MNX Draft Specification

3. Infrastructure

3.1. Terminology

3.1.1. Notational idioms

A notational idiom is a set of rules in the world for encoding music as some set of visual markings, which can be interpreted by musicians to produce an audible performance.

3.1.1.1. Conventional Western music notation (CWMN)

This notational idiom comprises a set of notational rules common to (but not limited to) Western European music from circa 1600 to the present day.

3.2. Common syntaxes

There are various places in MNX that accept particular data types, such as note values, numbers or durations. This section describes the conformance criteria for content in those formats, and how to parse them.

3.2.1. Common parser idioms

The space characters, for the purposes of this specification, are U+0020 SPACE, U+0009 CHARACTER TABULATION (tab), U+000A LINE FEED (LF), U+000C FORM FEED (FF), and U+000D CARRIAGE RETURN (CR).

The White_Space characters are those that have the Unicode property "White_Space" in the Unicode PropList.txt data file. [UNICODE]

This should not be confused with the "White_Space" value (abbreviated "WS") of the "Bidi_Class" property in the Unicode.txt data file.

The control characters are those whose Unicode "General_Category" property has the value "Cc" in the Unicode UnicodeData.txt data file. [UNICODE]

The uppercase ASCII letters are the characters in the range U+0041 LATIN CAPITAL LETTER A to U+005A LATIN CAPITAL LETTER Z.

The lowercase ASCII letters are the characters in the range U+0061 LATIN SMALL LETTER A to U+007A LATIN SMALL LETTER Z.

The ASCII letters are the characters that are either uppercase ASCII letters or lowercase ASCII letters.

The ASCII digits are the characters in the range U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9).

The alphanumeric ASCII characters are those that are either uppercase ASCII letters, lowercase ASCII letters, or ASCII digits.

The ASCII hex digits are the characters in the ranges U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9), U+0041 LATIN CAPITAL LETTER A to U+0046 LATIN CAPITAL LETTER F, and U+0061 LATIN SMALL LETTER A to U+0066 LATIN SMALL LETTER F.

The uppercase ASCII hex digits are the characters in the ranges U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9) and U+0041 LATIN CAPITAL LETTER A to U+0046 LATIN CAPITAL LETTER F only.

The lowercase ASCII hex digits are the characters in the ranges U+0030 DIGIT ZERO (0) to U+0039 DIGIT NINE (9) and U+0061 LATIN SMALL LETTER A to U+0066 LATIN SMALL LETTER F only.

Some of the micro-parsers described below follow the pattern of having an input variable that holds the string being parsed, and having a position variable pointing at the next character to parse in input.

For parsers based on this pattern, a step that requires the consumer to collect a sequence of characters means that the following algorithm must be run, with characters being the set of characters that can be collected:

1. Let input and position be the same variables as those of the same name in the algorithm that invoked these steps.

2. Let result be the empty string.

3. While position doesn’t point past the end of input and the character at position is one of the characters, append that character to the end of result and advance position to the next character in input.

4. Return result.

The step skip white space means that the consumer must collect a sequence of characters that are space characters. The collected characters are not used.

When a consumer is to strip line breaks from a string, the consumer must remove any U+000A LINE FEED (LF) and U+000D CARRIAGE RETURN (CR) characters from that string.

When a consumer is to strip leading and trailing white space from a string, the consumer must remove all space characters that are at the start or end of the string.

When a consumer is to strip and collapse white space in a string, it must replace any sequence of one or more consecutive space characters in that string with a single U+0020 SPACE character, and then strip leading and trailing white space from that string.

When a consumer has to strictly split a string on a particular delimiter character delimiter, it must use the following algorithm:

1. Let input be the string being parsed.

2. Let position be a pointer into input, initially pointing at the start of the string.

3. Let tokens be an ordered list of tokens, initially empty.

4. While position is not past the end of input:

   1. Collect a sequence of characters that are not the delimiter character.

   2. Append the string collected in the previous step to tokens.

   3. Advance position to the next character in input.

5. Return tokens.

For the special cases of splitting a string on spaces and on commas, this algorithm does not apply (those algorithms also perform white space trimming).

3.2.2. Numbers

3.2.2.1. Rational numbers

A string is a rational number if it is either an integer, or a pair of integers separated by a U+002F SLASH whose second element is nonzero.

The rules for parsing rational numbers are as given in the following algorithm. When invoked, the steps must be followed in the order given, aborting at the first step that returns a value. This algorithm will return a pair of integers, one for the numerator and one for the denominator which must be nonzero, or an error.

1. Let input be the string being parsed.

2. Let position be a pointer into input, initially pointing at the start of the string.

3. Let fraction be an initially empty list of integers.

4. Collect a sequence of characters that are space characters. These are skipped.

5. While position is not past the end of input, and fraction contains fewer than two elements:

   1. Collect a sequence of characters that are not space characters, ASCII digits, U+002D HYPHEN-MINUS or U+002F SLASH characters. This skips past leading garbage.

   2. Collect a sequence of characters that are not space characters or U+002F SLASH, and let unparsed number be the result.

   3. Let number be the result of parsing unparsed number using the rules for parsing signed integers.

   4. If number is an error, set number to zero.

   5. Append number to fraction.

   6. Collect a sequence of characters that are space characters, or U+002F SLASH.

6. If fraction has no elements, return zero.

7. If fraction has only one element, append 1 to fraction.

8. Return the first element of fraction as the numerator and the second element of fraction as the denominator.

3.2.3. Element locations

An element location constitutes a reference to a specific element in the document. It consists of the character #, immediately followed by the XML ID of the referenced element.

3.2.4. Style property lists

MNX supports a simple and compact style property list syntax, allowing a map of key-value pairs to be represented in a single string where the keys are names of style properties.

To parse a style property list:

1. Let input be the string being parsed.

2. Let defs be the result of strictly splitting the string input using U+003B SEMICOLON as a delimiter.

3. Let properties be an empty map.

4. While defs is not empty,

   1. Let definition be the first element of defs, and remove it from defs.

   2. Collect a sequence of characters from definition that are not U+003A COLON, and let property name be the result after stripping leading and trailing white space.

