Conference napalm::commusic_v1

    ... not to mention stuff like:
    
                The first string to be hit on each strum would have
    a higher velocity value...
    
                We could be dealing with 100% gate time...
    
                Individual pitch bend values...
    
    AAARRRGGGHHH!!
    
    Edd

    uhhh... work on your keyboard technique and learn to close arpeggiate.
    I've got a real nice Flamenco technique, especially on a 'clavinet'
    voicing.
    
    km the uninvited

    
    	Another vote for keyboard technique.  I guess I don't understand
    why you don't just play it.
    
    	Next, you can try finger-picking.
    
    						,another uninvited Karl

    Well, I jist go ahead and DO IT, you know?  I had to solve the same
    problem for my covers of If I Fell and Pipeline that some of you
    have heard.  There were no guitars in these "arrangements", but
    there ARE some very effective strums.
    
    They're basically "just" 64th note arpeggios.  I picked my chord
    voicings from a guitar chord chart, and picked 4 or 6 string voicings
    as appropriate.  I programmed them in step mode.  No overdubs
    necessary, but ability to count to large numbers will prove highly
    useful, as will as the ability to remember where you are and what to
    push next.
    
    Now, here's where things get interesting.  1st, my sequencers only
    go to 32nd resolution in step mode.  2nd, 32nd note arpeggios starting
    on the beat sound "late" as simulated strums (i.e, 4 32nd notes
    take as much time as an 8th note, and the "centroid" of the strum
    is 1/16th note after the beat.  Solution to problem 1; playback
    at twice the desired tempo, with everything else programmed at twice
    the time value.  I.e., all 8ths become 1/4s, 1/16ths become 8ths,
    etc..  Thus 64ths become 32nds, which the sequencers can deal with.
    If the nominal tempo is 120, playback at 240.  Voila!  64th notes.
    
    Solution to problem 2 - start each "strum" a little early.  How
    early?  And how do this conveniently?  (Strums nominally in the
    downbeat end up crossing the bar line, making copies (desparately
    useful in saving time and sanity) tricky.)  Solution - the MSQ-100
    remembers bar lines - you must tell it explicitly where they are.
    The MSQ-700 assumes bar lines - they come every "n" (typically,
    n=4) beats.  So - step program the strums into the MSQ-100, in nominal
    position.  Then replace the first bar with one that's been shortened
    by 1/16 (or a 32nd) note (usually a silent bar for tempo count anyway).
    Transfer from the MSQ-100 to the MSQ-700, which shifts all the subsequent
    bar lines 1/16th (or 1/32nd) earlier.  Now all the strums (4 or 6 32nd
    or 64th notes) straddle the beat they nominally occur on, and don't
    sound "late".  Voila!
    
    Combine both techniques as necessary.  Up strums and down strums
    are dealt with by suitably arranging the notes in the chords.  I
    didn't bother with any velocity or other refinements.
    
    Choose a suitably "plucky" voice.  Adjust your release/decay times
    to whatever sounds good.
    
    Does it work?  Damn straight!  It sounds too regular when listened
    to directly, all by itself, but in the mix it works just fine. 
    Worth the effort?  Well, judicious use of copying of bars (BEFORE
    displacement) can save an enormous amount of it, especially if there
    are only a few chords (Pipeline had 4 if I recall correctly, If I
    Fell had 12) and a few basic strum rhythms.  A computer could automate
    the whole process - you'd just have to specify the chord and the
    strum rhythm.  It took me about 2 hours to program 3 minutes worth
    of strumming.  Most of that was finding and correcting mistakes.
    If I hadn't made any mistakes the first time I input the sequences,
    it would have taken more like 15 minutes.  You do get good at it
    eventually.
    
    Even 64ths sound a little too "open" for a realistic strum, but
    you do what you can.
    
    len the fanatical
     

    ... and I recall getting major flames for stating that 'sequencers
    are the refuge of those with crappy (keyboard) technique'... I seem
    to see some tacit agreement from a) another keyboardist b) mr.
    sequencer himself !
    
