MusicScienceGuy

A jammer layout "sound-byte"

The Coles-notes version of how the jammer works
Explained in Easier to Learn? and Building a better keyboard are the detailed (many would say pedantic) rationale behind the design of the Jammer’s seemingly weird key layout.
Should you ever be possessed of the need to explain it quickly, here’s the TV-news-sound-bite way to do it:

The problem

First, In music’s all-important, major scale the important notes are the white notes on a piano. But; when laid out in a line like that, they are hard and slow to play, as:

• the hands have to move around a lot and beginners have to look at the keyboard
  
• its really inconsistent
  
• the keys that sound worst when played together are right next to each other, so anyone not an expert can and will make mistakes

It’s an honest-to-goodness elitist, albeit inadvertent, conspiracy.

Stack the keys

 

So so does one fix it?

 

First, you move the keys around a bit.

If you stack them so they are closer together, and especially put  notes that usually are played together, in various combinations, close together.

Added bonus #1: often they can be played with one finger.

Bonus #2: this also makes the fingering consistent, so you need only learn one fingering, not twelve.

Second, as another bonus granted by the fickle laws of mathematics, the troublesome black keys move out of the way:

Bonus #4:

Jammer theory sound-bite

We can put an on extra set of keys on the other side to make them easy to reach. Thus automatically springs into existence a "Sharp" and a "flat" section, because this thing is built around music theory.

Voila! All three problems are solved. Plus you get quintuple bonus points because it matches the way we hear music, as I've described in  "Building a better keyboard".

That's the simple, powerful idea - just move the keys to where they are more reachable, useful and one makes far less mistakes.

And that's why I'm so keen on the jammer: it's simple enough for even yours truly to understand.

Shortcuts to faster instrument learning

Among the ways to learn a musical instrument faster are these basic strategies:

• Find or create an instrument that is easier to learn
• Understand what you are, at more than a shallow level, trying to learn
  
• Leverage existing skills you might already have

The first I’m addressing on this website with the evolving incarnations of the jammer. The second I’m slowly gnawing away at, with postings on useful music theory.

The third strategy has now popped up. James (Jim) Fallgatter (of www.Aqwertian.com), a very interesting website, just suggested we leverage our ability to type on a qwerty keyboard as a stepping stone to learning to play music.

This idea has appeal! After all, if one can type at 40 words per minute, that’s 240 characters per minute (5 char per word + 1 space), or 4 characters per second. If a character was instead a quarter-note, that’s well in excess of the two quarter-note per second speed (120 BPM) of the typical song. Beginners could even rip out songs with eighth notes.

This leads to the question: Why The Heck (this is a family-oriented site) can’t good typists play “Twinkle-twinkle-little Star” on a computer keyboard when said keyboard’s keys are set to play musical notes? The answer is maybe they can, that perhaps with a couple of twists, the gap can be bridged.

One can't do this to easily by mapping a PC keyboard to a piano-like note layout because a piano lay is just plain too long. The folded layout of a jammer keyboard is much denser in useful notes.

The simplest way to do this would be to assign a "letter value" to each jammer key, perhaps with a clear plastic overlay. The ZipEx jammer's keys as derived from a M-Audio keyboard are 2.5 cm (1 inch) wide, and a PC keyboard is about 1/3 less, so it's a bit more of a stretch, but the fingers are used to such variation.

Novel Learning strategy- Qwerty overlay of the keys

 
The assignment for the right hand is shown to the right:
Green = middle C. So to play the major scale just say:
  'play "KL;UIOP8" '  and the song "Doe-a-deer, a female deer" would be "KL;-K;K;K;". This is a bit weird, but a bit of experimentation on this site shows it works. Such a translation would be simple to program.

Novel learning strategy- Qwerty overlay membrane

Or, one could use a finer overlay, and have a letter on every musically important  key and key combination, as shown at left. If nothing else, this shows just how "dense" this layout is - within the reach of the fingers, without moving the hand significantly (less than a centimeter or so), are 90% of the notes and combinations you need at a given time to play a piece.