   3. If property name is empty, return an error.

   4. If the next character of definition is not U+003A COLON, return an error.

   5. Skip the next character of definition.

   6. Let property value be the remaining characters of definition, after stripping leading and trailing white space.

   7. Add a new entry to properties with key property name and value property value.

5. Return properties.

Examples include:

color: red

A definition of the property color as having the value red.

color: green;

A definition of the property color as having the value green. Note that a terminal ; is provided in this case, but has no effect.

smufl-font: Bravura; color: red;

A definition of two properties: smufl-font with value Bravura, and color with value red.

3.3. Content models and categories

Each element in MNX falls into zero or more categories that group elements with similar characteristics together. Examples of content categories include event content and sequence content, among many others.

3.3.1. Element definitions

Each element in this specification has a definition that includes the following information:

Contexts

A non-normative description of where the element can be used. This information is redundant with the content models of elements that allow this one as a child, and is provided only as a convenience.

Content model

A normative description of what content must be included as children and descendants of the element.

Attributes

A normative list of attributes that may be specified on the element (except where otherwise disallowed), along with non-normative descriptions of those attributes. (The content to the left of the dash is normative, the content to the right of the dash is not.)

Style properties

A normative list of style properties that may be specified on the element (except where otherwise disallowed), along with non-normative descriptions of those attributes. Where these attributes may be inherited from ancestor elements, this is indicated.

This is then followed by a description of what the element represents, along with any additional normative conformance criteria that may apply to producers and consumers and implementations. Examples are sometimes also included.

4. Notational concepts

This section describes various foundational concepts in music notation that are frequently referenced by this specification.

4.1. Parts and staves

A score consists of multiple parts. Each part is a grouping of related musical material that relates to a single performer or set of performers. It has the same temporal extent as the score overall, but presents a slice of content that is relevant to a single instrument or a group of related instruments.

A part may employ one or more staves. Each staff supplies a pair of dimensions, usually one for pitch and one for time, within which notes may be placed. Conventionally, the time dimension is horizontally oriented; for pitched instruments, the pitch dimension is vertically oriented. All staves within a part share the same time dimension.

For unpitched instruments, the vertical dimension indicates a choice of sound rather than a pitch, governed by a set of conventions that map note placement to sound.

Every segment of a staff possesses a clef that determines the mapping between its pitch dimension and some set of performable pitches, with additional information supplied by a key signature. Accidental symbols on notes further modify this mapping on an ad-hoc basis.

Staves in CWMN are identified within a part by a unique staff index. The topmost staff in a part has a staff index of 1; staves below the topmost staff are identified with successively increasing indices.

4.2. Notated events

A notated event in CWMN is a discrete action in the score with a notated duration. It has an onset that is relative to the start of its containing sequence as well as to other elements in that sequence, subject to the conventions of CWMN. Events belong to a specific staff within a part, denoted by its staff index.

A notated event may include one or more notes possessing a pitch, or a rest indicating silence. Events including more than one note are referred to as chords.

In both cases, the event possesses an associated note value that indicates its notated duration. This value is not literal, but is subject to performance interpretation.

The content of a notated event includes only the specific notes and chords within it. In particular a notated event does not account for ties, ornamental interpretation or many other kinds of performance reading. As such, the onset, duration, pitch and other properties of notated events will often differ from those in the corresponding performance events.

Events may further possess articulations, additional properties that modulate their musical performance in commonly understood ways. In this specification, we use the term articulation in an expanded sense to cover all such additional properties.

Notated events are represented by the event element.

4.3. Metrical position

Most notated events possess a well-defined metrical position, giving a time onset expressed as a rational number of whole note durations after the start of its containing measure. This position may be thought of as the event’s "address" and plays a determining role in the normative rendering and performance of events.

4.4. Chromatic pitches

A chromatic pitch describes a pitch situated in a 12-tone temperament notated as per CWMN conventions. The description incorporates three elements:

• The diatonic step, which is one of the letters A through G, describing the corresponding diatonic steps.

• The octave, which is an integer giving the octave in which the step occurs. The assignment of pitches to octaves follows the Scientific Pitch Notation (SPN) convention, in which the octave number increases on the step from B to C. For example, the SPN for the range of steps around middle C on the piano keyboard includes the sequence: ...A3, B3, C4, D4....

• The alteration, which is a fractional alteration of pitch from the given step and octave, expressed in terms of the prevailing temperament of the work.

4.5. Directions

A direction is a discrete instruction in the score that applies to notated events.

Directions do not have a duration, although they have a specific location in relation to a containing measure or sequence. Like events, directions also belong to a specific staff within a part, denoted by its one-based staff index.

Directions come in the following flavors:

Single-ended directions begin at a point in time in some part and generally continue to apply until superseded by another direction, or by some notation in the score that is understood to terminate it. An example is a piano dynamic.

Span directions begin at a point in time in some part and end at a later point within the same part. One common example of a span direction is a slur.

Liaison directions begin on one note and end on an immediately succeeding note. A common example of a liaison direction is a tie.

4.6. Notated sequences

A notated sequence is a set of notated events whose notional time intervals do not overlap, which lie within the same measure and the same part, and which occur at progressively greater temporal offsets within a measure.

Notated sequences are represented by the sequence element.

4.7. Voices

Notated sequences may also belong to voices. A voice is a set of sequences in different measures, but within the same part. The sequences within this set can be thought of as constituting a single musical voice throughout the score. Thus they are an organizational construct, rather than a notational one.

A given voice need not be expressed in every single measure of a part. It may be present in some measures, and absent in others.

Even so, the assignment of sequences to voices has concrete implications for MNX implementations. Producer implementations may interpret a voice as affecting the way that musical material is organized, for example by cutting and pasting material from a given voice in one measure into the same voice in a different measure. Consumer implementations might allow users to isolate the playback of a single voice, including only the sequences that belong to it.

4.8. Performance Interpretation

A performance interpretation is the end result of deriving a set of performance events from a score. Human performers do this by reading music, while MNX consumers will generally do this algorithmically.

