Conference napalm::commusic_v1

RE: drum kits

   Being somewhat of a drummer (ie, not too good), I try to program drum
   boxes kinda like I would like to play.  When I play a real kit, I
   often find myself not concentrating and hitting a rim mistakenly, so I
   try to throw a rimshot in here and there.  Diddling with dynamics can
   make a difference, too.  If you have the good fortune to have a
   tunable box or two snares (latter like the 707), then stick pretty
   much to one snare, but use the other once in a while (like you hit too
   close to the rim). 

RE: sequencing

   I find that when I play the piano, I will fade in and out of time -
   one section will be squeaky clean, while the next phrase will be a bit
   hapazard.  Playing pieces of phrases (as opposed to a whole piece) and
   (not) quantizing every 3rd piece help to make a sequence sound a bit
   less mechanical.  Of course, if you really screw up and try to use
   this "rule", it sounds pretty nasty. 

brad

    
    Not that I would call myself an experienced sequencist,  but.....
    
    When sequencing keyboards,  there are alot of tricks that  you  can
    use to make your music more aesthetic or human.  And what it really
    comes down to is my wishlist for features  in  sequencing  software
    
    o scaling,  this includes both velocity and rate scaling.  this can
      be done over any series of events,  and could be either linear or
      logrithmic
    
    o phrase envelopes,  meaning accents on certain beats in  a  phrase,
      as in the first beat in every measure.  This does not only  apply
      to drums,  nor does it necessarily mean  a  one  measure envelope.
    
    o grace introduction, the breaking apart or rolling of  chords,  or
      group of notes played on the same beat in a measure.
    
    o duration,  and use of the sustain pedal.  This  is  critical, and
      there is really no way to do it right.
    
    Zach
    

    I have a lot of "tricks" that I use when sequencing drum parts to
    make them sound more "realistic".  I plan to discuss them at length,
    when I get around to finishing my note in MUSIC on "Drums and
    Drumming...".  I can give a quick overview of some of them here.
    
    First, I think about how a drummer would actually play a figure,
    and especially the overall pulse or "groove" of a piece.
    
    For example, there's a tendency to accent the downbeat and the backbeat
    of a pop drum part.  So I'll program the hihat part accordingly.
    Using MC500 conventions (volume levels of 1 to 8, 8 being loudest),
    I might program a hihat ride as
    
    	beat	1 . 2 . 3 . 4 .
    
    	hihat	5 2 4 2 3 2 4 2
    
    This exaggerates a bit for the sake of example (besides, I don't
    want to give *all* my secrets away).
    
    Similarly, the bass drum tends to be accented more on the beat than
    off, so a fairly standard bass/snare/ride pattern might be programmed
    as
    
    	beat	1 . 2 . 3 . 4 .
    
    	ride	4 2 4 2 3 2 4 2
    	snare   . . 6 . . . 6 .
    	bass	6 . . 3 5 . . .
    
    Sometimes, an offbeat bass is a key part of the rhythm, so I'll
    program that accordingly:
    
    	beat	1 . 2 . 3 . 4 .
    
    	ride	4 2 4 5 3 2 4 2
    	snare	. . 6 . . . 6 .
    	bass	5 . . 6 . . . .
    
    Note that such accents tend to be distributed across multiple voices,
    reflecting the tendency of human drummers to accent the time slot
    rather than a particular voice.
    
    Fills are where the effects of proper dynamics make the most difference.
    E.g.,
    
    	beat	1...2...3...4...
    
    	ride	5.2.4.2.3.......
    	snare	....6.....436353
    	bass	6.....4.5.......
                                                  
    One way I often approach tha accenting of fills is to consider how
    I would "stick" the fill, i.e., which hand plays which note.
    
    There are three factors that influence such a decision:
    
    	left or right hand - right handed drummers naturally tend to
    	   strike harder with their right hands
           
    	first or second ("bounce" stroke) - the first stroke of a pair
    	   in the same hand, or a solitary stroke, will usually be stronger
    	   than the second of a pair
    
    	accented or not - is it on the downbeat or the backbeat or on
    	   some accented beat?  A deliberately accented note will be
    	   louder than an unaccented one 
    
    This makes 8 possible combinations:
    
    	hand	stroke	accent?
    
    	  R        1       Y
    	  R        1       N
    	  R        2       Y
          R        2       N
    	  L        1       Y
          L        1       N
          L        2       Y
          L        2       N
    
    These are arranged more or less in descending order of loudness.
    There are many possible ways of assigning dynamic values to these
    combinations.
    
    Brad's suggestions about varying the sounds themselves is a good
    idea in principle, except I've found that in practice the variations
    introduced are too large for a competent drummer.  E.g., the two
    snare sounds on my 707 are too different for this to work; the rim
    click added to the snare doesn't sound remotely like a rimshot,
    etc..  A more effective trick is to run the snare through a chorus
    with a slow modulation that's out of sync with the beat; this will
    subtly modify the sound from one hit to the next.  The key here
    is subtlety; you don't want the chorusing to be obvious.
              
    Etc.
    
    len.
    

RE: .3 (len, & chorus)

   It's on my Christmas list, Len.  Honest.  BTW - the 707/727 only
   respond to 5 different velocity levels.  Whut chu usin' to git 8? 

brad_the_non-drummer

    A couple of basic things that either haven't been mentioned or I
    didn't notice:
    
    1) Don't quantize
    
       When entering drum patterns, I set the interval to the max (96)
       and enter the riff in real time.  I may enter it at a slower
       tempo than I play it back, but I never enter sequenced stuff
       or drum patterns in step mode (I may 'edit' them that way though).
    
    2) Play the part - don't enter the part in step mode or with any
       kind of notation system.  No player in this world always plays
       a quarter note with the exact duration of a quarter.  No one
       hits all the notes in a chord simultaneously, etc.
    
