Conference napalm::commusic_v1

    
    I'd say that # of sounds available should be in there.
    
    -Ed (Who likes his SoundCanvas)
    
    

                      <<< Note 2746.1 by FORTSC::CHABAN >>>
>    I'd say that # of sounds available should be in there.
    
    No argument here.. but what's a 'sound' ?  The definition changes between 
    samplers, let alone synth mfgrs..  Is a sound one sample, transposed to 
    cover several keys ?  Or is a sound an entire keyboard's worth of 
    crossfaded samples, a 'preset' in E-Mu parlance ?  Is an LA 'partial'
    a sound ?  What if you have 3 attacks and 3 loop portions, all usable
    with each other, does that make 9 LA 'sounds' ?
    
    I'm not being smartassed here, I just don't know the answer.  The
    number of memory-available sounds certainly IS a worthwhile parameter,
    but in a disk-loaded sampler, you could have one sound fill all of
    memory, or you could trim and loop these little guys down to the last 
    byte, and get hundreds of tiny samples in memory at once.  So I let 
    the number of sounds go, as it's undefinable or as meaningless as 
    available memory is.
    
    It's not how big it is, it's how you use it.  ;-)
    
    karl

    
    Yeah, I see your point.  
    
    I dunno, I consider a single preset to be a "sound". Samplers are, to me,
    a different beast altogether.  Most of the SGUs you listed are not really
    samplers because they don't record samples.
    
    To me, what makes a synth good is the ability to get more sounds.  Most
    machines made in the past 5 years or so are great at playing presets but
    are miserable to program.
    
    I'm going off into my Preset vs. programmable tirade again...
    
    -Ed
    

    Hmmm.  All I have now is a D70, 31 voice at $1700 -> $54.84/voice.
    However, when I had a TX81Z I had to use two voices to "pan", while
    with the D70 I can pan each voice individually.  And, with the TX81Z I
    had to double up voices to get FX, while they can be done with an
    internal processor.  So, if I were to compare with the TX81Z, I can do
    with the D70 with each voice something that would have required 4
    voices.  Thus, I effectively have 4*31=124 voices if I stack it up
    against that machine and want to do pan and delay FX.  Further, with
    the TX81Z, I could do a chorus effect on each voice, but it could
    consume all 8 voices for the effect.  With the D70 I can get panning
    (2X), delay (2X) and chorus (8X) on each voice.  So, in a way, my cost 
    per voice can be thought of as being around $1.71 ...
    
    Steve

    Don't think solely in terms of the technology, think in terms of how
    the technology is applied to make music.
     
    A sound is something that corresponds to a "part" or an "instrument"
    in the musical sense.  How it's generated, whether it shares "keyboard
    space" with another sound, etc., are not really relevant.
    
    So a multisampled instrument is a single sound - the multiple samples
    are necessary to cover the instrument's useful range without obvious
    transposition effects.  On the other hand,  multiple instruments with
    limited note ranges could partition the note number space, so these
    would be multiple "sounds".  A multitimbral "performance" with
    multiple channels assigned is multiple sounds; each can play its
    own part independent of the others.
     
    This kind of definition, which has real practical value, doesn't
    necessarily map directly to patches, presets, samples, partials, tones,
    etc..
    
    len.
    

	Is there still a high-end in the synthesizer market?
	What kind of synthesizer do you get for $20,000? $10,000? $100,000?
	Are Eventide and Karmon or something the high end in effects?
	What do they run?  Are they better than srv2000's and dep5's?
	How can we summarize criteria for effects boxes?  Number of
	really different effect types * average number of programmable
	parameters * (number of nybbles in a dsp word)/ price?
	Tom

    Choosing metrics that are meaningful to compare MIDI gear is probably
    impossible.  How much disk space did we waste in here talking about how
    many voices an MT32 *really* has?  Of course, simple "$ per voice"
    metrics overlook things like the *quality* of the voices, the voice
    architecture, etc.  
    