    Now, don't hit !
    
    karl 2

    re .5, .6 - Huh?
    
    Look, if you can't play keyboards you do the best you can.  I can
    do things with my sequencers that I couldn't do otherwise.  I'll
    be the first to admit I have no keyboard technique.  Do I make fun
    of you guys who can't play drums because you resort to drum machines?
    I don't see you guys making any effort to explain piano technique
    to those of us who don't have any.  How about being helpful instead
    of superior for a change?
    
    What's the point?  The question was how do you get a guitar strum
    effect if you can't play guitar?  What does keyboard technique have
    to do with this?  We had this argument about the merit of sequencers
    as tools (one man's tool is another man's crutch, I suppose) already.
    
    "Mr. Sequencer" indeed.  From "Mr. Piano"s I and II themselves no less!
    
    Some days this sort of stuff makes me just want to pack it up.
    
    It must be the rain.
    
    len.
    
    

    I CAN play the guitar. It's my synth that can't....
    
    Howzabout a test come August? We take someone with "keyboard chops"
    and give 'em a nice axe patch on the synth of their choice. Then
    we let them "strum" in real time and see if it can be done.
    
    Chops or not, I don't think it can be done WITHOUT a sequencer.
    It's just too damn fast...
    
    Edd

Re:.0, .8
    
    All whining aside, it CAN be done.  Unfortunately, it takes a lot of
    concentration on the part of the keyboardist, and a pretty decent
    understanding of what's going on on the guitar during a strum.
    
    For easy "strums" (i.e., an open guitar E or D chord), the best wy I've
    found to do it is to simply lock your wrists/fingers into the chord
    pattern and roll (very rapidly) together.  Of course, for reverse
    strums, simply reverse the roll.  A few times practicing the rolls and
    you'll get it down.  Make sure that you don't throw in a couple extra
    notes, or it won't sound like an axe anymore. One of the biggest
    mistakes most board players make when trying to emulate a strum is to
    throw in more notes to make the sound "fatter". Take my word for it, it
    just doesn't work. 

    WARNING - don't try to quantize these "strums", or you'll waste 'em.
    Most sequencers don't have the granularity needed to effectively
    perform quantization.  You'll need around (this is a WAG) 1/128 to
    perform this properly. 

    For more complex guitar work (such as simple finger picking), use your
    sustain pedal.  Pay special attention to which notes are played when on
    the guitar.  Then simply emulate that action on the board.  An
    interesting nit - eight voice units seem to be more realistic than do
    16 voice units - I guess that's because the voice/string ratio is
    closer (but I'm just guessing).  MAKE VERY SURE that your technique
    (note on placement) is very similar to that of the guitarist - it's of
    the utmost importance. 
    
    And finally, probably the single biggest mistake is to try to get that
    "hot lead" sound by trying to synthesize it.  Don't waste your time.
    Write (or find) a patch without a lot of modulation and a sharp attack.
    Then find a Fender Twin and fuzz the bee-jaybers out of it.  Voila! A
    lead guitar!  Incidentally, that's where most of Jan Hammer's hot leads
    come from - a Mimi-Moog thru a Sundown amp.  I use an Arp Axxe and a
    Carvin Amp w/EV driver for most of my "hot" lead sounds. 

    Now - isn't that better than a lot of whining?  Chet Atkins- eat
    your heart out ...  ;-}
    
The Dr.  8^)

    Whining, huh?
    
    bye.
    

    ... I got 2 NASTY guitar samples for the Mirage...
    
    A. "5ths" sounds like a wall of marshalls ala "you really got me"
       by the Kinks
    
    B. "Solo" Strat, complete with controllable feedback.
    
    I just cant do "Back on the chain-gang" by Chrisie and co....
    
    Edd

    
    	Hey Len, couldn't one think of a traditional musical score
    (plus orchestra) as a sequencer?  You write down what notes you
    want played when and the orchestra plays it back.
    
