The ideas and design goals behind LinnStrument

My (Roger's) design of LinnStrument was driven by a desire to create a new musical instrument for the modern era, an instrument that overcomes the limitations of both current acoustic and electronic instruments, and is not limited to the capabilities of centuries-old technology. Here's an overview of the design questions and design process I went through.

What's wrong with acoustic instruments? 

Acoustic instruments are marvelous and have produced centuries of beautiful music, but they have some problems. Here's a brief list of advantages and disadvantages that I find in some of the popular acoustic instruments:

Guitar

 + Good note expression, able to vary pitch over a note's duration by vibrato and bends

+ Polyphonic, up to 6 notes

+ Fairly easy to learn to play pleasing tones

+ Portable

+ Easy to transpose-- just slide your hand or capo up or down

+ Overlapping string pitch ranges allow multiple instances of each pitch, permitting multiple fingerings of each chord or scale

- Always produces a fast attack / long decay envelope

- Difficult to play chords and melody at the same time

- Because of equal temperament, somewhat difficult to play just 3rds or 5ths

- Inefficient use of hands: requires both hands to play a single note

- Because of the odd 3rd interval between the B and G strings, playing a specific chord or scale requires 3 different fingerings depending on which strings are used.

- Because of the string physics, notes are big at one end and small at the other.

- The strings go out of tune

 

Piano 

+ Polyphonic

+ Easy to play pleasing tones and to play in tune

+ Pitch arrangement emphasizes the C major scale and is therefore easy to learn and play simple music in the key of C major, or A or D minor.

+ Easily adapted to an electronic keyboard with on/off switches

- Always produces a fast attack / long decay envelope

- Poor note expression, unable to vary loudness, pitch or timbre over a note's duration

- Not isomorphic; playing a specific chord or scale requires a different fingering for each of the 12 transpositions.

- Only one instance of each pitch

- Because of equal temperament, can't play "just" 3rds or 5ths

- Because of the original design constraint of needing to place a string behind each key, the pitches are in a single, long line and therefore difficult to see both hands at once. Isn't it interesting that people are playing electronic piano keyboards, given that there is no longer a need for a string behind each key?

- Because the pitches are not equally spaced, the piano pitch layout is not well-suited for expressive pitch control. For example, a whole-tone pitch slide from C to D is the same physical distance as a half-tone physical slide from B to C, a pitch slide to or from an accidental requires a diagonal movement, and a vibrato performed on a "C" key will have more pitch change in the right direction (to D) than the left direction (to B).

- Not very portable

 

Bowed strings 

+ Excellent note expression, able to control loudness, pitch and timbre continuously over the note's duration.

+ Can slide continuously from note to note

+ Violin and viola are very portable

+ Overlapping string pitch ranges allow multiple instances of each pitch, permitting multiple fingerings of each chord or scale

- Difficult to learn to play pleasing tones

- Because of the lack of frets, it is difficult to play in tune

- Because of string physics, semitones are closer together at high pitches than low pitches

- On violin, because high notes requires very short strings, the pitches are very close together and therefore difficult to play in tune

- Can only play two notes at a time, usually only one

- Limited pitch range for each instrument (violin, viola, cello, double bass), requiring 4 different instruments to produce full pitch range.

- Violin and viola must be held uncomfortably between chin and shoulder

- In order to play low notes, the double bass is very large and not very portable

 

Wind instruments (brass and woodwinds)

+ Excellent note expression, able to control loudness, pitch and timbre continuously over the note's duration.

+ Portable, except for bass instruments like tuba

- Difficult to learn the reed or valve technique to play pleasing tones

- Limited pitch range, requires multiple instruments of each type to cover entire pitch range

- Can't play more than one note at a time

- Note layout is optimized for the opening and closing of holes in the instrument, not for ease of playing.

- Difficult to transpose; for a given scale, it is necessary to learn a different fingering for each of the 12 transpositions

 

In summary, the problems with acoustic instruments

In reading the points above, you may notice two recurring themes:

1) The human interface of mechanical instruments is severely compromised by the need to produce acoustic sound.

2) The human interface of each mechanical instrument is different, requiring that you learn the new instrument's interface and pitch locations in order to produce its sound.

In the age of synthesis, why must I learn a new instrument fingering merely to produce its sound?