In MNX, a consumer will typically employ a set of rules that mimic the actions of a human performer, overriding these with exact performance events where these are explicitly supplied within the document.

4.9. Performance events

A performance event is a description of a single timed element of a musical performance with specific attributes for its onset, duration, pitch, dynamics, articulation and instrument. Unlike a notated event, these attributes are specific and not subject to interpretation, and they are independent of any notational concepts.

Performance events in MNX are used to describe the exact performance interpretation of some or all of a score, as distinct from its notation.

4.10. Note values

There are a variety of situations in which the note value of a musical event needs to be described, in terms of some fraction or multiple of a CWMN whole-note unit.

In CWMN, fractions for undotted base note values are constrained to be exact powers of two. The most common note values of whole, half, quarter, etc. correspond to whole-note fractions expressed by the non-negative powers 2⁰, 2^-1, 2^-2. The less frequently used note values of breve, longa, etc. are expressed by the positive powers 2¹, 2², ...

In the broader case of general note values, some number of dots act as a multiplier on the base note value. These multipliers take the form (2ⁿ⁺¹-1) / 2ⁿ, where n is the number of dots.

4.11. Orientation

Events and sequences may possess an optional orientation that determines a the placement and rendering of content according to a complex set of CWMN conventions. For the purposes of MNX there are two orientations:

• An orientation of up orients note stems pointing upwards, and places beams, directions and articulations accordingly.
• An orientation of down orients note stems pointing downwards, and places beams, directions and articulations accordingly.

4.12. Written pitches

A note’s written pitch is the pitch that would sound if the note’s notation were performed by a concert-pitch instrument. Written pitch can be thought of as the note’s pitch from the musician’s instrument-specific perspective.

A note’s written pitch might be different from its sounded pitch, which is the pitch that an instrument generates. Written pitch differs from sounded pitch in case of notation written for a transposing instrument, such as a clarinet.

MNX considers ottava markings (e.g., "8va," "8vb") to be purely presentational. Hence, ottava markings have no effect on a note’s written pitch. For example, if a note is rendered on the bottom staff line without an ottava marking, and a second note is rendered on the top staff space with an "8vb" marking, the two notes have the same written pitch.

5. Notational syntaxes

5.1. Note value syntax

MNX provides a microsyntax for encoding note values whose syntactic constraints map to the above requirements. Its syntax is designed to be distinguishable from other syntaxes for integers, floating point numbers or rational numbers. The syntax for base note values consists of either of the following forms:

1. For values less than or equal to a whole note:

   1. The character U+002F SLASH

   2. One or more ASCII digits encoding the base note value as a power-of-two fractional denominator

2. For values greater than a whole note:

   1. The character U+002A ASTERISK

   2. One or more ASCII digits encoding the base note value as a power-of-two multiplying factor

The syntax for general note values consists of these components:

1. A base note value encoding.

2. Zero or more occurrences of U+0064 LOWERCASE D characters. The number of occurrences supply the number of dots.

To parse a note value, use the following procedure:

1. Let input be the string being parsed.

2. Let position be a pointer into input, initially pointing at the start of the string.

3. Let number of dots be 0.

4. If the character indicated by position is a U+002A ASTERISK character (*), let fractional be false and advance position by 1.

5. Else, if the character indicated by position is a U+002E SLASH character (/), let fractional be true and advance position by 1.

6. Else, return an error.

7. Collect a sequence of characters that are ASCII digits only and let unparsed number be the result.

8. Let base value be the result of parsing unparsed number using the rules for parsing integers.

9. If parsing a general note value, collect a sequence of characters that are U+0064 LOWERCASE D characters. Set number of dots to the length of this sequence.

10. If position is not at the end of the string, return an error.

11. If base value is not equal to a power of 2, return an error.

12. If base value is equal to 1 and fractional is false, return an error.

13. If fractional is true, set base value to (1 / base value).

14. Return base value and number of dots.

5.2. Note value quantity syntax

MNX allows the specification of a note value quantity, defined as an integer multiple of a note value. To parse a note value quantity, use the following procedure:

1. Let input be the string being parsed.

2. Let position be a pointer into input, initially pointing at the start of the string.

3. Let multiplier be 1.

4. Collect a sequence of characters that are ASCII digits only, and let unparsed number be the result.

5. If unparsed number is not empty, assign unparsed number to multiplier using the rules for parsing integers.

6. Let note value be the result of parsing the remainder of string beginning at position according to the rules to parse a note value.

7. Return multiplier and note value as the result.

Examples include:

/8

a single eighth note

6/8

six eighth notes

6/8d

six dotted eighth notes

5/1

five whole notes

5.3. Time signature syntax

MNX allows the specification of a time signature, consisting of a sum of ordered, undotted note value quantities defining the meter of a measure. The sum may optionally share a common denominator. To parse a time signature, use the following procedure:

1. Let input be the string being parsed.

2. Let tokens be the result of strictly splitting the string input using U+002B PLUS as a delimiter.

3. If tokens is empty, return an error.

4. Let shared denominator be true.

5. Let fractions be an empty list.

6. While tokens is not empty,

   1. Remove the first element of tokens and assign it to t after stripping leading and trailing white space.

   2. If t contains the characters U+002F SLASH or U+002A ASTERISK,

      1. Let nv be the result of parsing t as a note value quantity.

      2. If nv has a number of dots greater than zero, return an error.

      3. If shared denominator is true,

         1. Replace the denominator in each element of fractions with the denominator of nv.

         2. If more elements remain in tokens,

            1. Set shared denominator to false.

      4. Append nv to fractions.

   3. Else,

      1. If tokens is empty, return an error.

      2. If shared denominator is false, return an error.

      3. Let numerator be the result of parsing t as a valid integer.

      4. Append the fraction composed of numerator and the denominator 1 to fractions.

7. Return fractions and shared denominator as the result.

Examples include:

3/4

Three-quarters time

2+3+2/8

A compound time signature of 2/8, 3/8 and 2/8, with 2+3+2 over the shared denominator 8.