Thirdly, Jim has a more dynamic assignment system that I'm hoping to get him to explain and post here ;-)

Drawbacks: The "feel" and timing action of the PC key board is a lot different, perhaps too much so for this to work. With this in mind, I'll lighten the feel of my ZipEx jammer's keys.

Other than that and this post, I'll put this idea on the back burner.

Sigh so many ideas and so little time.

Music for Nerds

We’ve had “Music for Dummies”, but there’s a group (sub-species?) that needs to start off at an (ahem) slightly more elementary level. These individuals, whilst excelling in some simple areas of human endeavor such as computers, stereos, midi and Global Warming, need help in translating music into simple terms that that they can understand.
Yes, I'm talking about Nerds.

Being a nerd myself, blessed by marriage to a talented musician, whom I’ve tried very, very hard to understand, I’ll try to bridge the gap with a little series of articles for nerds that may stray in from time to time. Musicians, too, may find it interesting to peer into an alien mind.

Part I: what the heck are notes?
The “octave”
The sounds we hear can be broken down into about 7-9 “octaves”, where each octave is double the frequency of the previous one, starting at say 27.5 cycles/second, up to 14,080 or so. An octave in turn is split up into 12 notes, which are spaced out evenly into 12 proportional intervals. It is called an octave and not a dou-dectave as it obviously should be, because musicians can’t count (in school they were "numerically challenged").

Some notes are more equal than others
To non-musicians, most notes seem pretty much the same, except for sounding and bit higher or lower and perhaps needing expensive speakers to accurately reproduce the sweet clear sound of booms, crashes, gunshots and thuds on their simple digital home theaters.

To musicians there is a vast difference, depending on the “root note” or tonic, the special key note (the key for short) an arranger happens to favor to build a song around.

Some fundamental facts

1. Musicians, unlike nerds, are easily bored. This goes triple for directors, who’ve heard it all, and triple more so for arrangers, who’ve studied it all.  Arranger-directors can deduce an entire piece from the first dozen notes.
   
   
2. Starting from the root, another note played with it sounds good (consonant), if it is roughly 1/4, 1/3, 1/2, 2/3 or 2/1 higher, these notes resonate. Per musicians’ numeracy problems, these are called a third, fourth, fifth, sixth and octave intervals respectively. These are the main part of the major scale. The big gap between the third and the root is filled in with a second halfway between. Likewise the sixth and the octave are filled in with an seventh, almost as an afterthought.
   
   
3. Neighboring notes (semi-tone intervals) sound lousy (dissonant) because they make a funny wobble in the sound.
   
   
4. Notes tuned to exactly the square root of 2 above or below the root (a shade below a fifth interval) kind of cancel out the root note and sound weird. This is the dreaded tritone, despair of music directors everywhere, as it happens easily when jumping up and down fifths.
   Tritones are also the despair of singers everywhere, as an arranger may, on whim (see #1 above), sprinkle in a few cleverly hidden tritones for some added spice. (Hint: beware the diminished chord.) Tritones are hard to sing because (1) they are anti-resonant – there’s nothing to latch onto or match, and (2) if you spend 99.9 percent of the time avoiding something it's hard to hit it when a director decides "April Fool".
   
   
5. Notes spaced away from the root note by any interval approximating the 1/3 and 2/3 power of 2 (a shade below the sixth and third) partially cancel the root and sound interesting, but nothing you’d want a steady earful of. These and the notes just below the seventh and second are the minor intervals.
   
   
6. Minor intervals and the tritone were added later. Music terminology still reeks with distain for them; they don’t have their own names and are called accidentals. Arrangers love these guys (see #1 again) and mellow them by blending in other notes into chords.

So how does this relate to the musical staff and why is it so confusing?
Microsoft vs Apple;   Qwerty vs Dvorak;   Pianists vs everyone else
In the distant musical past – there was a rip-snorting battle for market share between piano and other musical arrangers. The pianists won. Pianos then were tuned around C, the fourth note of the octave, and only had white keys. The musical staff we use is basically an early piano keyboard turned sideways.
Lines and spaces on the staff correspond to the white keys of a piano tuned to the C major scale, and musical notation pivots around this weird fixation. Below is listed the notes, with the black keys/accidentals in bold print.