 

The promise of synthesis

The promise of synthesis was that you'd be able to produce any instrument sound polyphonically from a MIDI piano keyboard, a human interface that most of us already know. And synthesis has made tremendous advances in its note expression capabilities. However, if you've ever tried to play a convincing guitar, sax, violin, clarinet or cello solo on a MIDI piano keyboard, you've probably found it to sound static and lifeless. This is because keyboards can't do much more than turn sounds on and off at different volumes, and their bend and mod wheels (essentially sideways knobs) aren't very helpful. For truly expressive performance, you need to capture the same sort of human gestures that are used to play solo acoustic instruments, including continuous gestural control of each note's loudness, pitch and timbre.

Guitar-to-midi controllers have proven to have many of the same limitations in producing realistic non-guitar tones. And although playing synthesizers from electronic wind controllers like the Akai EWI/EVI can produce quality note expression in the right hands, these controllers have the same problems of their acoustic counterparts: only one note can be played at a time, and the note layout is difficult to learn and play.

 

My design goals in creating LinnStrument

My goals in designing LinnStrument were:

* The human interface should be optimized for human fingers and hands, not limited by the need for acoustic sound generation.

* It should be highly expressive, able to control each note's loudness, pitch and timbre continously over time. It should be able to perform convincing performances of expressive traditional acoustic instruments, and apply that expressive control to new imagined sounds.

* It should be polyphonic, with independent expressive touch control for each of simultaneous touches

* The note layout should be isomorphic-- for any given chord or scale, the fingering is the same for all 12 musical keys.

* Multiple instances of each pitch, to permit a) playing accompaniment and melody in the same pitch range, and b) using multiple alternate fingerings for a given chord or scale

* Efficient use of hand/finger gestures. Like a piano keyboard, only one finger should be required to produce a single note.

* Easy to play in tune, yet subtle pitch nuances are possible. You should be able to easily play notes in tune while also playing vibrato, bend and other pitch nuances.

Design questions and challenges

There were a variety of design questions in the design of LinnStrument.

Separate, moving keys vs. a continuous, flat playing surface

We are familiar with separate keys on piano, which work very well for turning sounds on and off. However, separate moving keys have two significant problems:

1) You can’t slide continuously from one key to another in order to perform pitch bends, pitch slides or any other expressive pitch gestures. Bend wheels are a poor and unrealistic workaround this omission.
2) Separate keys can be made to move a little in the horizontal direction in order to perform vibrato, but it is very difficult to hold it steady when you don’t want to change pitch, which results in unintended pitch variation, even if the effect on pitch is set low.

For this reason, I decided to use a flat continuous surface in LinnStrument:

1) You can slide continuously from one pitch to another. This is intuitive; to bend pitch from C to E, you play the C note pad then slide your finger up to the E note pad.
2) Like a violin, you can slide from any pitch to any other pitch, but once you find the correct pitch, it is easy to lock your finger to the playing surface in order to keep the pitch steady
3) LinnStrument's note pads are slight raised, so like separate keys, you can feel the physical location of each note pad.

A spongey surface vs. a hard surface

A spongey playing surface provide desirable tactile feedback when playing sounds that require pressure to control volume, but is less desirable when playing sharp-attack sounds. By comparison, a hard surface works well for playing sharp-attack sounds but provides only audible feedback for pressure-based sounds.

So LinnStrument uses a slightly compressible playing surface of only 2 millimeters thin. This provides a superior tactile feedback for sharp-attack sounds while providing a small amount of tactile feedback for pressure-based sounds, which already have feedback in the form of audible feedback-- the sound volume tells you how how you're pressing. The other advantage of LinnStrument's thin playing surface is that it transfers a much larger amount of your finger's pressure to the touch sensor below it, allowing for higher sensitivity to light touches.

How to arrange the pitches?

Given the popularity of the piano key arrangement, many would think that this would be best choice, but it doesn't work well for expressive performance of pitch gestures. So I preferred the arrangement of any stringed instrument-- muliple rows of consecutive semitones, which solves nearly all of my above-stated design goals. The argument between piano and string layout is discussed on this page, and as well as how I arrived at the default 4ths row tuning and the relative merits of a grid layout vs. a hex-key layout.

LinnStrument, my best guess for the instrument we'll be playing in 50 years

In summary, LinnStrument's design is my attempt at a new instrument, based on modern technlology, that achieves the above-stated design goals. And it iis my best guess at the human interface for musical performance that most people will be playing 50 years from now.