    	I have nothing against sequencers.  I think musicianship is
    way more important than performance ability.
    							- KmI

    re.7 :: listen, Leonard, do I have to pepper my replies with profuse
    smiley faces for you to know when I'm kidding ? 
    
    I haven't even looked at the intervening replies, I want you to
    know at least MY reply was firmly tongue in cheek. Guess I'll haveta
    learn how to make those ugly little symbols.
    
    Although there is one fallacy in your flame; that is, that none
    of us overeducated keyboard types can play drums. If there's a set
    available during my visit you can invite me to put up or shut up,
    okay ?
    
    karl moeller  
    

Re: .10 (Len)
    
    The "whining" remarks were in jest and not directed at anyone in
    particular - a "8^)" should have been inferred.  Sorry if I offended
    you.  I repent.  As for technique, that displayed on your tape
    submission leads me to believe that you're more than able to play in
    this manner.  Egad, you think your piano technique is bad?  You should
    hear me drum!
    
Re: .11
    
    I forgot - some of you rich people up north can afford things like
    samplers.  8-)  Since they only pay field people minimally, I'm still
    stuck with low-end stuff.  You should be able to use the techniques
    described to get some decent guitar-like effects on your Mirage,
    though.  Maybe someday when I'm rich, I can get rid of my old analog
    stuff and get a sampler ...  (now who's whining?)
    
the Pouter >8^{

    re -1... say, Brad, would you consider giving me 'little faces'
    lessons ?
    
    karl moeller

Re: .15
    
    You're not the first person who's asked me about this, so ... here's an
    edited copy of a note I posted in a local conference about 2 months
    ago.  Happy doodling ... 

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

    "Faces" as you see in the notesfiles (ahem - conferences) originated
    many, many moons ago on the USENET, which is another big user-group
    network.  DECfolks (as far as I can tell) picked up on this when the
    first versions of NOTES showed up. 


    The colon series:	:)  :-)  :^)	(most common face)
    			(:  (-:  (^:	(for Aussies)
    The eight series:	8)  8-)  8^)	(my preference)
    The B series:	B)  B-)  B^)	(he wears glasses)
    The % series:	%)  %-)  %^)	(a former boxer - face is munged)
    The & series:	&)  &-)  &^)	(a jovial little fellow)
    The = series:	=)  =-)  =^)	(rather narrow-eyed individual)
    The | series:	|)  |-)  |^)	(asleep or a squinter)


    Then there are various facial expressions.  For example,
    
	8^)		(grin)
	8^))		(horse laugh)
	8^]		(beaming)
	;-}		(wry grin)
	8-o		(surprise)
	8(		(bum out)
	>:^(		(angry)
	<|(		(pouter)
	8-{		(mustache or variation on pouter)
	8^)>		(beard - the van Dyke look)
	;^)		(the wink)
	D^(		(cyclops)
	8(=		(buck toothed varmit)

    Then there are hats.
    
	<8^)		(dunce cap)
    	*<8^)		(party hat)
    	-8)		(pin head)
	(8^)		(baldy)

    Of course, we can't forget noses ...
    
	8-)		(standard nose)
	8)		(no nose)
    	8^)		(for those like me, with large noses)
	8:)		(Dick Ames look)
	8o)		(Bozo)
	8`)  8')  8=)	(various others)
    
    This is getting ridiculous.  You get the idea.
    
8^)  8-)  >8-}  etc etc etc

    RE: 14
    
    Brad-
    
    Let me know the day that you decide to "get rid of your old analog
    stuff."
    
    I'd KILL for a Moog III or an Eu modular system.  Anyone want to
    get rid of a Buchla?
    
    MIDI be hanged.
    
    John M.

    OK, Ok, ok, my lack of sleep is showing.  Sorry for whining, whining
    about not whining, not inferring smiley faces, whatever.  It's not
    worth getting all worked up about.
    
