was not seen as marketable to many manufacturers.  And
rightfully so.  Splitting the timbres across different channels (or at least 
making that possible) is definately preferable!

	I never said the DX7II was multi-timbral without channel-implementation.

	Dan
    

    Calling it "two-voice multi-timbral" is not quite as objectionable
    as calling it "multi-timbral".  I have no big problem with that as
    long as those two places that use the word "multi-timbral" describe
    how that is achieved (splits, layers, etc).
    
    We can only speculate why Yamaha didn't call it multi-timbral.
    
    "Not Marketable" was one possible motive, but obviously I have
    my own theory.  I would expect they didn't call it "multi-timbral" 
    because the DX-21 did NOT have the functionality that most other 
    manufacturers describe as multi-timbral, and thus to have called
    it that would have been confusing, perhaps misleading.
    
    	db

RE < Note 1834.32 by DREGS::BLICKSTEIN "Yo!" >

>    Calling it "two-voice multi-timbral" is not quite as objectionable
>    as calling it "multi-timbral".  

	What's the difference?  The number of timbres available has less to do 
with it than the fact that two are available SIMULTANEOUSLY.  I'd have expected
you to be more bothered that they didn't call it 'bi-timbral'!

>    I have no big problem with that as
>    long as those two places that use the word "multi-timbral" describe
>    how that is achieved (splits, layers, etc).

	They mention single, dual and split modes.  BUT!!! they use the very 
same description for the synth(s) that DO split timbres across independent MIDI
channels!  Both are called 'two-voice multi-timbral', and both are described as
implementing 'single, dual and split modes'!  There is no further description
in either case.
    
	Dan

Since when has a manufacturer been worried if people would be confused or
mislead by their claims???? :-)

Chad

    re: .33  "Take That"
    
>>    Calling it "two-voice multi-timbral" is not quite as objectionable
>>    as calling it "multi-timbral".  

>	What's the difference?  The number of timbres available has less to do 
>with it than the fact that two are available SIMULTANEOUSLY.  I'd have expected
>you to be more bothered that they didn't call it 'bi-timbral'!
    
    The important distinction is not the quantity (two), but the unit
    of measure (voice).
    
    If they had called it "Bi-timbral" I would have objected.  My claim
    is that that implies two CHANNELS (as on the D-50).
    
    The reason why I find it less objection (but still somewhat
    objectionable) is that it IDENTIFIES (although vaguely) the unit
    of timbrality (voice) as being different from the (in my opinion)
    "implied" measure of channel.
    
    But....
    
    Given that they use the same term to identify multi-channel
    synths, this clearly was not their intention. 
    
    I think that is somewhat misleading.
    
    If I had bought the DX-21 on the basis of it being "two voice
    multi-timbral" on the assumption that that term meant THE SAME THING
    (which it doesn't) as it did for the Yamaha 2-channel synths, I think I'd
    have cause for feeling mislead, wouldn't you?
    
    So to summarize, I think that you have established NOT that Yamaha
    uses one particular meaning; you have established that Yamaha uses it
    inconsistently.
    
    	db

RE < Note 1834.35 by DREGS::BLICKSTEIN "Yo!" >

>    So to summarize, I think that you have established NOT that Yamaha
>    uses one particular meaning; you have established that Yamaha uses it
>    inconsistently.

	I disagree.  I believe it shows that Yamaha finds the concept of 
multi-timbrality a separate issue from the concept of timbres isolated by
MIDI channels.

	Dan
    

    Anyone still having any doubts about the definition of multi-timbrality
    might wish to check the "Technology" column in the Feb. 89 issue
    of Keyboard magazine (Randy Newman on the cover).
    
    They supply a fairly good (but not perfect) definition of
    "muti-timbrality": the ability of each "voice" to play a different sound.
    
    	db

RE < Note 1834.37 by DREGS::BLICKSTEIN "Conliberative" >

>    They supply a fairly good (but not perfect) definition of
>    "muti-timbrality": the ability of each "voice" to play a different sound.

	That is the definition I have defended all along.  Was there more to it
than that?  (I don't have my Feb issue handy)  Did they bring up the MIDI 
Channel issue?

	Dan

>>    They supply a fairly good (but not perfect) definition of
>>    "muti-timbrality": the ability of each "voice" to play a different sound.

>	That is the definition I have defended all along.  Was there more to it
>than that?  (I don't have my Feb issue handy)  Did they bring up the MIDI 
>Channel issue?
    