 

Note   Name    Interval from Root note    Also known as …    Singers name   Ratio   
+12    C       Octave                     Octave             Doh            2/1   
+11    B       Seventh                    Major Seventh      Ti             +7/8   
+10           (blue) Minor Seventh                                          +3/4
 +9    A       Sixth                      Major Sixth        La             +2/3    
 +8            Minor Sixth           
 +7    G       Fifth                      Perfect Fifth      So             +1/2   
 +6            Tritone                    arrgh              Heck
 +5    F       Fourth                     Perfect Fourth     fa             +1/3   
 +4    E       Third                      Major Third        mi             +1/4   
 +3            Minor Third                                                 ~+1/5   
 +2    D       Second                     Major Second       Rae            +1/8   
 +1            Minor Second             
  0    C       Root                       Boring as Hell     doh            +0   

 

Strange, isn’t it? They call it an “Octave”, but it only has seven “key” notes and it skips that cool minor seventh interval in favor of the dissonant “major” 7th. Go figure! Its hard to believe even early musicians were quite that numerically challenged.

If you go up the scale on the white keys (major & perfect notes), you are following the “Major Scale in the Key of C”. If you follow all 12 notes (semi-tones), that’s the “Chromatic scale”.

Anyway, the important thing is the pattern is simple: go up from the root note in intervals of 2 semi-tones (whole-tones), except for a semi-tone jog from the third to the perfect fourth and a correcting semi-tone jog from the seventh back to the octave. If any sane person (and even some musicians) were to design the staff notation today, it would look a lot better.

You would just put in a symbol marking where the root is, and always assume whole steps, except where some little little brackets would indicate that the whole series of 4 notes is dropped down (flatted) by a semitone and the root note indicated by a little carrot (naturally, Doc). This way, musicians would only need to learn to read two staffs, instead of 24 variations. (The same goes for the piano itself; if they had alternating white keys/black keys, people would only have to learn 2 fingerings, instead of 12! Further, one could … get a lot of blank stares from musicians, who are quite used to being notationally challenged.)
Don’t be too hard on them – do you use a Mac computer and a Dvorak keyboard?

Figuring out the tonic
Its nice to know where on that staff those little jogs are hiding – otherwise you can trip over them and be a semitone out – not a gentlemanly/ladylike thing to do; kind of like farting in public.

The normal musical staff notation assumes the notes correspond to the C-major scale, to get to the other scales, musicians shift the C major notes by the minimum number of semi-tone adjustments sharp or flat, (they don’t like mixtures), to get the major scale needed. In nerd terms, that little jog in the whole-tone spacing of notes is shifted up or down.  

For example, if one pastes on an F-sharp on the staff, that shifts that third-to-fourth jog up and gives us the right jog pattern for a G-major scale; the F-sharp becomes the seventh below the tonic. You can repeat this again, going up a perfect fifth each time.  Music theorists make a big deal of this, calling it the circle of fifths, but it’s just the way the math works. If you see sharps, just sharp the last one.

For flats, the trick works going by down in fifths. The important thing to remember here is that the second to last flat is the tonic, and when you see one flat, the tonic is F. (It’s simpler than counting up a fifth from the last flat.)

Caveats
I have hugely simplified a vast subject, thus violating the maxium:
  'Tis better to keep quiet and be thought a fool than to open one’s mouth and remove all doubt. -Anon.
Please be kind.

How to Build a Jammer

Update - January 2012 - Yes, I really did once upon a time build a Jammer by hand.

How I have better ways to do it.  See recent posts. Ken.

 

I've documented how useful a jammer is elsewhere . In short, it's like a super-improved piano, a 2.0. version.

Riddle: How do you make a 4-row hexagonal-key instrument from a 2-row piano-key one?

A bit more portable than a Piano?

Answer: You cut it in half and fold the keys back under!

Converting keyboard to Jammer

Amazingly Simple, ain't it?
Expand the picture to the left
for the details.