2/8 + 3/4 + 2/8

A compound time signature of 2/8, 3/4 and 2/8 as separate fractions (note that the spaces are ignored)

Shared denominators are all-or-nothing. So there’s currently no way to share a denominator for only some of a time signature’s fractions, which would require a grouping construct like 2/4+(2+3/8) or such.

5.4. Chromatic pitch syntax

MNX allows the specification of a chromatic pitch in a single string, by employing the rules for parsing a chromatic pitch. To parse this syntax, employ the following procedure:

1. Let input be the string being parsed.

2. Let position be a pointer into input, initially pointing at the start of the string.

3. Let alteration be 0.

4. If the character at position is not an uppercase ASCII letter in the range from U+0041 UPPERCASE A - U+0047 UPPERCASE G, return an error.

5. Let step be the character at position, and advance position by 1.

6. If the character at position is U+0023 HASH,

   1. While the character at position is U+0023 HASH,

      1. Increase alteration by 1.

      2. Advance position by 1.

7. Else, if the character at position is U+0062 b,

   1. While the character at position is U+0062 b,

      1. Decrease alteration by 1.

      2. Advance position by 1.

8. Collect a sequence of characters that are ASCII digits only and let unparsed number be the result.

9. Let octave be the result of parsing unparsed number using the rules for parsing integers.

10. Let alteration factor be 0.

11. If the character at position is U+002B PLUS,

    1. Set alteration factor to 1.

    2. Advance position by 1.

12. If the character at position is U+002D HYPHEN-MINUS,

    1. Set alteration factor to -1.

    2. Advance position by 1.

13. If alteration factor is not equal to zero,

    1. Collect a sequence of characters that are ASCII digits, U+002E FULL STOP, or U+002F SLASH and place the result in unparsed number.

    2. If the character at position is U+006F LOWERCASE o

       1. Multiply alteration factor by 12.

       2. Advance position by 1.

    3. Else, if the character at position is U+0077 LOWERCASE w

       1. Multiply alteration factor by 2.

       2. Advance position by 1.

    4. If unparsed number contains U+002F SLASH, parse it as a rational number, otherwise parse it as a valid floating-point number. Multiply the result by alteration factor and add this to alteration.

14. If position is not at the end of the string, return an error.

15. Return step, octave and alteration as the result.

Examples include:

C4

Middle C

C#4

The C-sharp above middle C

Db4

The D-flat above middle C

Dbb4

The D-double-flat very near middle C

C4+0.5

The pitch one quarter-tone above middle C

C4+0.25w

The pitch one quarter-tone above middle C (identical to the above, but expressed in whole tone units)

C4+1/4w

The pitch one quarter-tone above middle C (identical to the above, but expressed as whole tone fraction)

C4+1/24o

The pitch one quarter-tone above middle C (identical to the above, but expressed as octave fraction)

5.5. Measure location syntax

There are a variety of situations in which the measure location of a musical event needs to be described, in terms of the content of the measure.

The following cases exist for specifying measure locations:

1. If the measure location is a metrical position in the context of some containing measure, then it is specified as a valid floating-point number or note value quantity that gives the number of whole notes from the start of the measure.
2. If the measure location is a metrical position in the context of an arbitrary measure in the score, then it is specified as a pair of tokens separated by U+003A COLON. The first token is a measure index identifying the measure, and the second token is a valid floating-point number or note value quantity that gives the number of whole notes from the start of the identified measure. The identified measure must belong to the same measure content as the element in which the measure location is given. This requirement has the effect of ensuring that the measure index is unique, since measure locations cannot reference measures in other systems with different barring.
3. If the measure location is identical to the metrical position of some known event in the score, then it is specified as a element location identifying the event. The identified event must belong to the same measure content as the element in which the measure location is given.

5.6. SMuFL glyph name syntax

Some contexts, particularly the glyph property, permit the specification of a SMuFL glyph name from the catalog of glyph names defined in the SMuFL specification.

6. Document structure

6.1. Root structure and metadata

6.1.1. The head element

6.1.2. Metadata content

Metadata content may be included in many elements to supply bibliographic data and other descriptive information.

Many elements TBD. Need to harmonize with existing metadata and bibliographic standards.

6.1.3. The title element

6.1.4. The mnx element

<mnx>
  <global>
      ...measure content describing system-wide features...
  </global>
  <part>
      ...part description content...
      ...measure content for part 1...
  </part>
  <part>
      ...part description content...
      ...measure content for part 2...
  </part>
  ...additional parts...
</mnx>

<global>
  <measure index="1">
    <directions>
      <tempo bpm="120" value="4"/>
      <time signature="4/4"/>
    </directions>
  </measure>
  <measure index="2"/>
  <measure index="3"/>
  <measure index="4"/>
</global>

<part>
  <part-name>Violin</part-name>
  <part-abbreviation>Vln</part-abbreviation>
  <instrument-sound>strings.violin</instrument-sound>

  <measure>
    <sequence>...</sequence>
  </measure>
  <measure>
    <sequence>...</sequence>
  </measure>
  <measure>
    <sequence>...</sequence>
  </measure>
  <measure>
    <sequence>...</sequence>
  </measure>
</part>

6.2. Measure content

Measure content supplies a sequence of measure elements, each of which supplies musical content for a time interval within a score.

The placement of the measures in measure content constitutes their score order, which is the order in which they are logically presented to a reader. This is distinct from their performance order, which is the order in which they are played by a performer.

Each measure may bear the index attribute, which provides its unique measure index within the score order for this measure content. The first measure in a system has an index of 1.

6.2.1. The measure element

Contexts:

global, part

Content Model:

Metadata content

Zero or one directions elements

One or more sequence elements (for measures within part elements only)

Interpretation content

Attributes:

index - an optional integer index for the measure

number - an optional textual number to be displayed for the measure

barline - an optional ending barline type for the measure

The measure element encloses the direction and sequence content that together make up the majority of musical content in an MNX score.

The optional attribute index defines the one-based index of this measure within score order. This is used for cross- referencing measure elements in corresponding runs of measure content. The default value of index is 1 for the first measure element within a run of content; for all other measures its default is the index of the previous measure element plus 1.