    Things really get sticky once you leave the land of rack mounts and
    start loking at keyboards and workstations.  My SQ-80 was $1900 new,
    so the $ per voice metric is awful...$237.50 per voice.  Of course, it
    has an eight track sequencer, a disk drive, a poly-aftertouch keyboard,
    what's that worth?  A D-50 aslo listed at $1900, also 8 voices.  Same $
    per voice, but no sequencing, no poly-aftertouch..wait, it has digital
    reverb....etc., etc.
    
    							Brian

    Maybe the thing to do is to rate each synth according to a standard,
    similar to how computers are rated (remember VUPS?).  One standard
    might be the DX7.  Establish parameters based on the DX7 for the
    numerator.  All synths can then be mapped against this standard.
    So a similar standard with drum machines (use the TR-707?), sequencers
    (use the MC-500?), and samplers (use the Mirage?).
    
    Steve

    maybe this IS just a silly exercise, given the number of architectures
    available, plus the impossibility of prorating an SGU that just happens
    to have a keyboard/sequencer attached.  Plus Tom's point is valid, once
    we start quanitfying SGU's, we really *ought* to do the same for FX,
    etc..
    
    karl

In a measure like this, I think the number of parameters you can change 
`simultaneously' (e.g., without menu hopping) would be helpful.  It would
help compare the old analog monsters (with all those knobs and dials ready
to turn) to the more modern, 2x24 LCD boxes.  Of course, I'm willing to give
extra points to the menu-driven boxes that allow simultaneous control of
multiple parameters using a SysEx box like the Lexicon MRC.

A factor based on this would help drive up the points for some of my relics
like the Korg PolySix I have.  Perhaps this is some kind of usability factor.
The reason I feel so strongly about it is that I remember being able to sweep
the filter with my thumb on my brother's Mini-Korg to keep an otherwise blah,
monophonic, analog drone interesting... with one hand!

... and if we work one up for SGU, FX, etc. we'll have to figure out how to
compare `master controllers' so we can compare the MIDIboard, KX88, and A-80.
Though, `feel' will always be a subjective (and usually overriding) factor.

Back to what passes for reality...

- John -

    >In a measure like this, I think the number of parameters you can change
    >`simultaneously' (e.g., without menu hopping) would be helpful.  It
    >would help compare the old analog monsters (with all those knobs and
    >dials ready to turn) to the more modern, 2x24 LCD boxes. 
    
	
    In that case, the Oberheim Xpander/Matrix-12 machines blow just about
    everything else away- the front panel has six "soft knobs", and you can
    add a bunch (12?) of MIDI parameters as well.  The only thing that
    might be slightly more flexible would be an old Moog modular.
    
    	-Bill

Awright, awwright- if you wanna do this *right,* here's how:

(1) List every "positive" factor you can think of, such as effects processing,
MIDI processing, patch/voice layering and/or routing, display quality, user
interface architecture, number of keys, keyboard feel/touch, and such. Assign a
weighting factor to each. This is your "goodness numerator."

(2) Do the same for negatives, such as total weight, frequent component failure,
lack of available service shops, bad manuals, and so on. This is your "goodness
denominator."

(3) Create a spreadsheet incarnating the resulting values, and divide the
resulting measure of "good stuff" by the price paid, giving you your "drool
index."

(4) Now all you have to do is fight about whose weighting factors are
"right," and you've got an absolute measure of MIDI goodness. Piece of cake,
right? (Ain't science fun?)

Bob

    re .12 - certainly seems a fair, logical, straightforward method of
    equipment evaluation and comparison to me..
    
    thanks, Bob.
    
    karl

    re .11 - I think Roland sets the standard for programmers.  Before
    Roland went off the digital deep end, Roland's synths offered optional
    outboard programmers that have a knob/slider/switch for every
    parameter.  Well, not quite, they could usually only be applied a
    "tone" at a time, so if you had a complex layered patch you could only
    diddle one layer of it.  The MPG-80 for the MKS-80 Super Jupiter, the
    PG-800 for the JX-10/MKS-70, the PG-1000 for the D-50/550 all have
    more controls than a 24 channel mixer.
    
    len.
    

   >>> ...the PG-1000 for the D-50/550 all have more controls than a 24 
   >>> channel mixer.

   Now, if only they had a heads-up display with virtual reality
   instrumentation response (with a dollop of hypertext on a useful manual).

   - John -