I’ve made two, the first, a light wear-able version:

and the second, a stand-top version:

A very playable jammer

The first was naturally harder to build. Design improved with the second; it can be made by anyone with basic carpentry skills (no electronics skills needed).

Update: I'm seriously investigating making modified keys and the colored key-caps (as seen in the photo to the right) made. I've created prototypes, next the question is the cost. If so, I can create a simple conversion kit. Drop me a call here if interested.

Here goes a basic summary (if you find yourself confused contact me, and I’ll happily elaborate).

Background (optional reading):

A typical electronic keyboard has 2 parts, one is the music control electronics, and the second the key holder assembly itself.

I have examined just 3 keyboards, A Yamaha DjX, a Technics P50 and a M-Audio 88es, but I suspect most are pretty similar.

The key board electronics is on two loooong printed (Image)  circuit boards (PCB) clamped into the key assembly. The two boards are connected to each other by a single 6-inch, 12-lead wire, and one of the boards in turn connected by a 10 inch or so, 6-lead connector to the controller. The controller in turn has connectors running to switches mounted on the case.

Incidental details on how the keyboard works: each long PCB has clamped onto it a bunch of rubber buttons which make a little dome about 2-3 mm over the PCB. These buttons are depressed when the plastic keys of the keyboard assembly are pushed down (I'm not sure what makes them happy).

The underside of the little dome buttons, the part clamped to the PCB, has a little bit of conductive black graphite paint on it, and when it hits the PCB, it closes a circuit, and the circuit generates a signal that the controller interprets.

On the Technics and M-Audio keyboard, each key has two buttons it depresses, presumably at slightly different times to denote a key's velocity.  The Yamaha has a double contact set within each button and the contacts are of slightly different length, thus denoting the key's velocity.

The important thing here is that you don't need to know further detail on how it works, or electronics. Your can mod an existing keyboard slightly.

In my first design, I used just the PC boards and the rubber buttons, and rebuilt the whole key holder assembly. For the second one I borrowed a simple idea from a nice guy in Chile and added to it.

List of materials

• A keyboard to cannibalize - I recommend the M-Audio Keystation 88es (now $300US), but Keystation 61es ($200) might do the trick. This is an excellent, rugged keyboard, without flashy buttons that are useless to our purposes. Critically, it is also USB powered – it can work with only a laptop. 
  

  M-Audio Keystation 88es
• A laptop that can run Max/Msp or some other midi remapping software and the instrument samples you desire. If you are buying one, check that it has enough volume. 
  If your laptop is quiet, a small pair of battery-powered “iPod” speakers works well.  
• Some wood for the box to contain it. I used 8 feet of 3” hemlock coving.
• Clear plastic or plywood sheet about 3 mm thick for the base –
  I suggest a piece from a vinyl window. One should be able to get this from a glass window shop.
• 133 keycaps  – allow extra 10 for wastage - a set of 1” or 1 1/16 plastic disks.
• double–sided tape. to attach the keys caps were attached to the key-bases.
• 2 spray cans of black paint, or one black and one varathane varnish for a more exotic pearly effect
• Model cement to stick wood to plastic (Weldon #16 worked well)
• Wood glue

Tools

• Drill press with depth control
• Drill bits in 1/64" gradations
• Work shop Disk Sander is handy – these are $50-70 and this is a good time to get one.
• A fine-tipped felt pen to mark up the plastic or wood base.
• Table Saw to make the keys (a hand saw may do)

Software

• Needed is some way to remap the keys to different notes. I used MAX/MAP, an excellent but expensive program @ $500 (I have lots of uses in mind for it, so felt it's price justified). One might be able to use Java, or in a pinch, I will compile my Max version and send you a runtime exe.   
• A nice instrument pack – I have Garritan’s GPO ($200), I have also heard that the MODARTT Pianoteq piano emulation package (Mac/Win) works well.

1. Preparing the keyboard

1. Before proceeding, plug in the keyboard into your desktop or laptop and confirm that it works before you start modding it. This will greatly ease your mind as you work.
 