The optional attribute number provides an non-negative integer which is to be used as a visual label for the measure. It is not required to be unique. If omitted, its default value is the same as index.

The optional attribute barline defines a barline type for the barline drawn at the end of the measure. Allowed values include:

regular

dotted

dashed

heavy

light-light

light-heavy

heavy-light

heavy-heavy

tick

short

none

If barline is not given, then the barline should be interpreted as follows:

1. If the measure is the last in the document, use light-heavy.

2. Otherwise, use regular.

The following example shows both direction content and a single-voice sequence with two monophonic half notes:

<measure index="2">
  <directions>
    <dynamics type="f" location="0"/>
    <wedge type="diminuendo" location="0" end="1/2"/>
    <dynamics type="p" location="1/2"/>
  </directions>
  <sequence>
    <event value="/2">
      <note pitch="C4"/>
    </event>
    <event value="/2">
      <note pitch="C5"/>
    </event>
  </sequence>
</measure>

6.2.2. The sequence element

Contexts:

measure

Content Model:

Metadata content

Zero or one directions elements

Zero or one beams elements

Sequence content

Interpretation content

Attributes:

orient - default orientation of direction and sequence content

staff - default staff index of direction or sequence content

voice - optional cross-measure voice identifier

The sequence element organizes a set of musical events within a measure into a strict temporal sequence, accompanied by relevant directions. The assignment of measure positions to these events is accomplished by sequencing the content, with a starting position of 0 and a time modification ratio of 1.

The sequence content within each sequence supplies the music for a single polyphonic voice within its containing measure, including notes, chords, rests, beam groups, tuplets and grace note runs.

The optional orient attribute provides a default orientation for all content within this sequence. If not provided, the orientation is determined automatically according to the implementation’s rendering rules, and may also be overridden by descendants.

The optional staff attribute provides a default staff index for all content within this sequence. If not provided, the staff index is determined automatically according to the implementation’s rendering rules, and may also be overridden by descendants.

The optional voice attribute supplies a string that identifies the voice to which this sequence belongs. All sequence elements in a given part having the same value of voice belong to the same voice. Within a given measure element, no two sequence elements may share the same value for voice. The value of voice is an opaque identifier that does not supply information from producers to consumers.

This example shows a sequence with a single monophonic voice including a series of events that together comprise a 4/4 measure. A single staff is used, so no staff attribute is present.

<measure>
    <sequence>
        <event value="/2">...</event>
        <event value="/4">...</event>
        <event value="/4">...</event>
    </sequence>
</measure>

Here’s a more complex measure that shows two melodic voices with independent rhythms on different staves, each represented by a sequence element:

<measure>
    <sequence staff="1">
        <event value="/2">...</event>
        <event value="/4">...</event>
        <event value="/4">...</event>
    </sequence>
    <sequence staff="2">
        <event value="/2d">...</event>
        <tuplet inner="3/8" outer="1/4">
            <event value="/8">...</event>
            <event value="/8">...</event>
            <event value="/8">...</event>
        </tuplet>
    </sequence>
</measure>

If the voices in the previous example shared a single polyphonic staff, it might look like this instead:

<measure>
    <sequence orient="up">
        <event value="/2">...</event>
        <event value="/4">...</event>
        <event value="/4">...</event>
    </sequence>
    <sequence orient="down">
        <event value="/2d">...</event>
        <tuplet inner="3/8" outer="1/4">
            <event value="/8">...</event>
            <event value="/8">...</event>
            <event value="/8">...</event>
        </tuplet>
    </sequence>
</measure>

The following example shows a typical organization of sequences within a measure for a SATB-style grand staff with four voices, two on each staff:

<measure>
    <sequence orient="up" staff="1">...</sequence>
    <sequence orient="down" staff="1">...</sequence>
    <sequence orient="up" staff="2">...</sequence>
    <sequence orient="down" staff="2">...</sequence>
</measure>

When a direction content element is included in a sequence, it acquires a measure location identical to that of the following event, or to the end of the measure if there are no more events. The following example specifies a dynamics element which applies to the start of the following note:

<sequence>
  <directions>
    <dynamics location="0" type="mp"/>
  </directions>
  ...preceding sequence content...
  <event value="/4">
    <note pitch="C4"/>
  </event>
  ...following sequence content
</sequence>

Note that directions may also occur within a directions element at the start of a sequence, in which case they are assigned explicit locations. In this example, dynamic changes are specified over the course of a single whole note which occupies the entire measure:

<sequence>
  <directions>
    <dynamics type="f" location="0"/>
    <wedge type="diminuendo" location="0" end="1/2"/>
    <dynamics type="p" location="1/2"/>
  </directions>
  <event value="/1">
    <note pitch="C4"/>
  </event>
</sequence>

6.2.3. The directions element

6.2.4. The beams element

6.2.4.1. The beam element

6.2.4.2. The beam-hook element

6.3. Sequence content

Sequence content supplies a series of musical events that are both presented and performed in a given order, each at a distinct time. Such events express the concepts of chords, notes and rests.

Sequence content possesses a starting position. This is the metrical position within a containing measure of the content’s first element.

Sequence content also possesses a time modification ratio. This is a rational number scale factor which implicitly applies to all positions and durations within the content.

Sequence content also permits interspersed direction content whose directions are injected into the sequence either adjacent to events, or at explicitly given measure locations.