    Perhaps you did, but I've reread some of your notes and I don't think
    so.
    
    The key word is EACH as in "EACH voice".
    
    If you ever have described an instrument that only does splits as
    "multi-timbral" than you certainly don't agree with that definition
    as EACH voice can NOT produce a different sound in a split�.
    
    	db
    
    � Two-voice synths don't count ;-)
    

    I think this is a bit sematical, no?

    The phrase "EACH voice" is technically *incorrect*.  Consider the
    Proteus, for example - 32 voices, but only *16* "voice" multi-timbral.
    In terms of the earlier definitions, the Proteus isn't truly
    "multi-timbral", no? 

    I didn't read the article, but I've always viewed timbrality as
    follows: 

	mono-timbral: one patch� at a time, period

	bi-timbral: two patches may play at a time, regardless of MIDI
		    channel assignment/interaction.  For example, my
		    OB-Xa is "bi-timbral"; it can play two sounds at
		    once, but cannot respond to more than one MIDI channel
		    at once.

	multi-timbral: can play several patches simultaneously.

-b

�   (I've always wanted to use a superscript ...) patch refers to a
    distinct sound played by any one pressed MIDI note.  Layers don't
    count. 

RE < Note 1834.39 by DREGS::BLICKSTEIN "Conliberative" >

>    The key word is EACH as in "EACH voice".

	O.K., you caught me on a technicality...  My purpose, though, was to
point out the absense of the reference to MIDI channels.  I just wanted to pull
your chain...

	Brad's got a good point, there, in the last reply.  Along with the 
Proteus, you could include the TX802, which has 16 voices, but can only put out
eight unique patches.

	Dan
    

    The point Brad raised is why I had the caveat in parenthesis in the
    original reply:
    
>>    They supply a fairly good (but not perfect) definition
    
>Proteus, you could include the TX802, which has 16 voices, but can only put out
>eight unique patches.

    Same is true of many other synths.  The MT-32 has that same problem
    and yet it says "Multi-timbral" right on the box.
    
    I think the intention was clear however.   And in fact, the terms
    "splits" and "layers" ARE addressed in the article but separately and
    distinctly from "multi-timbral".
    
    	db

    
    	Rathole Revisited.......
    
      C'mon boys. If you can get a minimum of 4 different voices out
    of a box simultaneously, it's multi-timbral. Period.

    ...and what if they're all the same timbre?
    
    Edd

	Well, if you're all having enough fun with THIS little problem 
of definitions, you probably don't want to try answering a question 
that a (pipe) organist friend of mine asked ME this week-end.

(para-phrased)	How many pipes is the Proteus box equal to ?

Yikes;  Poly_multi_dynamically_reallocated_voice_stealing_109_ways_til_sunday_
Try_it,_You'll_like_it,_don't_think_in_terms_of_#_of_pipes_or_stops

	R

    
    I think you have to resort to the legal idea of what's 'reasonable'
    in cases like this. 
    
    So the answer to 'Is it multitimbral?' is in most cases 'Yes', these
    days.
    
    And the answer to 'How many Pipes is it equal to?' is 'helluvalot'.
    
    Richard.

    I think it's pointless and perhaps even impossible to come up with a
    failproof definition.  I doubt I'd have any problem shooting down any
    definition under 4 paragraphs.  That's why we compromise the word
    "multi-timbral" to some extent - to have a useful familiar term.
    
    	db

    Uhm, it's nice taking this definition out of Keyboard, but now
    I've got to ask, "what's a voice?".
    
    Matter of fact, I think we need a set of commonly agreed to definitions
    for at least the following:
    
    	patch
    	timbre
    	voice
    	part
    
    	etc.
    
    len.
    

    re: .50
    
    > What is a voice
    
    The article defines that too (in a most bizarre analogy).   And as 
    one would expect, the "definition" (in this case, I should really
    say "explanation") for "voice" is also full of flaws even though the 
    meaning is clear.
    
    Anyway, I'll pass on any attempt to give rigorous definitions for 
    these terms, although my original statements about "multi-mode"
    and equivalent concepts seem to be very consistent with the terms
    used in the literature and manufacturers terms.
    