    I don't use sequencing for much more than "recording a performance".
    
    	db

    The 707/727 mostly respond to all 128 MIDI velocities.  I've checked.
    You're thinking of the 505.  Also, the 8 levels are at the MC500, not
    the TRs.  The MC500 allows you to map the 8 rhythm part levels to
    any MIDI velocities.
            
    I say "mostly" because some of the 727 voices are clearly more coarsely
    resolved with respect to dynamics.  I routinely program 8 step
    crescendos into the MC500 and the 707 faithfully tracks them.  At
    least *I* hear 8 distinct volume levels.  I have also done regular
    tracks with full MIDI dynamics driving the TRs and they have a *lot*
    more than 5 discrete levels.
    
    len.
    
     

    re .5 - I routinely program everything in step time.  Sloppy playing
    sounds like sloppy playing.  With the exception of some fairly subtle
    timing considerations (see the excellent article in Electronic Musician
    last month or the month before), dynamics has *far* more of an effect
    on "robotic" playing than random or sloppy timing.
    
    How can you recommend playing to a metronome in one context and
    then suggest that precise timing is a liability?  I believe the
    kind of difference we're talking about here are a few MIDI clocks
    one way or the other, and my observations of real recorded MIDI
    tracks show that most nonprofessional human players are far sloppier 
    than that.  Enough so to warrant correcting their "playing".
    
    len.
    

RE: .6

   I don't *think* it is.  Maybe my ear is not as discriminating as
   yours.  I thought that I read this in the 707 manual.  Oh well, if you
   say it is, it must be.


brad_the_confused

    re .8 - I'm pretty sure you're thinking of the 505 which confesses
    to only having 5 discrete dynamics levels even via MIDI.  I haven't
    exactly memorized my 707 manual, but I'm pretty sure it doesn't
    say anything like that.  I'll check to be sure; if I'm wrong, I'll
    turn in my ears.
    
    len.
    

Edit bass patches to have the following envelope...

          / \
         /    \
        |       .
        |          .

Hmmm... maybe better to describe. Use a fast, but not gated attack. 28 or
29 out of a possible 30 is cool. Use a moderately slow decay rate. Do not
use any sustain level and have your release rate fairly quick. The decay rate
will often need adjustment depending on the tempo.

A sustained 'bass guitar' is a clear tip-off (to me) of poor sequencing
technique.

I do not use 100% gate-time on bass lines *except* in one instance, that
being when I am moving only one semi-tone down from a note of a shorter
duration than the one I am moving to. I've set most of my bass patches
to MONO and use a fast FINGERED PORTAMENTO to simulate sliding my finger
one fret towards the nut. Use fulltime portamento to sound like a sloppy
bassist...;^)

Use more velocity on the first note of a pair of shuffled 8th note
'triplets'....

Edd          

    
    re: .6, .9
    
      The 707 (mine at least) responds to all 256 accent levels. My
    Octapad records them fine. It's the damn keypad that's the limitation.
    
      As far as humanizing my sequences, I do the following:
    
    I try not to quantize - I prefer to record by hand, and I practice
    the stuff before I record it. You all are right. Sloppy playing
    sounds like sloppy playing.
    
    I use dynamics as much as possible - especially now that I have
    a 'board which understands key velocity. Also the Octapad.
    
    I don't do impossible drum parts. I'm sure you've heard the type
    - the hi-hat keeps going during the incredible drum roll and the
    snare still hits giving the impression that the drummer has four
    hands.
    
    Less is more - If I fill up an arrangement, I don't continue recording
    tracks. If it's too busy it starts going robotic. That is, only
    if I'm trying to NOT sound robotic. 
    
    /paul

    Gee, I find myself uncomfortable with the notion of arguing with
    "the guy who wrote the notes book on drumming".  However, I am rather 
    confident (i.e. "opinionated") in this particular view of mine.

    IMO, good drumming does NOT imply playing things right on the beat.
    Part of the "feel" is when you (and hopefully the rest of the band)
    play slightly ahead of or behind the beat.  (Isn't this the main
    motivation for the "swing" feature?)  This is far more true for
    accents (cymbals, certain snare usages, etc.) than for basic beat
    things (bass drum, hi-hat).

    When you enter things in step mode, there is a tendency to program
    patterns with resolutions of exactly 1/4, 1/8 and 1/16 notes.
    I think you lose the "feel" that a real drummer has when you do
    that.  
        
    To me, the "feel" can most directly attained by running the thing in real 
    time and hitting the keys according to the feel of the piece.  In
    the stuff I do, I like to have other instruments following rhythmic
    patterns suggested by the drums.  It's very hard to do this in
    step mode and get it right.  
    
>    How can you recommend playing to a metronome in one context and
>    then suggest that precise timing is a liability?
    
    That's easy.  The purpose of practicing to a metronome is not just
    to hit things at precisely the right time.  It's to keep a steady
    tempo over time, which is really what provides the solid time feel.

    Actually, there are many more reasons to practice to a metronome.
    This is but one.
    
    I don't think anyone could ever convince me that practicing to a
    metronome isn't the best method.  I've played with only a few drummers
    that I consider exceptionally solid, and every one of them practices
    with a metronome.  I'm sure some people can develop solid timing
    without it, but I'm willing to bet most can't.
    
    BTW, I practice guitar and keyboard to a metronome as well.  The
    major reason for that is to develope the ability to sync to an external
    time source which is important for playing in a band.  I also think
    this is both helpful and important from my own experience.
 
    	db

	When I programmed the snare for the 'Desert' piece for last Sunday,
I did it all in step time via MSQ-100.  The first thing I noticed was that
I didn't keep a consistent velocity (or even a musically useful one) when I
played it back.  Some of you sequencists with event editing of velocity can
take care of that after the fact.  I can't.

	I think in the future, when I program drums, I'm going to lay down the
kick in step time as a 'metronome' to keep the rest of the drums in line, and 
then sequence the rest in real time.  Maybe that'll take care of the step time 
glitches.

	Would that be a decent compromise between youse guys?