88es Controls 2592x1944

2. Take apart the keyboard , note that you need to keep only the screws that hold the keyboard assembly to the shell, and the ones that hold the switches, volume control and pitch bend wheels to the shell.

3. Unplug the electronics from the keyboard assembly and put them in Ziploc bags and set them aside, safe from damage.

88es Controls back 2592x1944

4. Discard the plastic cover, except for the small area that the controls and wheels mount on. Roughly saw that out (you will trim it down later) and keep it with the electronics.

5. Saw off the excess part of the white keys. I just left the keys mounted on the assembly and taped the assembly with wood and masking tape so the keys could not move and the plastic chips would not get into the workings of the key assembly (some did anyway). I then just passed it through the table saw several times, cutting through the white keys a groove a ¼ inch (3-4 mm) deeper each time (easier on the nerves, less risk, and avoided melting plastic).  The cut was made about 1/8 inch (2 mm)  beyond the tip of the black keys.
Note: a better way to do this would be to make new wooden keys with the right shape, (or have them pressed out of plastic should we have enough demand). I will do this for my next version. If you are interested, I’ll tell you how to do it (basically make a block of wood with the right cross-section, then slice off keys, drill 2 holes) you can make several while you are at it (very little extra work) and we will trade favours.

6. Level (top) the black keys so they are flush with the white keys. I did this by taking off each black key and clamping it into a piece of wood that I cut out / grooved to hold the black key and pass through the table saw correctly. This was slow and some needed building up and trimming to make them flush.
I suggest instead fixing the black keys so they can’t move and turning the key assembly sideways (on edge) and using the saw blade to trim off the top of  the black keys.

88es Key Assembly top during hacksawing 1944x2592

7. Remove the wrapping and vacuum off the keys thoroughly.

8. Look over the keyboard assembly and find the place where the two printed circuit boards (PCBs) abut; this is where the keyboard will be split. Mark this point on the steel rails that make the bottom of the keyboard rigid (the M-Audio keyboards are definitely good units), that is mark the  rail’s Front-left, Front-right, Back-right, Back-left sections. Unscrew the rails and saw them at the marked point.

9. Remove some keys around the PCB abutment point (I used a small flat-bladed screwdriver to gently lift the spring out of the slot, then inserted the screwdriver into the fulcrum access hole/attachment point – generally a gentle twist disengaged the catch.

10. disconnect the little cable connecting the two PCBs together & tape the floppy end down
11. with a fresh, high quality (the top part is steel – remember this is a good keyboard) blade, hack-saw the keyboard into 2 parts.
You are done the demolition. From here on it’s all building. I'm out of time, but here's a summary:

Portable jammer assembly side view

1. Cut a mounting-board (dimensions to be posted later). 
   
   
2. Mount the key-assembly halves on it, facing each other and almost touching. Use the rails to determine where to place the holes.
   
   
3. Tilt them at a 10-degree angle to each other. this is to stop the pinching one would otherwise get. I used the back-rail to prop it up, plus a little extra.
   
   
4. I suggest re-drilling the black keys spring holes a bit deeper. You need to  hold the keys in a clamp to do this (otherwise they will jump out the side slot). Cut them to the same extension (Not the same depth) as the white keys. make all the hole to the exact same depth.  Consider also doing the white keys down a bit, 2 mm (1/8 inch), other wise they are a bit firm.
   
   
5. Cut a wedge of wood for the key-posts at 10 degrees, then slice it in three, one stick for each post-key height, then cross-cut the sticks to the exact width of a key. This gives you the key-posts.Through the magic of designers luck
   

   ZipEx2 assembly - cutting key-posts
   , this works out well.
   
   
   
6. Mount the key-posts with wood-to-plastic glue; Weldon #13 worked well.
   
   
   
7. Make or buy the Key-Caps, I used 1/8 inch plastic sheets, cut to a hex shape, then shaped and coloured. this worked well, but took more time than all the rest of the work.
   I suggest seeing if you can get 1-inch plasitic disks.
   