Within sequence content, nested event content is assigned metrical positions according to the following procedure, called sequencing the content:

1. Let sequence cursor be the starting position of the sequence content.
2. Let content to the list of elements comprising the sequence content.
3. While content is not empty:
   1. Let next be the initial element of content, and remove it from the head of content.
   2. If next is an event element:
      1. If next has a measure value of yes,
         1. If sequence cursor is greater than zero, throw a processing error.
         2. Set sequence cursor to the end of the measure as defined by its time signature.
      2. Else,
         1. Set the metrical position of next to sequence cursor.
         2. If next has a duration attribute, assign it to event duration.
         3. Else, set event duration to next’s value attribute.
         4. Multiply event duration by the time modification ratio, and add the result to sequence cursor.
   3. If next is a forward element:
      1. Set the metrical position of next to sequence cursor.
      2. Add the duration of next, multiplied by the time modification ratio, to sequence cursor.
   4. Else, if next is a tuplet element:
      1. Sequence the content of next, using sequence cursor as the starting position, and multiplying the time modification ratio by the tuplet's outer / inner ratio for the processing of the tuplet.
      2. Add the total duration of next as given by outer, multiplied by the time modification ratio, to sequence cursor.
   5. Else, if next is a grace element:
      1. Process the contents of next, assigning them a non-metrical ordering relative to preceding or following elements as appropriate.
   6. If sequence cursor exceeds the specified duration for the enclosing element (time signature for a measure, inner attribute for a tuplet), throw a processing error.

6.3.1. The event element

Contexts:

sequence, tuplet

Content Model:

Metadata content

Either zero or more note elements, or one rest element.

Event content

Interpretation content

Attributes:

value - the notated metrical duration of this event

measure - optional flag indicating that the event occupies the entire measure.

orient - optional orientation of this event

staff - optional staff index of this event

duration - optional performed metrical duration, if different from value

Style properties:

stem-direction - the stem direction of this event

The event element represents a notated event: a discrete period of time within a sequence during which one or more notes are performed, or in which a rest occurs.

All events other than whole-measure events require a value attribute to provide their duration as a note value. This duration is implicitly multiplied by the current time modification ratio, as specified by the process of sequencing the content of the event’s containing element.

Here’s an example of a simple event representing a half note:

<event value="/2">
  <note pitch="C4"/>
</event>

With more than one note, the event becomes a chord:

<event value="/2">
  <note pitch="C4"/>
  <note pitch="E4"/>
  <note pitch="G4"/>
</event>

With a rest element alone, the event is a rest:

<event value="/2">
  <rest/>
</event>

NOTE: It is legal for an event to have neither notes nor a rest. The result is functionally identical to forward, but is more constrained since an event must have a valid note value, while a space can have a multiple thereof.

If the optional measure attribute is given as yes, then the event is a whole-measure event which occupies the entire measure. Whole-measure events may not specify a value attribute, and must either be empty or contain exactly one rest. Here’s an example:

<event measure="yes">
  <rest/>
</event>

The optional orient attribute provides a specific orientation for this event. If not provided, the orientation is inherited from any sequence or tuplet ancestor which specified it. If no ancestor did so, it is determined automatically according to the implementation’s rendering rules.

The optional staff attribute provides a specific staff index for this event. If not provided, the orientation is inherited from any sequence or tuplet ancestor which specified it. If no ancestor did so, it is determined automatically according to the implementation’s rendering rules.

While staff could be used on a per-event basis, its primary purpose is for overriding a default staff assignment at the sequence level, as in cross-staff keyboard notation. The following example illustrates a lower-staff keyboard voice that temporarily crosses into the upper staff:

<sequence staff="2">
  <event value="/4">
    <note pitch="C3"/>
  </event>
  <event value="/4">
    <note pitch="G3"/>
  </event>
  <event value="/4" staff="1">
    <note pitch="E4"/>
  </event>
  <event value="/4">
    <note pitch="C3"/>
  </event>
</sequence>

The optional duration attribute specifies the actual, performed duration of the event, if different from the notated value given by value. The duration is specified as a note value quantity, which is less constrained than note value: it may be any desired multiple of a note value.

The duration attribute not only alters the performance of the event, but determines the amount of time the event takes up in the measure, affecting the location of subsequent events in the containing sequence. This can change the layout of the measure, as subsequent events will occur earlier or later as a result and be positioned accordingly (see sequencing the content). It also affects the validation of the measure for metrical correctness.

This attribute is unlikely to be frequently encountered, but is needed to handle cases in which composers employ notated values that are not interpreted literally. Some examples follow.

One case occurs in Brahms, in which a dotted half is used to specify a note that is traditionally understood as occupying the space of 11 sixteenth notes:

<sequence>
  <event value="/16">
    <rest/>
  </event>
  <event value="/2d" duration="11/16">
    <note pitch="E4"/>
  </event>
</sequence>

Because duration must be specified as a legal note value multiple, an enclosing tuplet element may be required in some cases to impose an appropriate time modification ratio.

6.3.2. The tuplet element

<sequence>
    ...preceding events in sequence...
    <tuplet inner="3/8" outer="1/4">
        <event value="/8">...</event>
        <event value="/8">...</event>
        <event value="/8">...</event>
    </tuplet>
    ...remaining events in sequence...
</sequence>

6.3.3. The forward element

6.3.4. The grace element

<sequence>
    ...preceding event content...
    <grace>
      <event value="/8">...</event>
      <event value="/8">...</event>
    </grace>
    <event value="/4">...</event>
    ...following event content...
</sequence>

6.4. Event content

Event content comprises elements that describe the musical content of a single event, that is performed at a distinct time.

6.4.1. The note element

6.4.2. The rest element

6.4.3. The articulations element

6.4.4. The lyric element

6.4.5. The ornaments element

6.4.6. The technical element

6.5. Note content

Note content comprises elements that describe the musical nature of a single note within an event.

6.5.1. The notehead element

6.5.2. The fret element

6.5.3. The string element

6.6. Liaison content

Liaison content comprises elements that describe the liaison or connection between a single note within an event, and some other note.

6.6.1. Liaison attributes

Liaison elements share a set of common liaison attributes in an attribute group.

Attributes:

target - the optional element ID of the note at which the liaison ends

Liaisons in general must be provided with the ID of a note on which they end, given via the target attribute. The constraints on this attribute vary from one liaison to another.

6.6.2. The tied element

<sequence>
  <event value="/4">
    <note pitch="C4">
      <tied target="note2"/>
    </note>
  </event>
  <event value="/4">
    <note id="note2" pitch="C4"/>
  </event>

6.6.3. The arpeggiate element

6.6.4. The glissando element

6.6.5. The slide element

6.6.6. The bend element

6.6.7. The hammer-on element

6.6.8. The pull-off element

6.7. Event liaison content

Event liaison content comprises elements that describe the liaison or connection between a single event and some other event.