	Dan

    
    A couple of things that I do to kentize things a bit are.
                                                            
    Always get the bass drum and bass parts quantized.      
                                                            
    I record  the snare part with sticks and a pad.         
                                                            
    I record the hi-hat in real time and then overlay the dynamics in
    real time. (bot the last 2 are possible on the DDD-1 which I think
    is a little underestimated in light of the current HR-16 controversy)
                                                            
    The DDD also allows programming of the DECAY part of a drum hit
    so Ivary this to emulate the opening and closing of the Hi_hat and
    the choking of cymbals.
    
    I don't quantize any of the other parts at all. But will sometimes
    move the first bass drum beat of a bar forward every now and again
    just to drive it on a bit.            
    
   Re metronomes. I always practice with a drum machine these days.
    			
    					Paul.	
                                                               
                                                            

Just to defend the poor little TR505 - it does respond to more than 5 velocity
levels via MIDI.  It just happens to have 5 software selectable levels of 
volume for each drum on playback (similar to the playback volume sliders on the
TR707).

Drum machine feel: There are two types of popular drum beat sounds, quantized
and unquantized.  If you want a techno pop rock tune, program a sequencer.
Quantization of dynamics and timing is part of that sound.  If you want to 
do a tune that has mucho feel and dynamics, enter the drums in real time.

Some day I'm gonna enter a submission to a COMMUSIC tape using the TR505
played in real time from pads, and put the whole discussion of "drum machines
sound mechanical" to bed forever.  Tell 'em, Won!

Todd.

    re: .14
    
    Actually, that's EXACTLY what I do also.
    
    Some drum parts are to keep strict time (usually the bass drum,
    closed hi-hat, snare) and while I program them in real time, they
    are quantized to the right interval.
    
    Most other drum parts, particular those that 'follow' the 'feel' of what
    the other instruments are doing, I do in real time with no
    quantization.  I'm thinking of getting a drum pad for this, although
    the HR-16 does have velocity sensitive buttons.
    
    It's reassuring to hear that the approach I use isn't quite as
    'renegade' as my original impression had it.
    
    	db

    Dan,
    
    If you can edit a drum pattern in step time, why can't you edit
    the velocity?
    
    What I do, is enter the part in real time, then if I don't like
    the dynamics of a particular note but I'm happy with its timing, 
    I find the note in step time and put in a new note with the right 
    velocity.
    
    I mean, yeah, you're not really 'editing' the velocity, but it's
    really the same thing, right?
    
    I'm misunderstanding something right?
    
    	db

RE < Note 1091.17 by DREGS::BLICKSTEIN "Dave" >
   
	No, I can't edit in step time, only enter in step time.

	Another pitfall of my setup is that neither of my drum boxes recieve
MIDI volume as a continuous controller.  So what I enter is what I get.

    Dan
    

    Can you add a note in step time?  That's all I need to do to accomplish
    this.
    
    I.E., I find that the cymbal crash is too loud.  I step to the time
    interval where it happens, and enter a softer (velocity, not MIDI
    volume) crash cymbal note.
    
    That effectively replaces the loud crash.
    
    My guess is that I'm just misunderstanding what you're trying to
    do.
    
    	db

RE < Note 1091.19 by DREGS::BLICKSTEIN "Dave" >

>    Can you add a note in step time?  

	No, my sequencer only allows the initial track to be entered via step
time.  After that, you can only overdub.

	Dan

    If you are sequencing everything and then mixing down to tape, try
    fiddling with running the individual outs of the drum machine to
    a digital delay (not to be confused with a DIGITAL delay).  You
    can mix the delays on the hi-hat and sound a little like stewart
    copeland, or, alternatively, have a total wet mix, and fiddle with
    the knobs (ever so slightly) on the delay WHILE MIXING DOWN.  If
    all your drums are on 1 track, you're limited, but if you can add
    a 3ms delay to your hi-hat, and an 8 ms delay to your snare, and
    so on, it can sound fairly realistic.
    
    When a drummer plays, especially jazz, he doesn't necessarily hit
    the instrument being played the same way twice.  He may ping on
    the cymbal, then hit with the shoulder of the stick, then play with
    the butt, then move to the outer edge of the cymbal, et al.
    Same deal with a snare.
    
    P.S.  Have a cool yule

    I certainly don't want folks to think that I'm manic about metronomic
    drum parts, but I find a lot of this discussion about humanizing
    drum parts in the time domain to be a lot of talking through hats.
    If you ever bothered to carefully examine drum parts played into
    sequencers by non-professional drummers (i.e., people who don't play *all
    the time*, *for a living*), you will find that their "time" wanders
    all over the map.  Now, I step time program *all* my drum parts
    *all* the time, and not one of the many many people (including
    drummers) who have listened to them has ever complained about their
    lack of feeling.  I invest the bulk of my efforts in dynamics. 
    There's a lesson here for me, but I'll admit perhaps it's not for
    you.  That is, time is important, but not in the way most people
    think, and no where near as important as dynamics.  Yeah, if your
    time is so bad it's obvious, then it's more imprtant than dynamics,
    but you know what I mean.  
    
    I occasionally wish I could move the backbeat up or back a MIDI clock
    or two, so I could hear what it sounded like, but I am a firm believer
    in "if it don't itch, don't scratch", and my sequenced drum parts just
    plain don't itch, ever.  Maybe 20 years of drumming have destroyed
    my sense of rhythm and you guys (generic guys, no wish to exclude women
    programmers here) have more refined time perception than I and my
    many friendly listeners...
    