    Actually, I know Karl has some facility with drums, as he's told
    me that before.  So do I, and we both use drum machines, so it must
    be OK.
    
    As long as we're off the subject, I'll explain why I haven't been
    getting any sleep - I just got a Mandelbrot set program for my Amiga,
    so I've been up 'til the wee hours exploring the complex plane between
    (-2, -2) and (+2, +2), finding all kinds of interesting
    fractal/recursive structures.  Now, if there were only some way
    to base a composition on it...  (Oh no, not *another* Rosette?!)
    
    And I have to admit that a fair amount of my ill-natured reaction
    was just plain envy at not being able to play keyboards even a tenth
    as well as Karl or Tom.
    
    Now, about this "MIDI be hanged" stuff...
    
    len the sheepish.
    

    Uh oh. A Mandelbrot set to be converted for composition writing...
      
    Think of what you could do with the "paradiddle" experiment Len!
    
    Todd. :-)
    

    'whine cooler' indeed.
    
    please refer me.. there's some unresolved referents in the previous
    two replies... 'another ROSETTE' and 'the paraddidle experiment'...
    
    sounds interesting. MUSIC V1 ?
    
    loretta

    The 'paradiddle experiment' I'm refering to is a song designed and
    recorded by Len.  He was nice enought to "sneak preview" it for
    those who attended the CZ101 demo.  It is based on a mathematical
    model in which all possible paradiddle permutations are derived
    from the set of his available electronic percussive voices from
    his three drum machines.  Am I right Len?  I believe the Rosette
    is the visual interpretation of the permutation model as it appears
    on paper.  Is this right too Len?
    
    Todd the mouthish.
    

    Ayuh.  "Rosette" and the "paradiddle experiment" (sounds like a
    grade B SF flick; or is it the name of the next big group -
    "Rosette and the Paradiddle Experiment"...?) are actually one and
    the same.  "Rosette" is an (interminable) 8 minute percussion piece
    on my COMMUSIC tape submission.  I don't know that I'd call it a
    "song".  Basically, the idea goes like this (I'll assume you know
    what paradiddles are):  define an adjacency relationship for
    paradiddles that differ in one stroke (e.g., lrllrlrr and lrlrrlrr
    are adjacent - they differ only in the 4th stroke).  Now consider
    all paradiddles that can be repeated indefinitely and not violate
    "paradiddleness" (i.e., no more than 2 ls or rs in sequence, ever).
    Call such paradiddles "perfect".  (E.g., lrllrlrr is perfect; llrllrrl
    is not.)  It turns out (if you're interested, I'll send you a copy
    of my "Ph.D. thesis" on paradiddles, which explains all this at
    rather more length) there are 40 perfect 8 stroke paradiddles. 
    Now work out all the adjacency relationships for these 40 paradiddles,
    and plot them as a graph (adjacent paradiddles are connected nodes
    on the graph).  The graph is an 8-fold symmetric four layer rosette,
    hence the name of the piece, which enumerates the paradiddles by
    "walking" this rosette.  This particular enumeration requires some
    duplication, so it takes 64 bars to hit all 40 paradiddles.  Rosette
    (the piece) then goes on to repeat the enumeration backwards and
    inverted (ls turned to rs and vice versa) and backwards-inverted.
    Mirabile dictu, the same sequences of paradiddles show up in the
    inversions and reverses.  A clever voice assignment algorithm is
    applied to all this, based on membership in rotation equivalence
    classes.
    
    It's actually much more interesting to talk about and look at than
    listen to, but it was the first fully finished piece of a series
    that I have been working on based on my "research" into
    paradiddle-space.
    
    You *did* ask.
    
    len.
     

    Oops, one little error.  There are actually 46 perfect 8 stroke
    paradiddles, but only 40 of them are strongly connected by the
    adjacency relationship.  The other 6 have no adjacencies (e.g.,
    lrlrlrlr is perfect but has no adjacencies).
    
    len.
    

Re: .23
    
    I was going to point that out, but I thought I would give you time
    to redeem yourself.  I see that you did.  Odz bodkinz ...
    