6.7.1. Event liaison attributes

Event liaison elements share a set of common event liaison attributes in an attribute group.

Attributes:

target - (optional) the element ID of the event at which the liaison ends

6.7.2. The slur element

6.8. Direction content

Direction content consists of some number of musical directions that modify or accompany the performance of events in one or more measures. Directions may be included in an MNX score in two ways:

• Within a directions element below a measure. The direction is considered to apply to all sequences.

• Within a directions element below a sequence. The direction applies only to the content within this sequence.

6.8.1. Direction attributes

Directions share a set of common direction attributes in an attribute group.

Attributes:

location - the measure location of the direction

staff - an optional staff index

orient - an optional orientation

All directions within a directions parent element must be given an explicit measure location by supplying a location attribute. The default measure location is zero.

Conversely, directions occurring within sequence content must omit this attribute as their location is determined during the procedure of sequencing the content.

The optional staff attribute designates the staff index to which this direction applies, if such a designation makes sense. If not provided, the orientation is inherited from any sequence ancestor which specified it. If no ancestor did so, it is determined automatically according to the implementation’s rendering rules.

The optional orient attribute provides a specific orientation for this direction. If not provided, the orientation is inherited from any sequence ancestor which specified it. If no ancestor did so, it is determined automatically according to the implementation’s rendering rules.

6.8.2. The dynamics element

6.8.3. The instruction element

<measure>
  <directions>
    <instruction>Molto adagio</instruction>
  </directions>
  ...following measure content...
</measure>

6.8.4. The expression element

<sequence>
  <directions>
    <dirgroup location="0">
      <expression>subito</expression>
      <dynamics>p</dynamics>
    </dirgroup>
  </directions>
  ...following sequence content...
</sequence>

6.8.5. The ending element

6.8.6. The repeat element

6.8.7. The coda element

6.8.8. The segno element

6.8.9. The harmony element

6.8.10. The symbol element

6.8.11. The dirgroup element

6.9. Spanning directions

Spanning directions are directions whose temporal extent within a part is characterized by a span.

6.9.1. Span attributes

Spanning directions share a set of common span attributes in an attribute group.

Attributes:

end - the measure location of the direction

Spanning directions MUST be given a measure location as their endpoint, by supplying a end attribute. This measure location must lie within the same run of measure content as the location given for the start of the spanning direction.

6.9.2. The octave-shift element

6.9.3. The wedge element

6.9.4. The cresc element

6.9.5. The dim element

6.9.6. The pedal element

6.9.7. The bracket element

Other spans exist in MusicXML and need to be migrated.

6.10. Staff directions

Staff directions are directions that apply as a whole to the one or more musical staves in a part, and which determine the interpretation of other notations within some applicable range of those staves.

Because it delineates disjoint ranges of staves, any staff direction has the effect of partitioning the events in a measure, such that all events lie either before or after the given directions. For example, consider a clef element describing a clef change. No matter how many polyphonic voices exist in a measure, all notes in all voices either lie to the left or to the right of this clef.

Staff directions that modify the interpretation or layout of the staff, apply from the start of that measure to all subsequent measures within the same global or part element until changed.

Most staff directions have a fixed location, typically the beginning or end of the measure in which they occur.

6.10.1. The key element

6.10.2. The time element

<global>
  <measure>
    <directions>
      <time signature="4/4"/>
    </directions>
    ...
  </measure>
  ...
</global>

<global>
  <measure>
    <directions>
      <time signature="3/4" measure="1/4"/>
    </directions>
    ...this pickup measure contains only 1 beat...
  </measure>
  <measure>
    ...the following measure contains 3 beats...
  </measure>
  ...
</global>

6.10.3. The tempo element

6.10.4. The staves element

6.10.5. The clef element

6.10.6. The staff-details element

6.10.7. The barline element

6.10.8. The use-instrument element

6.10.9. The transpose element

Other staff direction elements exist in MusicXML and need to be migrated.

6.11. Part description content

Part description content consists of elements that supply information describing a part, and occur at the beginning of a part element.

6.11.1. The part-name element

6.11.2. The part-abbreviation element

6.11.3. The instrument-sound element

6.12. Interpretation content

Interpretation content may be included in some elements to control the specific way in which the element is interpreted as a musical performance.

6.12.1. The interpret element

<sequence>
  ...preceding sequence content...
  <event value="/8">
    <note pitch="C4"/>
    <interpret>
      <performance-event start="0" duration="1/6"/>
    </interpret>
  </event>
  <event value="/8">
    <note pitch="D4"/>
    <interpret>
      <performance-event start="1/24" duration="1/12"/>
    </interpret>
  </event>
  ...following sequence content...
</sequence>

<sequence>
  <directions>
    <instruction class="tempo">
      <interpret>
        <performance-tempo beat="/4" bpm="60"/>
      </interpret>
      Langsamer
    </instruction>
  </directions>
  ...following sequence content...
</sequence>

6.13. Synchronization content

The following elements describe the synchronization of a semantic CWMN score with one or more audio media representing recordings of the score.

6.13.1. The score-audio element

6.13.2. The score-audio-media element

6.13.3. The score-audio-mapping element

6.13.4. The score-audio-region element

<mnx>
  <global id="global1">...</global>
  <part>...</part>
  <part>...</part>
  <score-audio system="global1">
     <score-audio-media src="recording.mp4"/>
     <score-audio-mapping start="1.43" end="2.43" start="1:0" end="1:1"/>
     <score-audio-mapping start="2.43" end="2.98" start="2:0" end="2:0.5"/>
     <score-audio-mapping start="2.98" end="3.53" start="2:0.5" end="2:1"/>  ...slowing down...
     <score-audio-mapping start="3.53" end="7.53" start="3:0" end="6:1"/>    ...1st repeat + ending...
     <score-audio-mapping start="7.53" end="10.53" start="3:0" end="5:1"/>   ...2nd repeat...
     <score-audio-mapping start="10.53" end="11.53" start="7:0" end="7:1"/>  ...2nd and final ending.
  </score-audio>
</mnx>

The class attribute may be used on any MNX element, and supplies the value of a style class definition which applies to a that element as per the rules of style property computation.

