Purpose and Meaning of Musical Style Sheets

Musical Style Sheets is a new style sheet language used for describing the performance of musical works on digital musical instruments.

As a stylesheet language, Musical Style Sheets are related to the well-known Cascading Style Sheets (CSS) and the  Extensible Stylesheet Language (XSL). Similarly, Musical Style Sheets separate contents from presentation, in this case, musical scores from their performance. A single piece of music can be performed in many different ways, with different tempi, velocities, and articulations.

With Musical Style Sheets, musical scores are transformed to a musical performance by means of the Musical Instrument Digital Interface (MIDI). Musical Style Sheets may exist as saved stylesheet documents but also generated on the fly during a musical performance. It is also possible to combine predefined and generated styles or sequentially record different styles from performances to achieve artistically pleasing interpretations.

Processing Musical Style Sheets

In order to apply Musical Style Sheets, a stylesheet player is required, which receives and processes musical input and generates musical output.

Figure 1 shows the necessary components of a system that processes Musical Style Sheets. On the input side there are musical score documents. These are typically created by music notation or scorewriter programs and available from many musical score repositories. Musical Style Sheets are XML-encoded documents but can also be created on-the-fly by a stylesheet controller. The stylesheet player processes these input sources and generates MIDI commands which are processed by a MIDI player. The resulting sound is then created by means of soundfonts.

Optionally, the musical scores shall also be displayed. The stylesheet player is also used to synchronize  the score display with the currently playing part and other dynamic information such as the playing time or the currently played notes.

A stylesheet player might also infer stylesheets from musical markings and signs. Since these are often missing, incomplete, or imprecise, such stylesheets might rather serve as a first approach.

Selectors for Musical Styles

Since stylesheets separate contents from presentation, selectors are needed to assign different styles to different parts of a score. With musical stylesheets, there are two selectors:

• Measure selector
• Voice selector

The measure is a number which refers to the chronological position of all musical signs of a score. A voice can be a voice in a song, an instrument, a staff, or, as in the case of keyboard instruments, a layer from a staff.

A Different Measure

In printed scores, bars are assigned numbers in increasing order. The numbers serve musicians to refer to specific locations in a score. To further subdivide bars, beat numbers are used.

The traditional way of counting bars, measures and beats is well-suited for communication among musicians, but is difficult to handle in a digital environment. For example, incomplete bars such as up-beats or different ending are excluded from counting.

The measure we are using as selectors is a unified measure as it combines bar numbers and beat numbers. Specifically, our measure number shows the following properties:

• Counting measures always starts at measure 0
• All measures count, even up-beats and different endings
• All positions within a measure are expressed as a fraction of the whole measure
• Positions within a bar do not depend to the number of beats
• Repeated sections continue to count

This unified measure considerably simplifies the use of the measure selector.

Musical Styles

We consider tempo, velocity, and articulation as the basic musical styles, as they are indispensable for the performance of music and apply to all musical instruments. Furthermore, these styles are available at any time from the first to the last note of a piece of music.

In contrast to the basic musical styles, ornaments are optional. A performer may omit or add ornaments according to the own taste. Pedals apply to just a few instruments; their effects can generally be simulated by other styles.

Some musical instruments such as string and wind instruments allow continuous changes of the sound such as vibrato or pitch bending. This style is called effects.

[Should effects be included in stylesheets or or should they be left to MIDI commands?]

Tempo

Tempo is the speed at which sounds are played. To precisely indicate the musical tempo of a piece of music, the measure of beats per minute (bpm) is used. Variations in tempo—for example at the end of a piece of music (fermata) or to enhance musical expression or rhythm—are denoted by percentage deviations from the general tempo.

Musical tempo markings are attached to specific positions within a musical bar, expressed as decimal fractions of a measure. Tempo values between the markings are linearly interpolated which ensures a constant acceleration or deceleration within the marked period.

The absolute values of the tempo also depend on the note value a beat refers to. This value typically corresponds to the denominator of the time signature, but other values might be used as well. We use the beat number as number of beats per measure.

Alternatively, the tempo can be calculated from the given duration of the piece of music.

Velocity 

Velocity relates to the volume or loudness of musical sounds. The word also means “speed”, namely the speed of the hammer hitting a string as in the case of a piano. This velocity greatly influences the specter of the harmonics. To achieve most naturally sounding instruments, it is vital to use an instrument with different sound samples for different classes of velocities.

In the MIDI universe, velocity is expressed as a 7-bit value. MSS adopts this measure as a basic unit, but uses percentage deviations to express different velocities over time.

MSS defines a general velocity and a layer-specific velocity. The resulting velocity for each note equals the sum of both velocity types.

Articulation

Not all sounds are exactly played as the notes suggest. Staccato for example is a musical expression which plays notes shorter than actually notated.

Again, MSS uses percentage deviation to shorten the audible length of notes.

There is no general articulation style, only layer-specific articulations.

Ornaments

There are two types of ornaments: Ornaments which introduce new notes and ornaments which change existing notes. Examples for the former include trills, mordents, or turns, examples for the latter are grace notes.

Pedal

Pedals are instrument-specific extensions and influence the overall sound of of an instrument. A piano typically has a sustain and a soft pedal, and occasionally a sostenuto pedal.

MSS also supports pedals but does not require any actions to be taken. Pedal events are typically forwarded to a MIDI instrument.