    My main position here has been that the split and layer features do not
    a properly described "multi-timbral synth" make.  I think most people
    tacitly accept that now.
    
    	db

    
    Yep. Trouble is, even if you could arrive at rigorous definition of
    these and other terms, you'd still be no closer to finding out what 
    any one synth actually does for you. There's always going to be room
    for the description to 'float' a little with regard to the practical
    application of the machine. Witness Roland's 32 note polyphony, claims
    for programmability via a front panel, 'limited only by your own
    imagination' (ha ha)...
    
    Richard

    > I think most people tacitly accept that now.
    
      Not for a moment!!!! But, in the interest of not beating a dead
      horse, I've forsaken argument.
    
      I read the article also. (A box full of Rich Littles) I don't
      remember the concept of multi-channality (I made that up) even
      being addressed.
    
      Edd

    Gee Edd.
    
    I thought we had at least agreed that the use of the term by
    manufacturers as a synth feature is consistent with what I've said,
    even though it may be a bastardization of the true meaning 
    of "Multi-timbral".
    
    I thought that was well established by observation, and not merely by
    opinion.
    
    If you agree with that, then we really don't disagree at all cause
    that's the only point I've really tried to make.  If you disagree
    with that (empirical evidence), then IMO you truly would be beating
    a dead horse.  ;-)
    
    	db

    Hi-ho Silver...
    
    ;^)
    
    Edd

    
    
    	Re: 44,45
      Four voice(sounds) because of your MKS(10), which has four voices.
    The next step down is 2 voices, which is of couse bi-timbral.
    No one's marketed a 3 voice machine. Therefore I postulate that
    if simulteneous sounds > 2, the machine is multitimbral.

                 			Rich
    

    re: .56
    
    That definition is flawed.  By that definition, any synth that can do
    independent layers on both sides of a split is "multi-timbral" because
    that's 4 simultaneous sounds. But in fact, no manufacturer would call
    that "multi-timbral".
    
    Take my word for it - don't attempt a rigorous definition.
    
    	db

I'm just back from a trip, so I didn't get to argue with db early :-)

>     Take my word for it - don't attempt a rigorous definition.

I never take db's advice :-) :-)

I think the problem is that polyphony and multi-timbral are conceptually
orthogonal but technologically linked (today).  Other than limitations in
their ROMss and ROMs, there is no conceptual reason why a DX11 should have
to trade polyphonic for multi-timbral, for example. 

I may be way off, but if you use a pipe organ as a reference point, it is
(61 note x number of manuals) + (number of foot keys) polyphonic.  Now (I
am getting more wet here) there are n number of ranks on an organ which
can be combined in some number of combinations.  A given combination of
ranks produces an identifiable instrument.  What we would call
multi-timbral is, I believe, the combined (union, inclusive or, or
whaever) ranks for each instrument.  (There may be some kind of weighting
factor, baffle etc,. for  combining different ranks, but the idea is the
same).  If you start pulling out knobs (or equivalent), you will at some
point have all ranks engaged.  Is that the multi-timbral level of the pipe 
organ?  Or is it the total number of different knobs (each of which could
theoretically be pulled out at once)?    (I'll claim it's the latter.)

In a pure sense, multi-timbral is the number of identifiable instruments 
that can be simultaneously associated with a single key (i.e., a single
key press yields how many instruments at the same time).  

Without splits, pressing more than one key should yield the same set of
instruments playing the other note/tone.  Stated in a different way, 
if you tape recorded one key pressed, and then rewound and overdubbed a
second key pressed, the recorded sound should be no different than if you
had pressed the two keys at the same time. 

With splits, the definition becomes harder.  I think you could make a case 
for calling the multi-timbral level to be the number of simultaneous
instruments supported by keys played simultaneously.  So if a right split
had piano and flute and the left split had piano and acordian, and you
could play left and right splits simultaneously, you have a 3-level
multi-timbral instrument.  Sound right?

The problem with the "n simultaneous sounds" is that it mixes the concept
of polyphonic and multi-timbral.  An autoharmonize function plays
simultaneous sounds and I'm sure that's not what anyone means!

Alex (I live in a glass house)

	OK, it's time to flip over all the cards, and reveal the TRUE
	definition of Multi-Timbral............



	It means.......


	Utilizing more than one type of wood (or tree). ;*)


	Ummmmmm......yeh...........sorry.