The value of this attribute supplies the names of one or more style class definitions which apply to the containing element as an ordered set of space-separated tokens. All style property values supplied by each class definition are applied to the element, in the order in which they were defined.

For example, the following applies a class named emphasis to a note:

<note class="emphasis" pitch="C4"/>

In this case, two different classes emphasis and alternate are applied, along with a local overriding color:

<note class="emphasis alternate" style="color: blue;" pitch="C4"/>

Applies to:

part

Sequence content

Direction content

Event content

Note content

Liaison content

Value:

Simple color

Inherited:

yes

Simple colors as per the HTML5 spec don’t support an alpha property, so perhaps we should adopt a separate syntactical definition here.

Applies to:

part

Sequence content

Direction content

Event content

Note content

Liaison content

Value:

A SMuFL font name

Inherited:

yes

Applies to:

Direction content

Note content

Value:

SMuFL glyph name

Inherited:

no

Applies to:

part

Sequence content

Direction content

Event content

Note content

Liaison content

Value:

normal | none

Inherited:

yes

The display property controls the way in which an element and its descendants interact with the layout of the document.

Permitted values include:

• normal the element is displayed normally.

• none the element is not processed for display and layout proceeds as if it did not exist.

Other values are likely, making this property into more than simply a way of hiding content.

Applies to:

Sequence content

Direction content

Event content

Note content

Liaison content

Value:

visible | hidden

Inherited:

yes

The visibility property controls whether an element’s contents (including all of its descendants) are displayed by a consumer or not.

In contrast to display with a value of none, a visibility value of hidden does not affect the layout of any other elements in the document. For example, the contents of a hidden event will not be shown, but the place where the event would have appeared will still occupy space in the containing measure.

Permitted values include:

• visible: the element is visible

• hidden: the element is hidden

Applies to:

Sequence content

Direction content

Event content

Note content

Liaison content

Value:

normal | none

Inherited:

yes

Permitted values include:

• normal the element is performed

• none the element is omitted from performance

Applies to:

Direction content, Sequence content

Value:

staff position

Inherited:

no

The x property places the horizontal anchor point of an event or direction at a given graphical offset relative to its specified measure location, given as a valid floating-point number of staff lines starting from the measure location’s assigned position in the layout and proceeding to the right in a positive direction.

Applies to:

Direction content

Value:

staff position

Inherited:

no

The end-x property places the horizontal anchor point of the end of a spanning direction at a given graphical offset relative to its specified measure location, given as a valid floating-point number of staff lines starting from the measure location’s assigned position in the layout and proceeding to the right in a positive direction.

Applies to:

Direction content

Value:

staff position | above | below

Inherited:

no

The special values above and below delegate exact positioning of the direction to the implementation and request that the direction be placed respectively above or below the staff.

TBD: describe rendering model and registration

Applies to:

Direction content, Sequence content

Value:

location

Inherited:

no

For layout purposes only, the location property overrides the measure location of an event or direction determined by sequencing the content, replacing it with the value location expressed as a measure location within the containing measure.

This is useful for forcing the positioning or alignment of objects that would otherwise be placed elsewhere, without otherwise altering their semantics.

Applies to:

Event content

Value:

up | down

Inherited:

no

The stem-direction property controls the direction of any rendered stem associated with this event. If omitted, the stem direction is determined automatically by the implementation, in accordance with the orientation of the event.

The value up causes an event’s stem to be rendered pointing upwards.

The value down causes an event’s stem to be rendered pointing downwards.

Applies to:

grace

Value:

yes | no

Inherited:

no

The grace-slash property controls whether or not a slash is rendered in conjunction with the flag or beam of a grace note. If omitted, the stem direction is determined automatically by the implementation based on the nature of the grace note.

Contexts:

head, mnx

Content Model:

Stylesheet definitions.

The stylesheet element serves to place stylesheet definition elements under a single container to support clean document organization and validation.

Contexts:

stylesheet

Content Model:

None.

Attributes:

name - the name of this style class definition

style - the style property list applied by this definition

The style-class element supplies a style class definition, which associates a list of style property values with a class that can be referenced elsewhere using its name alone, as per the rules of style property computation.

The name attribute supplies the name of this style class definition; the style attribute supplies the properties that make up the content of the definition.

Multiple occurrences of style-class are permitted to share the same value of name. These are equivalent to a single occurrence of style-class with the same constituent definitions of style property values in the same order.

Here is a style class definition that creates a class called emphasis, intended to color its target objects bright red and apply a thicker stem width in the case of events:

<style-class name="emphasis" style="color: #FF0000; stem-width: 0.05;"/>

Contexts:

stylesheet

Content Model:

None.

Attributes:

rule - a set of element names to which this style selector definition applies

style - the style property list applied by this definition

The style-selector element supplies a style selector definition, which defines a list of style property values applying to all elements matching a style selector rule, as per the rules of style property computation.

The rule attribute supplies a style selector rule that is automatically applied to all semantic MNX elements to determine a set of implied style properties. The following rule syntax is supported:

• An unordered set of space-separated tokens, each of which is the name of an MNX element. Elements whose names belong to this set are considered to match the rule.

The style attribute supplies the properties that make up the content of the rule.

Multiple occurrences of style-selector are permitted to share the same value of rule. These are equivalent to a single occurrence of style-selector with the same constituent definitions of style property values in the same order.

Note: The scope of a defined rule is not limited to descendants of the element in which the style-selector element occurs, but is global to the entire document.

For example, the following specifies that all event elements in the document will employ a given stem width:

<style-selector rule="event" style="stem-width: 0.05;"/>

6.16. Rendering

TBD: section describing normative MNX rendering procedure, leaving room for implementation decisions. The intent is to set out the normative constraints that MNX rendering must follow, including at least:

6.17. Interpretation

TBD: section describing normative MNX performance interpretation, leaving room for implementation decisions.