    Dave's right, you don't want random variations, you want a consistent
    variation from drum voice to drum voice (e.g., the snare is *always*
    "just a little late", etc.).  You can get this by careful programming
    (one of the things I have asked Roland for in the next MC500 software
    update is a way to introduce just such changes into the rhythm track
    on a pattern by pattern basis - you can do it now, but it requires
    programming the drums as a regular synth track, a tedious process),
    or as someone else noted, by routing the voice's *audio* through
    a delay (I have also used a delayed sync track while multitracking).
    Again, I refer you to the recent Electronic Musician article.  I
    note, however, that the kinds of variation typically introduced
    by "real timing" the drums rather than "step timing" them are usually
    less consistent and less subtle than the kind suggested by this
    article.                                                       
    
    The following observation certainly doesn't apply to any contributor
    to this conference, but it's been my experience (and dealers', and
    manufacturers' reps) that the reason most people want real time
    input of drum patterns is not to get a more human feel but because
    their rhythmic literacy is so poor that they can't map what they
    hear into traditional rhythmic notation, from which it is just a
    small step to step time programming.  Of course, lacking any
    understanding of dynamics, pattern variation and structural variation
    (i.e., details), they program up these dreadfully wooden parts in step
    time, and then assume that real time input ("gee, if only I could
    just bash about like a *real* drummer") is the answer to all their
    problems.
    
    len.
     
    
    

                               GREAT NOTE .
    
    I'm abashed to say that I uh, go both ways..
    
    I sometimes *quantize* (never step time, and always when I'm alone)..
    
    and sometimes I merely "bash about like a *real* drummer".. I once
    owned a Zildjian cymbal.. do I qualify ?
    
    actually, my two Commusic *IV* contributions show my two approaches.
    One, in 7/8, was played real-time no quantization.. snare/bass on
    one track, hihat/crash on another.. to a click. The other, in 4/4,
    contains a clever 16-bar drum cycle, completely quantized. They're
    both appropriate to the feel of the pieces for which they were made.
    

    karl

RE < Note 1091.23 by SALSA::MOELLER "good credibility..stop laughing!" >

>    I sometimes *quantize* (never step time, and always when I'm alone)..

	Gee, I wonder - is quatization an attempt to 'computerize' that which 
is human?  Kinda the opposite of 'humanizing' something that is computerized?

	Dan

>	Gee, I wonder - is quatization an attempt to 'computerize' that which 
>is human?  Kinda the opposite of 'humanizing' something that is computerized?
>

I think "control" is important...and that one can gain control by
practicing, or by computerization (quantization) followed by humanization.

/Mitch

    
    re .23
    
    I for one have pretty much dismissed this note altogether.  I  guess
    for you drummer types,  it truly is a great note,  but there is more
    to music than percussion.
    
    IMBO (In my biased opinion) drums come last.   I usually find myself
    grinding out the melody and chords first.  The bass line comes  next
    and is followed by background or filler such as strings.  Drums  are
    then added to accent and highlight,  not to keep a beat.
    
    Zach

    re .26
    
    I don't see how this note is limited to drums.  Perhaps most of
    the responses so far have been about programming drum machines,
    but instead of "dismissing this note altogether", or minimizing the
    role of percussion, why not contribute something about "humanizing"
    other parts?  I don't think you'll find a single "drumer type" here
    who believes there's nothing more to music than drums.
            
    len.
    

RE: .27

    I do.  Nuke synths.  Nuke pianos.  Nuke ALL ACOUSTIC INSTRUMENTS THAT
    CAN'T BE PLAYED WITH STICKS! 

    Gee, this sounds almost as stupid as ...


brad_the_ridiculous

My dulcimer is acoustic and played with sticks.  (Well, hammers, but they
look like sticks from a distance.)

The discussion about the fundamental role of rhythm is in *performance*
when real musicians have to stay together.  So one instrument (could be
drums, or piano, or bass) sets the tempo and evryone else follows.  If
the whole thing is sequenced, this is not an issue.  Of course parts can
be composed in any order.

But given that it is sequenced, how to make it sound like it isn't?  I've
been playing a lot of Michael Praetorius stuff lately, and my wife comments
on how dull and monotonous it sounds.  I agree with Len: the key thing to
watch is dynamics.  These Praetorius pieces arrived with not a whole lot
of dynamic information, so it comes out sounding like a pipe organ played
by somebody with stiff wrists.  Too mechanical.

    The way you get stuff besides the drums to sound human is to actually
    PLAY THEM and not enter them in step mode or quantize them.
    
    Unfortuantely, this isn't an option with drum sounds for most of
    us.
    
    	db

I like to sequence both the drums and bass using the MAC and Performer. The
drummer is a KORG DDD-5. Every instrument is tuneable. The volume of each
note for each instrument can be step edited. The decay of each instrument
can be varied from 0-15. So tuning, decay, volume, and variety seem to be
the key factors for humanizing drum machines. You can also get some bizarre
effects like playing a chromatic scale on cowbells in two part harmony. There
is a demo pattern in the DDD-5 that does that. For variety, I add subtle
things like using a partially open hi-hat instead of keeping it closed all
the time. Changing the tuning on the toms can simulate roto-toms for certain
passages. A crash cymbal crescendo can be simulated by using the highest
resolution setting on the drum machine (96 ticks per quarter note) and step
mode programming in the 96 hits starting on the 4th beat of the measure. Start
out very soft for the first 48 hits, and then ramp up quickly to the loudest
volume, allowing the crash to ring and decay (no more hits) for about the
last 1/16th note of the measure.

For bass, I use the S50 Roland sampler. I sampled a friends bass to get a nice
mellow bass, then combined that with slap bass and picked bass samples from
Roland. I have the samples assigned to about 1 1/2 octaves each on the same
keyboard patch. This provides all of the electric bass sounds you need by
just calling up one patch. Now you can simulate plucked, thumb slap, and
picked bass styles. Roland does magic with their timbre and decay parameters,
and I made the mellow bass that I sampled have a long sustain and slow decay.
The velocity sensitivity and aftertouch, combined with pitch bend, allow for
some really funky bass lines. I can easily make most any useful bass guitar
sound using this setup, and so, therefore, can the sequencer.

Dave

OK, I've got a problem here about humanization. 

After a couple of weeks of diddling with my CZ, I decided to get down to
business and do some real music.  Being a keyboard clod, I chose to
step-enter the music.  On the CZ, that's easy, and the only way to be able
to edit reasonably (the real-time mode is a tape emulator, with no looping,
quantizing, bouncing or punch-outs allowed). 