The definition is easy.  I don't care to hear all the rebuffs of this either
as they are all full of hot air ;-) ;-) ;-) ;-) :-) :-) :-)

Each unique sound able to be produced from the unit simultaneously adds
one to the "timbralness".  A split with flute high part and bass low part
is two voices, hence multitimbral.  A layer is in a pure sense itself not
multitimbralness ass it is a new thord timbre composed of two or more others,
but practically, it requires a multitimbral unit to produce these sorts of
layers.  Two ( or n) different sounds from two (or n) channels is also
multi-timbral.


for db:

Show me a multi-timbral keyboard (say n multitimbral) that can play a split
of two different timbres top and bottom and  ***still*** play n or (n-1)
other timbres or sounds according to your definition (the n-1 other sounds
plus the split = n) across channels and I will concede, otherwise we will
all have to admit that using such splits on multi-timbral boxes uses
up multi-timbrality and hence makes them multi-timbral because they can do such
splits ( 6 oranges minus 2 apples doesn't equaal 4 oranges ).


Chad

> Show me a multi-timbral keyboard (say n multitimbral) that can play a split
> of two different timbres top and bottom and  ***still*** play n or (n-1)
> other timbres or sounds according to your definition (the n-1 other sounds
> plus the split = n) across channels and I will concede, 
    
    I showed you one about a year or two ago: my SQ-80.
    
    > otherwise we will
> all have to admit that using such splits on multi-timbral boxes uses
> up multi-timbrality and hence makes them multi-timbral because they can do such
> splits ( 6 oranges minus 2 apples doesn't equaal 4 oranges ).
    
    I don't follow this logic at all (I think your mixing two usages of the
    term: the pure word, and the synthesizer feature), but fortunately,
    since you have effectively conceded, I don't have to.
    
    	db

I haven't conceded yet.  You just can't say, "I showed you a year ago.  
The SQ80 can do that" or whatever.  You need exhaustive proof :-)

First, what "multi-timbrality" does the SQ80 have?  I'll assume 8 for this
discussion (where 8 is my n).

So, on channel 1 I have a split with electric piano lower half and flute
upper half (playing or sounding).

Now, simultaneously
can you on this SQ80 have on channel 2 a marimba souding?
                                     3 a split with trumpet top and 
	                                 trombone bottom sounding?
                                     4 a metal-guitar sounding?
                                     5 a string-bass sounding?
                                     6 a string-section sounding?
	                             7 a split with vox-choir top and
                                         harmonica bottom sounding?
                                     8 a solo tuba sounding?

In other words, can you have on this SQ80 with n=8 multitimbrality

11 different sounds sounding?

If the honest answer to this is yes, then I'll concede (and I want 
to see it in person :-).  If not, then you'll have to concede.

>    > otherwise we will
>> all have to admit that using such splits on multi-timbral boxes uses
>> up multi-timbrality and hence makes them multi-timbral because they can do 
>> such
>> splits ( 6 oranges minus 2 apples doesn't equaal 4 oranges ).
>    
>    I don't follow this logic at all (I think your mixing two usages of the
>    term: the pure word, and the synthesizer feature), but fortunately,
>    since you have effectively conceded, I don't have to.
>    
>    	db
>

Simple (and I'm not mixing two usages at all).  Look at above SQ80 example.
If manufacturer XYZZY says box NX-80r is 8 multi-timbral and using a split
with two separate sounds (ex.  EP and flute) only uses one of those 
multi-timbral slots, then you are right, otherwise you are wrong.  I think
you'll find that using such splits reduces your multi-timbrality.  Therefore, as
these splits are subtracting from your multi-timbrality, they must be
multi-timmbral "objects" as 6 oranges (multi-timbral slots) minus 2 apples
(split timbres) doesn't equal 4 oranges.  ONly if split timbres are
oranges = multi-timbral objects or slots does it = 4.

Chad

RE: .62

I would have thought that in your example, you would have called the
multitimbrality 11, because at the same time there are 11 different
instruments that can be playing.  

Isn't MIDI channel a red herring?  Can't a non-MIDI instrument be
multi-timbral?

Alex

re: .63

Bingo!
That is my point exactly -- it should be 11.

The point was made to show that splits are a part of multi-timbrality.
And yes, a non-MIDI box can be multi-timbral.  Multi-timbral to me is a 
big generic definition.  It assumes nothing either way about channel, splits,
etc.  That is what those words are for.  To clarify it further.

Chad

> If the honest answer to this is yes, then I'll concede (and I want 
> to see it in person :-).  If not, then you'll have to concede.
    