   
8. Stick them on with double-edged tape (use the best)
   
   

   Adding Key-posts
   
   
   
9. Make the protective cover (dimensions to be added here)
   
   

   ZipEx2 assembly - after painting
10. Paint it all.
11. Mount the electronics.
    

    Portable Jammer note layout
12. plug it in.
13. I'll send you a run -time to map the keys, as shown:

That's all for now - if you want to build your own, post a comment, and I will gladly add detail.

====================================
Here's the end product to the left. 

A portable jammer, finished

Currently the pitch-bend and mod wheels are hidden. I plan to made them into thumb controls.

The extra keys at the sides (one PCB was 10 notes longer than the other, so there was overhang) are octave-shifting function keys, used to make it pretend to have more rows and keys. In effect I have a 96-key keyboard that pretends, when needed, to be two keyboards, one for each hand. This all is documented here.

The blue-black keys in the center divide the two sides of the instrument, and are dynamically assigned to whichever hand is playing near them, see Colouring. The coloring pattern was an educated guess, but it seems I hit true gold.

With the special function keys, it's easy to play in any key.

Notes on the Zipex Jammer

Discoveries:
Fingers have to learn timing: ‘Tis non-linear pressure and timing needed to create the right sound at the right time. The nervous system must learn the exact blend in order to make a uniform sound.

Pianos must be standardized for this reason, so jammers will have to be too.
Notes on the Zipex Mark II jammer
How well the current design works and what I would change.
Key size and spacing,
The keys are larger / spaced out more than optimum – about 26-27 mm – bigger than the piano – a 12-15% reduction to slightly less than the width of a piano key would be good – the lower limit would be the spacing of a computer keyboard’s keys, to 19 mm  or 25% less. This would allow one to span an octave, and put more keys in easy reach, particularly of the little finger.
I suggest 22 mm spacing would be about right when we start making our own units from scratch.
Key length and angling.
The original keys were angled towards the center by 10 degrees, to reduce the pinching motion of keys coming together at the center. In conjunction with rounding the key-caps, this worked.

Weighting
The keys, because they are missing the weight on the end pulling it down have a shorter key, are a little harder to depress than a piano, but still quite workable, as the keys are uniform in touch and position.

Note that I drilled the white key’s spring-hole down so that the spring pull matches the black key’s (they were not the same; the black keys were set significantly lighter than the longer white keys), using my drill press set to an exact depth, and, crucially, a clamp to hold it in exactly the right place (it wanted to wander out of the hole); this was quite successful. I am much temped to drill the spring-holes down a bit more to lighten the touch.

It was noted in another posting that if the keys are made more pc-keyboard-like in feel, then a new player might be able to use their extensive experience with typing on a computer to learn to type.  I would suggest the keys should have an upward force a bit lighter than the tip of a piano. See http://www.stanwoodpiano.com/PolarRuleInstructions.pdf and http://theorem.ca/~my2iu/workshop/piano/piano.html  - about 7-9  grams.

Shape of the key-caps
The “spindle” profile for the key caps worked well. With a pure hexagon, the side corners can catch the finger (it’s a bit sharp) and also make the gap a bit small. Grinding off 1 mm gave a 12% gap increase, and stopped my fingers from getting jabbed.

The top and bottom point can be left on; it fills in the hole nicely.
The high-quality, thick double-sided tape (3M brand) works well. I’d planned on using it as a temporary measure, but will delay this indefinitely. The keys do move a bit on the tape, but can be gently pushed back.

On the C-thru Axis key board the keys have beveled edges – this may help in centering and is an idea that should be tried. http://www.keyboardmag.com/pop-up.aspx?content_id=27426&section_code=5

Streamlining the process
The most labour-intensive part was in making the keys and key-caps.
Making wooden keys would perhaps not be hard. It would be made out of a single block, cross-cut to the right shape, and then sliced up, just like a piano key is made. This could be much better. The trick would be getting the pivot snap and press-point points.

Making the keycaps was a royal pain – Alternative techniques need to be developed.

• could make a rod of the right cross-section, either of wood, or acrylic, and slice off sections.
• Use a  molds and a setting resin
• Have a few million made in China