So I break out the sheet music and enter it.  Great!!!  The CZ suports 
notes from whole to 1/32 , dotting, rests of the same durations, ties, 
repeats, first and second endings, etc.

I soon find out what it does *not* support.....slurs.  The portamento is 
either on or off.  Portamento "on" ties any consecutive notes of the same 
pitch, ugh.  Tieing two notes of unequal pitch for slurring is a no-no.  
The CZ ignores the pitch value after the tie.  OK, so no slurs, all right, 
I'll live with that.

But the feel!!!  The sequence plays correctly but sounds stiff.  And not 
just from lack of dynamics, but due to *phrasing*.  The music is Irish 
fiddle tunes and of course, fiddle tunes are notorious for their phrasing
being *very* important.  This phrasing, of course, is something that cannot 
be notated on sheet music.  After years of playing, musicians tend to 
"humanize" what is on the sheet automatically, while the poor sequencer 
cannot do that.  

Now I can step edit, so I might be able to do things like subtle dottings, 
i.e take the phrase of a quarter note plus eight note and convert it to 
quarter note tied to a 1/32 followed by a dotted sixteenth, etc. in order 
to more closely approximate the phrasing or (gag) play it in real time...

Now that we are starting to see more sequencer/music copyist programs 
available, this becomes an interesting problem, how *do* you *notate* 
phrasing subtleties to allow the sequencer to reproduce them??

Comments?????

    re .32 - yep, phrasing (and articulation) is(/are) important.  Step
    time sequencing is dangerous, because it encourages an oversimplified
    notion of what's going on.  For some styles it works fine, but for
    others it's a disaster.
    
    The primitive sequencers that are built into most keyboards (as
    opposed to dedicated or PC-based sequencers) do not allow the kind
    of editing that's necessary to take raw step time programming and
    breathe some expression into it.  You have be able to do at least
    two things:
    
    	1) change the relative position of notes, and
    
    	2) change the duration of notes.
    
    And not by nice cleanly quantized units, either.
    
    Phrasing marks and slurs often require a legato style, which you
    can simulate by programming the notes so the note offs overlap the
    subsequent note ons.  Your synth will have something to say about
    how this is interpreted.  Some synths will retrigger the envelopes
    whenever they see a note on; others will use the same envelope,
    just changing the pitch, if there's already a note playing.  Some
    even give you a choice, but you may only get to exercise it in
    unison or mono mode.  Portamento may not always be the right thing
    to do - many instruments don't give you that option (e.g., piano),
    though you can still phrase them nicely.  The synth's rules for
    application of portamento may also be mode and overlap dependent.  
    
    You can also bound/delimit phrases by editing note durations so
    there is a "pause for breathing" at the end of the phrase - the
    last note is a little shorter (relatively speaking) than the rest
    of the notes in the phrase.
    
    Moving notes off the pulse can also have a dramatic effect; not
    randomly, but consistently early or late or whatever's appropriate
    for the feel you want.  Late notes will give things a more laid
    back feel; early notes will lend an urgency to things.  But don't
    overdo it, or things will sound just plain early (rushed) or late.
    A sequence of nominally rhythmically uniform notes can be phrased
    by moving some of those notes closer together or further apart, as
    well as applying a little dynamic variation to them.
             
    I never bother notating any of this stuff;  I just keep editing
    it until it sounds right.
          
    len.
    

RE: .32, CZ sequencing...

>I soon find out what it does *not* support.....slurs.  The portamento is 
>either on or off.  Portamento "on" ties any consecutive notes of the same 
>pitch, ugh.  Tieing two notes of unequal pitch for slurring is a no-no.  
>The CZ ignores the pitch value after the tie.  OK, so no slurs, all right, 
>I'll live with that.

I've had to work around this same problem. Part of the work-around is that
you can turn portamento on and off throughout the sequence. However, that
didn't always seem to "take effect" when I wanted it to. But what I found was
that I could force a "new note" by doing a null patch change - IE, put in a
change to some random voice, followed immediately by a change back to the one
you're really using. Yech. Seems to work, though.

>Now I can step edit, so I might be able to do things like subtle dottings, 
>i.e take the phrase of a quarter note plus eight note and convert it to 
>quarter note tied to a 1/32 followed by a dotted sixteenth, etc. in order 
>to more closely approximate the phrasing...

I do exactly that sort of thing. I'm about ready to move up to a "real"
sequencer so I can enter things in real time, and still be able to repeat
them (not to mention copy, etc!).

Tom

    I spent a lot of time this weekend sequencing using an ESQ-1.
    
    I found that quantizing helps in certain parts, but definitely
    sounds stiff in others.
    
    The major harmonic instrument I used was a electric piano.  The
    ESQ-1 has this wonderful feature that whenever you quantize a track,
    you can easily compare the unquantized track with the quantized
    track by pushing a button.
    
    After recording each basic track I tried quantizing it and in all
    cases I felt that it produces something very stiff and mechanical.
    On the other hand, I found that quantizing the synthesized bass
    guitar track almost always improved things (kept it in sync with
    the drums).
    
    Regarding drums, I generally found that the timekeeping notes (kick,
    snare, HH) generally benefitted from quantizing, but the fills (toms,
    ride, crash, etc) made it sound mechanical, ESPECIALLY when they
    were 'following' a rhythm created by some other instrument.  Oddly
    enough, when they were quantized, it sounded like the FILLS were
    not in the correct time, rather than what they were following (which
    was actually ahead or behind the true beat).
    
    	db
    

    
    Re: .33
    
    I understand all that, Len, but what I want to know is if anyone
    hs ever attempted to seriously notate phrasing beyond the "stock"
    sheet music.
    
    For instance, I have a series of bass books by Carol Kaye where
    she extensively notates some funk bass lines, and reading them,
    they look incredibly complex.  What is going on is usually not that
    difficult to play, but it *is* difficult to notate such lines, where
    there is lots of syncopation, glisses, etc.     
    