    It can do what you said, but you can't have 11 things sounding
    simultaneously because the thing only has 8 voices.
    
> If manufacturer XYZZY says box NX-80r is 8 multi-timbral and using a split
> with two separate sounds (ex.  EP and flute) only uses one of those 
> multi-timbral slots...
    
    Chad, can you find me a piece of manufacturer literature that uses this
    quantitative measure of "8 multi-timbrality" or discusses
    "multi-timbral slots"?
    
    I think these are of your own invention.
    
    	db

    re: .63
    
    We've agreed that "multi-timbral" may not be the best word for
    MIDI synths that respond to several channels, however you will
    find that "multi-timbral" is what is used to describe
    that ability, and it is NOT used to describe the ability to do a 
    split or layer.
    
    That a non-MIDI multi-timbral synth can be imagined does not
    diminish what I have said.  I have only made claims about
    how the word is used.
    
    	db

> The point was made to show that splits are a part of multi-timbrality.
    
    Chad,
    
    Whether you think so or not, you are mixing contexts.
    
    I have always agreed that a split is an example of multi-timbrality. 
    What I have said is that manufacturers don't describe synths that do
    ONLY splits (or layers) as "mult-timbral synths".
    
    	db

>> If the honest answer to this is yes, then I'll concede (and I want 
>> to see it in person :-).  If not, then you'll have to concede.
>    
>    It can do what you said, but you can't have 11 things sounding
>    simultaneously because the thing only has 8 voices.
>

So then it can't do it.

    
>> If manufacturer XYZZY says box NX-80r is 8 multi-timbral and using a split
>> with two separate sounds (ex.  EP and flute) only uses one of those 
>> multi-timbral slots...
>    
>    Chad, can you find me a piece of manufacturer literature that uses this
>    quantitative measure of "8 multi-timbrality" or discusses
>    "multi-timbral slots"?
>    
>    I think these are of your own invention.
>    
>    	db
>

multi-timbral-slot -- most machines have a certain limit as to the distinct
number of different voices that can be played simultaneously.  I called
this n-multi-timbrality and each occurance of a unique sound/voice being used
subtracted from the pool of available to use capability of the box.  Example,

A box that is 8 multi-timbral  plays a piano.  That means up to 7 more
unique voices/sounds could be played at once.  Polyphony plays no role here.
The piano used up one of my "slots".  ex.: If while the piano is still sounding,
an acoustic bass plays, it uses a second slot and that leaves 6 left.  Another
occurance of that same piano while the first ist still sounding
wouldn't use up one of these slots. (but it would take away from the 
total polyphony). 

I don't care if it is in literature.  I am discussing concepts of what
it means when it says multi-timbral, not nit-picking the words marketing types
choose to describe things.

I never said those terms aren't of my own invention.  I am trying to discuss
something generically, to apply a name to a concept.  Except in the
case of "n multi-timbrality"

That is in everyone's literature in some form.  Often it is described as 
multi-mode or some such.  There isn't much support for your case in the
literature that multi-timbral means that it has to be across several
MIDI channels.  Every piece of literature I have in my office (from Kurzweil, 
Ensoniq, Korg, and Yamaha) talks about MULTI-MODE, not multi-timbralness.
Most of the literature didn't even mention the word multi-timbral.  In fact,
one on the Korg P3/SYMPHONY modules about MULTI-MODE and then talks about
splits and layers.


db, if it can sound two or more different instruments, timbres, whatever you 
want to call them at once, even if only in a split, it has multi-timbral
capabilities. You have to read further (the fine print) to see exactly what
the box can do.  You can't assume anything from the word multi-timbral except
that it should be able to play two different sounds/unique voices/patches
at once somehow.


Chad

I don't know why I got involved in this discussion as a rookie, but I did :-)

First, I think Chad's definition sounds good to me.  It's easy to specify
(i.e., is consistent) and makes sense with and without MIDI.  I even agree 
that it is better that a Casio toy with a split be called multi-timbral 
(at 2, or bi-timbral) than to try to explain it away with some hand
waving.

If Chad is right and multi-timbral is hardly used at all, maybe we can use 
the term in this notes file and know what we mean (for a change).

Then we might talk about MIDI-timbrality or something like that to reflect 
the limitations caused by channel assignments, polyphony, or whatever.  