8^)	(the fibber)

    Thanks, Brad, I'm glad somebody's checking up on me.  It's lonely
    out here in paradiddle-space, and sometimes I lose the thread...
    
    ;^)  (how'd you know Fibonacci's involved?)
    
    (len's theorem - there are 2*F(n) paradiddles of length n, where
    F(n) is the n'th Fibonacci number; F(0) = 1, F(1) = 1, F(n) =
    F(n-1)+F(n-2) for n>1.)
    
    len.
    

    Tom, I tried that method, but the strings in my synthesizers are
    obviously real small, 'cuz I couldn't see them. I tried plucking
    the little black things with the gold legs but it didn't sound
    too much like a guitar.
    
    Thanks for your input though.
    
    Edd

Re: .26
    
    Tom - feel free to use these faces anywhere you like.  I hereby make
    them available to the public at large.
    
8) 8^) 8-) ;-) >8-} 

    ... and I play just what I feel
    
     
    Enough with these happy faces, back to the topic...
    
    I figured if Len did it his way I would too. I picked a suitably
    "strummy" song ("Deacon Blues", Steely Dan) and started.
    
    The method...
    
    Firstly, I learned not to trust the sheet music for the proper
    inversions for all instruments. So I wrote in the proper 3 note
    chords on the sheet. I set my QX21 to 32nd note resolution and played
    the lowest note (my inversion) of the first chord, then tie, tie,
    tie.... 15 times, as the chord lasted 2 beats. On 17/32 I sound
    the lowest note of the second chord and tie*15. Then measure 2 gets
    the same process. Onward till the first 4 measures have one note
    playing. Track it down and onward to the second note of the chord.
    
    Starting at measure 1, beat one, I enter a 32nd rest and then the
    next highest note followed by tie*15. By starting each note a 32nd
    after the one before it, you get the strum. By making them all the
    same length, you get the effect of a string vibrating till it's
    hit again.
    
    Continue on in this manner for the third and final note of the chord.
    Note 3:== 16th rest, note, tie*15, track down. More than 3 notes in a 
    chord started getting "slow" at 32nd N.R.
    
    After doing all this on MIDI ch 2, (where my DX lives) I repeat
    the whole process on 1 for the JX. Why repeat? Even though they
    are playing the same thing for the first 4 bars, on the 5th they
    get separate parts. Then back to 2 for the bass line. Put the DX
    in split mode, Fretless Bass patch on lower, Jazz Guitar on upper
    and JX gets the Electric Piano. Program the drums and VOILA!! It's
    2:15 AM, I've been here 6 hours and I've got 4 bars of Steely Dan.
    However, alot of the time was spent in R&D. I'm doing up the whole
    tune and it's sounding GREAT!!! Slow, but rewarding....
    
    Next we work in that nice Mirage sax sample...
    
    Edd
    
    P.S. Oh Leonard, can you program a drum machine to sound like Steve
         Gadd?
    
    
    

    Edd - I think you could save some time by manipulating your patches
    so you don't have to program the ties.  Just set your DECAY
    and RELEASE parameters for the patch in question so a single 32nd
    decays as slowly as you want the chords to "ring" for.  You may
    have to adjust this as a function of tempo.  If you set the RELEASE
    right a 32nd long NOTE ON/NOTE OFF pair will sound just like the
    same note sustained all bar long (which would be based on the DECAY
    time).  Note that this patch modification will get in the way if
    you DO sustain the notes (i.e., bar long NOTE ON/NOTE OFF pair),
    because when you finally release the notes they'll ring for another
    bar's worth.
    