    Since my sequencer only allows editing of *notated* things (i.e.
    I can't go shift timing, etc. as on some PC-based sequencers), the
    only ways I have of entering correctly phrased parts are to have
    them notated, or to play them in real time.
    
    BTW, I realized when I bought it that the CZ sequencer had its
    limitations and decided they were worth living with.  If in the
    future, my budget allows it, i fully intend to get a more advanced
    sequencer, so the answer is not a ssimple as "change your sequencer".
    
    

    The answer is NOT NECESSARILY "Ill get a 'real' sequencer".

    first: Portamento is not sluring. Sluring is 'not tonguing
    the notes under the slur sign'. This corresponds (mostly,er,
    more or less) to Not re-triggering the envelope(s) on the
    sound generating circuitry. 
    
    Even the most expensive sequencer (by itself) may not have
    provisions to "notate" (if you will), this action. You will
    probably need to send system exclusive packs to:
    
   	a. disable multiple triggering for this machine
    		[then send note on/offs for the slurred notes]
    	c. enable multiple triggering again

    The point is that this (specific) problem is one of the sound
    generating unit, not the sequencer. Note that some gen's will
    respond to [note x on, note y on, note z on, note x,y,z off]
    as a slur. THEN were talkin about the problem  of how to tell
    the sequencer THAT instead of on/off on/off on/off pairs. You
    see what I mean.
    
    Now 'slured' notes (usually over bar lines) to the same note
    (pitchwise) is easily accomodated because it is still only
    one attack and one end...alll youve done is 'stretch' the 
    duration of the note over a bar line. This I dont consider
    phrasing.
    
    ya know?
        

    
    Re: .38
    
    You are right about the use of slurs, particularly with wind
    instruments,I guess the correct term for strings is glissando???
    Anyway, portamento would be close enough for me  :-)  :-)  :-)
    Or buy a Poly 800  8^)  8^)  8^)
    
    And I agree, it's not necessarily the sequencer, it's a combination
    of the sequencer and the generator (which in my case are the same
    beast).
    
    I will admit that I had not really thought about this, because the
    phrasing is not really a problem when playing in real time (assuming
    one can play the part, even if not at full tempo) but only when
    doing step entry.
    
    If I had a sequencer that supported looping or bouncing or cutting
    and pasting of real-time sequences, I could enter the tune phrase by 
    phrase in real time and cut and paste to get what I want.  
    
    

    re .38 and .39 - Slurring, portamento and glissando are all different
    effects, and are notated differently.  A slur is a curved line over
    two notes, and is a phrasing mark (i.e., the notes should be played
    as part of the same phrase, whatever that means to you).  A portamento
    (and I think glissando; Tom, we need your notational expertise)
    is notated by a straight line between the bodies/heads of the notes
    (If you want a glissando, you ask for it explicitly by writing
    "glissando").  Portamento and glissando differ, as Brian clearly
    recognizes, in that the former is a continuous variation in pitch
    while the latter is quantized into discrete halfsteps.  Some
    instruments, e.g., piano, can *only* play glissandi, they ar enot
    capable of pure portamento.
    
    Thus a portamento effect on strings would be played by sliding the
    stopping finger along the string, while a slur would probably be
    played more like a hammer on without another bowing attack. 
    Any string players out there who can get this right for me?
    
    len.
     

    Well, the Whiteman orchestration of "Rhapsody in Blue" starts with
    a portamento and has a piano glissando at the end of the piano
    solo.  I have a score, so I'll try and remember to check tonite.
    
    A slur is what happens when you try and do a gliss when you're drunk.

    	db

RE < Note 1091.40 by DRUMS::FEHSKENS >

	When I used to play trombone, a gliss was notated by a squiggly line
from one note to another, not a straight line.  Often it would have the 
abbreviation 'gliss.' somewhere nearby.  Is this unique to trombone?  Also, 
because of the trombone's method of pitch change (not counting embouchure)
it is indeed capable of a 'potamento' type gliss.

	Dan

    On electronic instruments, you can do anything, so do anything.
    
    On a violin, portamento is across two strings; you hear a slide
    out of a note on one string and a slide into the next note on the
    new string, no slide in the middle.  A violin glissando is done
    by sliding the hand.  You can imitate doing a scale as you slide
    by bouncing the bow on the string, or using a heavy vibrato WHILE
    you slide your finger so as to stop briefly at intermmediate positions;
    neither is an accurate scale, but an illusion of one.
    i gotta go
    Tom

    my favorite reference says a gliss is a wavy line OR a straight
    line with the word "gliss." written there.
    whatever bakes your cookies. You're not tied to acoustic concepts.
    Tom

    I looked up portamento and glissando in a couple of music dictionaries
    last night, and there's not a whole lot of agreement about what
    these words mean specifically.  In general they seem to mean more
    or less the same thing.  If you read into things, I probably got
    the distinction between portamento and glissando backwards.  I couldn't
    find anything about notational conventions.  It seems that glissando
    applies to instruments with quantized pitch (e.g., keyboards, valved
    wind instruments excepting the trombone, harp, etc.) and portamento
    to continuously pitched instruments (fretless strings, trombones,
    etc.), but the effect is considered the same.
    
    As Tom says, whatever works for you.  Who needs authority?
    
    
    len.
    

One technique that I have come across lately has to do with tempo changes
throughout a song:

My experimentation seems to indicate that a realistic feel results when
"fast fingered" or "busy" parts are given slightly slower tempos and
"less busy" parts are given slightly faster tempos.

My theory says that while playing, human's will react to "faster" (note-wise)
parts by easing up on the overall tempo, while "emptier" (note-wise) parts will
invoke increased tempos.

Of course, I may be totally wrong...

/Mitch

    This is an old topic I forgot about but have developed some new
    ideas on.
    