Listen, if Roland can call a patch a tone, I say all bets are off. ~/~ :-}

Alex

    Howsabout splitting the definition of "multi-timbral" into two:
    "multi-timbral" and "poly-timbral" ?
    
    Any use ?
    
    Jim Burke
    
    (has someone proposed this before ?)

RE: < Note 1834.67 by DREGS::BLICKSTEIN "Conliberative" >

>    What I have said is that manufacturers don't describe synths that do
>    ONLY splits (or layers) as "mult-timbral synths".

	Except Yamaha.
    
	Dan

> I don't care if it is in literature.  I am discussing concepts of what
> it means when it says multi-timbral, not nit-picking the words marketing types
> choose to describe things.
    
    What you are NOT discussing, is the ONLY thing I HAVE been discussing
    and I have stated that at least a half a dozen times now.
    
    If we are not discussing the same thing (as I have said several times
    and you have denied), why then do you have this elaborate idea on what 
    we each have to concede?

    And speaking of concession:
    
>>> Show me a multi-timbral keyboard (say n multitimbral) that can play a split
>>> of two different timbres top and bottom and  ***still*** play n or (n-1)
>>>other timbres or sounds according to your definition (the n-1 other sounds
>>>plus the split = n) across channels and I will concede,
    
>>    It (the SQ-80) can do what you said, but you can't have 11 things sounding
>>    simultaneously because the thing only has 8 voices.

>So then it can't do it.
    
    No actually, it can do it for a different "n".  It just happened that
    the ESQ-1 did not have enough voices for the example you chose in which
    "n" was 8.
    
    However, no matter.
    
    The new Ensoniq VFX can do EXACTLY what you described (for n = 8)
    because it does the same thing as the ESQ-1 but has more voices.
    
    So you have to concede anyway.
    
    	db

I HAVE been discussing multi-timbrality (which is the note title).
The reason  for discussing this is so that we don't assume things when we
read in the literature about multi-timbrality. 

Basically when a piece of literature says "multi-timbral" you can't read 
anything into that more than "can produce more than one timbre/sound
simultaneously".  You have to read more to find out anything more.

The whole point of my discussion was to show that the idea of splits has
nothing to do with the word multi-timbral when reading (meaning that
units that are multi-timbral are such regardless of whether they
can do splits.  Splits themselves rely on multi-timbralit).   I showed that
in order to do a split, a unit uses multi-timbral slots when in a split
mode with different timbres on each half.  Therefore splits are multi-timbral
entities.  Therefore, keyboards that can do splits must be multi-timbral
(multi > 1), even if they cannot produce multiple timbres on more than
one channel simultaneously.  And I looked in several brands literature and
they don't support your notion and Dan Eaton found a Y* that said
multi-timbral for s only-splittable keyboard.  (and I found a Korg that
discussed something similar).

Chad


And yes, I invented a few terms here to describe concepts.

About the VFX...  PLease explain in exhaustive detail.

Merely saying, ha ha :-), the VFX can do it isn't worth beans.

Chad

> About the VFX...  PLease explain in exhaustive detail.

> Merely saying, ha ha :-), the VFX can do it isn't worth beans.
    
    This answer speaks for itself.
    
    	db

    > I am discussing concepts of what it means when it says multi-timbral, 
    > not nit-picking the words marketing types choose to describe things.
    
    Perhaps you can explain what the difference is between "what it means
    when it says multi-timbral" and "the words marketing types
    choose to describe things"?
    
    	db

    
    
    In the last band in which I played we had a Farfisa (Remember the
    forks). It had 2 consoles and a set of detachable bass pedals. 1973
    no midi !. The top console could play one "timbre". The Lower console
    could play another. And with the bass pedals attached we could get
    it to break wind.
    
    Multi timbral no?
    
    					Paul.

	re    a few back

	Does anyone happen to know if  "Timbre"  does have any origin 
in  "timber".  Its conceivable at least, different tones from 
different woods, etc.


	re  the rest of this arguement

	It seems to be seasonal, maybe bi-annual or bi-ennual, but 
futile at best.  We could as easily be argueing about the term  
"hot water heater",  and BTW, I think we SHOULD !

	R

    re: .78
    
    > The top console could play one "timbre".  The lower console could
    > play another.
    
    > Multi-timbral no?
    
    The relevant question is: did any of the literature that came with
    it describe it as "multi-timbral"?
    
    	db

    
    
    re -1
    
    No *I* did !
    
    					Paul. 