    If for whatever reason you don't want to do this, you can still
    save time by only using 32nd note resolution at the beginning and
    end of the bar (I assume the QX21 allows you to change resolution
    in mid-bar..."barkeep, gimme a bourbon instead of this ginger-ale,
    I've changed my resolution...").  E.g., The first note of the strum
    can be done as 2 half notes; the second as a 1/32nd rest followed
    by a half note tied to a quarter tied to an 1/8th tied to a 1/16th
    tied to a 1/32nd;  and the third as a 1/16th rest followed by a half
    note tied to a quarter tied to an 1/8th tied to a 1/16th.  While
    this requires fewer keystrokes, it *does* require more concentration,
    a commodity often lacking as the "evening" wears on.  I'll be honest
    with you, I often find it easier to just keep hitting the tie button.
    
    I can't make your drum machine sound like Steve Gadd, but I can
    certainly program the parts he plays.  *Transcribing* the parts
    so I can program them is a bit more work.  When the millenium finally
    comes and I get some decent sequencing software for my Amiga,
    specifically allowing note by note control of dynamics (I no longer
    have a velocity sensitive keyboard, but my JX-10 should be here
    soon), or my MC500 finally shows up, we can get a little closer.
    The reason most drum machines sound mechanical is they don't have
    enough dynamic levels (e.g., my 707 has only 3 - no accent, small
    accent, big accent, even though over the MIDI input it will track
    the entire 0 - 127 velocity space).
    
    len.
    

    Correct me if I'm wrong, but isn't the line "shine up the battle
    apple" from the song, "Josie?"
    
    I never have understood that line.  Oh well.
    
    JM

    Yeah, it's from "Josie" alright. (Don't get anything confused between
    my personal_name and the content of my note. No relation.)
    
    "Battle Apple" is (I believe) an obscure reference to a Candy Apple
    colored automobile. Plus, it had to rhyme with "...scrapple".
    Obviously when Josie hits town, the boys like to put on their best
    show. "... she'll never say NO! No?"
    
    Edd
    

    Leonard, I found an even faster way....
    
    Assume 4 bars, strums on the downbeat...
    
    Lowest note (1) Set resolution to 1/2 note and play the 8 notes
    (2 per bar, four bars.)
    
    Next note.  32nd rest. Set resolution to 1/2 again and play 8 notes.
    
    3rd note.   16th rest.  "      "       "  "    "    "    "  "  " 
                                                                    
    Delete bar 5. Done.
    
    Edd

    Took me a few seconds to figure out what you were doing, but yes,
    that'll work, if your sequencer automatically inserts bar lines.
    My MSQ-100 doesn't, while the MSQ-700 does.
    
    What you've described in .33 will give you two strums per bar, on
    1 and 3 (you may still have the problem of the strums "feeling late"
    because of the spread over 3/32nds).
    
    One other thing I forgot to mention last time.  Again, this may
    be unique to the Roland sequencers, but...  The sequencer's
    articulation is a function of the selected resolution.  I.e., NOTE
    OFFs are sent just a bit before the "end" of the resoution unit,
    and just how far before seems to be proportional to the resolution
    unit.  E.g., suppose the sequencer sends the NOTE OFF after 0.98
    of the note's nominal time value has elapsed.  .98 of a half note
    is much larger than .98 of a 1/16th note, so step mode sequencing
    a bar as two tied half notes will result in a more "open" articulation
    than doing the same as 16 tied 16th notes, i.e., more absolute time
    between the NOTE OFF and the next NOTE ON.

    len.
    

    The QX21 refers to that function as "Gate-time". It has 2 settings;
    85% (default) or 100%. 100% gate time gets weird at times, like
    when playing the same note twice in a row. It gets a note on and
    a note off at the "same" time.
    
    So far, my arrangement of "Deacon Blues" isn't suffering from strum-
    lag, but I'm only using 3 note chords.
    
    Edd

    The 100% gate time may be there to give you the effect of legato
    playing.  The MSQ-100 can do this (you hold down the tie button
    while you key in the next note), and it actually overlaps the NOTE
    OFF of the previous note and the NOTE ON of the next note.  If the
    synth has a "retrigger" option for its envelopes, you can do real
    legato phrasing - e.g., take breaths on a flute line in the "right"
    places.
    
    len.