    One of the things I've recently discovered that seems to improve things
    is to make a conscious effort to use the pitch bend wheel, even if
    only very subtley, as well as the modulation wheel.
    
    It may eat up sequencer memory, but that doesn't seem to be a problem
    on my ESQ-1 (now an SQ-80).
    
    I try to add pitch bend and modulation to every track that doesn't
    have a keyboard sound to it.  Sometimes it's nothing more than
    "falling off" the last note, or "bending up" to the first note,
    adding slight vibrato to a sustained note, etc.  I also try to
    over-"dramatize" the dynamics (essentailly velocity) as I find
    that when you mix in all the other parts, the dynamics of one part
    tends to get lost.
    
    Even these subtle touches seem to make it sound much more human
    and less "keyboardy".
    
    Having been raised as a classical pianist, this is all new to me
    and I sorta have to REMIND myself to do it because it's not natural.
    
    One technique I do on the ESQ-1 is to play the basic keyboard part
    straight, and then add the modulation and pitch bend on a separate
    track (I found a way of doing this).  So one track is all note-ons
    and offs, the other track is all modulation controllers.
    
    	db

    re .46 - yeah that tends to happen.  Also, the more insistent the
    beat (e.g., 4 snare hits to the bar is more insistent than just the
    2 and 4 backbeat) the faster it sounds.
    
    Also, songs that start fast tend to slow down towards the end unless
    the rhythm section really has their act together.
    
    len.
    

    There was this recording of some beethoven symphony, maybe the ninth,
    in which the conductor played the busy parts faster than the
    low-activity parts.
    Tom

    My gut feeling to my own playing is that I play the BZ parts
    faster than the dirges...
    
    Or maybe it just seems that way??
    
    I'm so confused...
    
    Edd

    Something that I'm doing now is to map the parts by track and use
    the QX to bring all velocities on a track to zero, then decide where
    pp, mf, ff, etc. will be and use the QX to add the crescendos, 
    decrescendos, and such.  It doesn't require much time to do this,
    really.  It turns out that if the notes are played irregularly (not
    just a stream of quarter notes or whatever) the crescendo is accurate
    but not mechanical.  And, if care is taken to not have the same
    dynamics at once it becomes more like a separate musician on each
    instrument rather than somebody mixing everybody up or down.  The
    final touch is to vary velocities on particular notes just enough
    to add or attenuate as a 'real' musician would do.
    
    Steve

As a drummer, I have noticed that subtle changes and/or anticipations in tempo
are done to complement the mood of a piece.  Thus, a drummer's job is to
process mood information and make the proper adjustments to tempo, dynamics,
and accents.  Seems to me (and I'm sure Len will agree) that small changes in
dynamics and accents are much more complementary than small changes in tempo.

Todd.

    re .52 - len agrees.  But every little bit helps.
    
    len.
    

    A couple of comments here:
    
    First I'm currently working on my CM VI submission, the first thing
    I will have recorded with my HR-16. My first impression is that
    the feel seems to be much more lifelike than anything I got out
    of my TR-707, this may be due to the machine or it may be due to
    the hours I spent studying (among other things) Len's classic on
    drumming from music.note....along with several hours listening to
    old Santanna records to get an idea how to use the latin percussion,
    when CM VI comes out you'll see why I chose Santanna for this study...
    
    Now I have a question for the musically literate (in other words
    not me). When a band ends a song sometimes the song will end on
    a diussonant chord, usually there is a slight timing hesitation
    before hitting the chord...is there a standard or fairly standard
    value to this hesitation...I tried numerous things and it never
    came out right for me....hints suggestions etc. would be helpful...
                                                         
    dbII

    Is this the root 7th that bluesy toons often slump into?
    
    Richard.

    re .54 - I think the time unit you're looking for is the "just a hair".
    As in "just a hair" late, or "just a hair" early.  This unit is
    not convertible into MIDI clock ticks.
    
    len.
    

    Yep -- arises from various members of the band realizing that the
    song is about to end and jumping up in the air. Rate of descent
    is never wholly uniform, so as each hits the stage he/she plays
    a chord/hits a drum/ plucks a note. Hence randomn chord/timing factor.
    
    Richard.

    Yeah I could never figure that out...
    
    When the guitar player jumps, does he start the final chord...
     
                 (A) At lift-off
                 (B) Apex
                 (C) Touchdown
    
    Do they teach orchestral leaping techniques at leading schools?
    
    Edd

    The jumping tactic is a good way to suss who's the boss in the band.
    At the Mandela concert at Wembley, Knopfler and Clapton shared the
    stage and I was laying odds whose jump would end the song. Sure
    nuff, end of that 'move these refrigerators' tune both giants leaped
    into the air, and to my dismay Clapton appeared to watch Knopfler
    as if to judge the height and thus not underjump him. In fact, slowhand
    couldn't resist hitting the floor first, and he let rip with a classic
    sustained distension of a distraught flattened 17th only to discover
    than Knoppo had tipped off the soundman and had Clapper's gittar
    turned down. 
    
    So all bets were off.
    
    Richard.

    Ok I think Len's right it's the just a hair late I'm looking for...sigh
    quantizing will never replace humans....
                                                                  
    Back when I was in practice with LA East I used to time the jump
    so I landed (on my feet) for the chord,m we were pretty consistant
    once....but now I have to just be ready to hit it no matter what
    my altitude....
                                            
    dbII

    
    One really cool thing about Sequencer Plus Mark III is that there
    are many different "randomization transforms" that you can do to
    humanize your sequences. There is a start time randomization transform
    that, at low values (1 to 5 % or so) really helps make things sound
    more human. Using the velocity, duration, etc. randomization 
    transforms are also very effective humanization techniques.
    
    Mike D

>    "randomization transforms" that you can do to humanize your sequences. 
>    that, at low values (1 to 5 % or so) really helps make things sound
>    more human. Using the velocity, duration, etc. randomization 
>    transforms are also very effective humanization techniques.