The difference between the concepts in general use and the nit-picking
of literature is like the difference between engineer lingo and sales
lingo describing the same piece of equipment.  I read literature
(and I've read a lot) and multi-timbral to me in the literature never
meant multi-channel.

regarding VFX

My literature on it seems to indicate that it is 21 polyphonic and also
21 multi-timbral (it is not clear though).  Therefore, as multi-timbrality is 
only dependent on polyphony it doesn't satisfy my condition.

I've said enough.  db can read the literature how he wants :-)

CHad

>Show me a multi-timbral keyboard (say n multitimbral) that can play a split
>of two different timbres top and bottom and  ***still*** play n or (n-1)
>other timbres or sounds according to your definition (the n-1 other sounds
>plus the split = n) across channels and I will concede, 
    
    The Ensoniq ESQ-1 can play a split of two different timbres and
    sitll play play across up to 8 channels.   
    
    > otherwise we will
>all have to admit that using such splits on multi-timbral boxes uses
>up multi-timbrality and hence makes them multi-timbral because they can do such
>splits ( 6 oranges minus 2 apples doesn't equaal 4 oranges ).
    
    Not that it matters but this is a tautology.  You define
    multi-timbralness to include splits, and then say that if I have
    to admit that splits are multi-timbral because using them takes
    away from the multi-timbralness ("your" definition of it, of course).
    
    In other words, what you are saying is that "n" is the maximum number
    of distinct sounds that can be generated via various mechanisms
    including splits.   Since using a split (obviously) takes away
    from a definition of multi-timbral that includes splits, you say
    that I have to concede that multi-timbralness includes splits.
    
    Obviously a proof of a theorem that requires presupposition of the
    theorem is hardly convincing.
    
    	db

    re: .71 (Dan Eaton)
    
>>    What I have said is that manufacturers don't describe synths that do
>>    ONLY splits (or layers) as "mult-timbral synths".

>	Except Yamaha.
    
    I just finished talking to a Yamaha rep at the event that was held
    at the Nashua Shertaon Tara.
    
    According to him "Yamaha wouldn't call a synth that only did splits
    or layers 'multi-timbral'".
    
    When I him about the exception you have noted, he said that back in
    the days when the term was sorta new and the meaning in the context
    of synthesizers was not well estasblished, they might have used that
    term in conjunction with splits and layers but they wouldn't do it
    NOW!
    
    A Kawai rep agreed and noted that their literature was consistent
    with my interpretation (not splits and layers).
    
    I have been given no reason to doubt that TODAY the general usage of the
    term multi-timbral does NOT include splits and layers.
    
    	db

    timbre is from old french and means small bell.
    Tom

    re .85:
    
    In that case, my cat's multitimbral.
    
    Richard.
    
    Oh -- sorry -- small *bell*.

RE: .84

>    According to him "Yamaha wouldn't call a synth that only did splits
>    or layers 'multi-timbral'".

Okay, db, how about a rigorous definition of layer?  I can understand the
objection to splits.  The only way I can interpret the rest of the
discussion is to define multi-timberal in terms of MIDI channels and
to define layering as "the number of simultaneous instruments (patches�)
assignable to keys within a given split region, independent of polyphony."  
Correct?

Alex

�not in the Roland sense, Len

> Okay, db, how about a rigorous definition of layer?
    
    That area is even more confused than multi-timbral.
    
    Whereas I feel too much is being called "multi-timbral" in this
    notesfile, I feel that not enough is called "layer" in the product
    literature (Roland's "partials" immediately come to mind.)
    
    I wouldn't touch that with a 10 foot pole.  Besides you have
    lots of other nomenclature you have to worry about before you
    even START talking about "layers" such as the things that you
    "layer" (partial, voice, patch, sound, tone, etc.)
    
    	db

    Split:==      <--- pinano    bazooki ---->
    
    Layer:==        ^  pinano    bazooki  |
                    |                     V
    
    Edd

    This is the first time I've responded to any of the incredible volume
    of "wot's multitimbrality" replies.  I'm not interested in creating
    a definition that will fit all synth architectures forever and ever..
    
    to me, a timbre=voice=(digital)oscillator . I have 8 voices/oscillators
    available in the EMAX.  I have 24 available in the 1000PX.  Both
    machines will stack or layer voices (2 per note on Emax, 4 per note
    on 1000PX).  As long as I'm aware of the voice per 'patch' allocations
    I'm using, I don't get surprised.  Both machines will dynamically 
    allocate output oscillators based on demand, up to the voice limits 
    of the SGU.  A split keyboard setup is merely another way of demanding 
    voices.
    