Does it really work?  Master Tracks Pro also has these features, but I have
not really experimented with them yet. I would tend to think that
"random" is very different the "human".

/Mitch

    Please post a pointer to "Len's Classic" in MUSIC.
    
    re: takeoff/apex/landing
    
    The Peter Townsend School definitely hits the last chord on the
    landing.  Of course the swinging arm is directing the band at that
    point, so that's what you have to watch.
    
    Now if you're Nils Lofgren, you could play the last verse while
    "aloft".
    
    						-Jeff

    My "Drums and Drumming for the Musically Literate Nondrummer" is
    note 103 in MUSIC V2.  There are 9 (of 11) "lessons" plus rambling
    commentary.  The remaining two lessons (on "Embellishing the Backbeat"
    and "Accents and Fills") are yet to be written.
    
    I have an online copy that I can mail to anyone interested if that's
    easier than accessing the archived MUSIC V2.
    
    len.
    

    I can HIGHLY recommend Len's classic!  I printed a copy off to read
    at home and it's really made the rounds among my friends, even a
    drummer.  Now, if I could just get it back...
    
    Greg  Who's_looking_forward_to_more_"lessons"

    
    My copy of Len's opus is labeled DMLN and sits in a DECnotebook
    right next to VAX C, LISP, and V5 PPL :-)  Handy to grab and flip
    into when someone comes in and askes a truly silly question...
    
    Seriously, it's great.  Len should get a medal for it.  Read it.
    
    	-Bill                                                
    
    P.S.  Who wants to do one for electronic synthesis?  :-) 
    

    re: drums & drumming by Len F.
    
    I ate it up.  I can hardly wait for "embellishing the back beat",
    etc.  My 505 sat up and begged, and now I'm hearing the ride beat
    in every song on the radio like I never did before.
    
    Len shoud *publish* it. 
    
    
    								-Jeff

    In encouraged him, even offered to program all the examples on my
    HR for inclusion in the marketed version...
    
    come on Len, everybody's waiting for the finale
    
    dbII

    Yeah yeah yeah. Why doesn't somebody else write such a treatise.
    I've got to get my COMMUSIC VI stuff together so I can get beat
    up.  I don't see any other stuff on how to play keyboards idiomatically
    or the sax or whatever.  You guys shouldn't complain unless you've
    done it yourself [;^)].
    
    Seriously, though, writing this stuff takes time, something I haven't
    had a whole lot of to spare lately.  I would like to finish the thing,
    and then maybe try to get it published, but it's easier said then done.

    And I'm perfectly capable of programming the examples myself.
        
    Still, thanks for the enthusiastic support.  It really does make
    it all worthwhile (but apparently not worthwhile enough for me to
    get off my duff and finish it, you're all saying, right?)
    
    len.
    

           <<< CITZEN::DUA1:[NOTES$LIBRARY]CLASSICAL_MUSIC.NOTE;1 >>>
                              -< Classical Music >-
================================================================================
Note 271.0  Now Mr. Bilson if you could play that again unmusically   No replies
MARVIN::SCOTT "BArry A. Scott"                       74 lines   8-MAR-1989 15:55
--------------------------------------------------------------------------------


	Last nights  Horizon  program,  (BBC  England)  was about
	computers   and  music.   They  where  looking  into  the
	question "Can computer make music now?".

	They looked  at a number of things including, how to make
	a  good violin, how to synthesis instruments, what does a
	human  contribute  to the playing of music, what is it to
	be  musical,  how  is  music  perceived and lastly how do
	emotions relate to music.

	Mr.  Bilson  was  featured helping out research into what
	it  is  to  be  musical.  He was sitting at a MIDI master
	keyboard  and  being  recorded by computer.  They had him
	play lot of style of music as well as he could.  Then had
	him  play  the  same pieces again without adding anything
	that  he considered to be musical, just play the scroe as
	written. An amazing difference in sound.

	They concluded  that  a  good performer did the following
	types   of   things.    Instead  of  playying  each  note
	individually  they  run  into  each other.  The timing to
	changed  from bar to bar, speeding up and slowing down to
	match  the  mood  of the piece.  And then melody is often
	played  out of sync with the accompaniment, which hilites
	the melody.

	A bit  later in the program they had Mr.  Hogwood talking
	about interpretation of music.  He talked about how whats
	a  good  interpretation  changes  with time.  Saying that
	there is no definitive performance, as what is liked will
	change with time.

	The later  parts  of  the  program on perception of music
	where  very interesting.  THey did a experiment where one
	simple  tune  is  played into the subjects left ear and a
	second  simple  tune is played into the right.  They then
	ask the subject to humm the tune they heard.  The suprise
	is  that its not either of the two simple tunes but a new
	third  tune that is made up of parts from both of the two
	simple tunes.

	Then they  cut to a full orchestra, had he second violins
	cross  the  stage to there 19th centry position.  Now the
	first  violins  on  their own played some tchikofsky (gee
	thats  a bad spelling).  Then the second violins on their
	own  play  there  part.  Not that interesting a melody so
	far.  Then they play together a new third melody appears,
	oh thats smart...

	The end  of  the  program  looked  at  research  done  in
	austrialia  into  the  relationship  of emotion to music.
	What  was  found  was  that  there  are  a  number of key
	patterns   to   peoples  emotional  response  to  certain
	situations,  they  came  up  with a number of catogories,
	which  I think where anger, love, hate, sex, fear, happy,
	sad,  grief.   To measure the response to an emotion they
	used  the  pressure  of  a figure on a button.  From this
	they got a shape for each emotion.

	Now these shapes match up with patterns in music.  So for
	example  the  funeral  march  in  the  Eroica matches the
	pattern for grief.

	There was  lots  of  other  interesting  things  in  the
	program  that  I  did  not  remember enough details of to
	write  about  here.   (I going to be taping the repeat if
	there is a deatil that interestes you)

	Also as  it was a BBC/WGBH production, you may be able to
	see it locally in New England on PBS.  (Or has it been on
	already?)

				BArry