    I can't believe so much space is spent on this issue !
    
    karl

    The problem with binding timbres to oscillators is that many timbres
    require multiple oscillators shaped by different amplitude envelopes,
    for example, one providing the attack sound and another the sustain
    sound.  I suppose you could call this a "layer", but I think most
    people have more "complete" component sounds in mind when they talk
    about layering.
    
    I'll buy the voice to timbre binding.  I usually think about this
    issue in terms of staves on a score.  Each staff represents a
    particular instrument or timbre.  Several staves may specify the
    same notes ("layering").  Each staff/instrument/timbre/voice has its
    own required degree of polyphony.  The number of staves corresponds
    to the maximum amount of multitimbrality required by the score.
    
    len.
    

    How does an orchestra size up?
Instrument #	#voices Instrument
	1    	2 	Flute I, Flute II/piccolo
	2    	2	Oboe I, Oboe II/English Horn?
	3    	2	Clarinet I, Clarinet II /Bass Clarinet etc.
	4    	2	Bassoon I, Bassoon II/ContraBassoon
	5    	2	Trumpet I,II
	6    	2	Trombone I,II
	7    	1	Tuba
	8    	4	Horn I,II,III,IV
	9	8	Chorus, SATB
    	10 	16	Piano
	11 	4 	Violins I (ca. 18) Violins II (ca. 16)
	12 	2	Violas (ca. 14)
	13 	2 	Violin Celli (ca. 12)
	14 	1 	Bassi (ca. 10)

    13 instruments colors, up to a total of 30 voices at once plus piano.
    (some pieces demand much more, say 8 horns, strings split up the
    kazoo, etc.)

    How would you describe this?
    Tom    

    < Note 1834.94 by ANT::JANZEN "cf. ANT::CIRCUITS,ANT::UWAVES" >
>    How does an orchestra size up?
>    13 instruments colors, up to a total of 30 voices at once plus piano.
>    (some pieces demand much more, say 8 horns, strings split up the
>    kazoo, etc.)
>    How would you describe this?

    I'd say a Kurzweil 1000PX (24 voice) for grandpiano and one or two string
    sections,  plus a 1000SX (20 voice) for the rest of the strings,
    plus a 1000HX for all the horns you can eat, oughta do it.
    
    Or a Mirage and an 8-track.
    
    karl

Just to add to the confusion.

From page 78, lower left corner, AUGUST 1989 KEYBOARD from an article
entitled: "How to Buy a Synth"

"In the old days, a typical synthesizer would only make one type of sound at a 
time.  A few offered keyboard splits or layers, with which one instrument
could make two separate sounds.  These days, many instruments--especially those
with built-in sequencers--are *multitimbral*.  Some instruments aren't 
multitimbral, though.  When in doubt, ask, "Is it multitimbral?

"A multitimbral instrument is one that can generate several types of sounds at 
the same time--for example, a bass sound in the low register, a trumpet
sound in the high register, piano chords in the middle, and also a drum
pattern. (In this case, at least two of the sounds will normally be played
by some sort of built-in sequencer.  Some multitimbral instruments allow
more than two sounds to be active on the keyboard at once, but some don't.)
Multitimbral operation is very useful for getting more independent lines
of music out of one synth.

"All synths these days are *polyphonic*, which means that you can play chords
on the keyboard and all of the notes will sound at once.  Don't try playing
cluster-chords with your forearm though:  All synths have a fixed upper
limit on how many notes will play at once--usually eight or 16. (Sometimes the 
upper limit for an instrument will change depending on how it is set up,
but the limit is always there.)  The fact that an instrument is multitimbral
*does not* increase its polyphonicity.  The nuilt-in sequencer may let you 
record eight voices of strings, eight voices of piano, and eight voices of 
sirens and whistles in the same bar, but if the synth is only eight voice 
polyphonic, some of the sequenced notes will not be played back, or will be cut
off by other notes.



--------------

NOt a mention of MIDI but it implies in the first paragraph that splits and 
layers are a part of multitimbrality.

---

Chad

    Don't think that "adds" to the confusion at all.  There's room for
    debate but it mostly seems to support what I've been saying all
    along.
    
    	db