kern: Kern pitch representation

In Computational-Cognitive-Musicology-Lab/humdrumR: humdrumR

Kern pitch representation

Description

Kern (⁠**kern⁠) is the most common humdrum interpretation for representing "notes" in the style of traditional Western scores. However! In humdrumR, the kern function outputs the pitch part of the ⁠**kern⁠ interpretation. ⁠**kern⁠ rhythms are instead created using the recip() function.

If kern() is applied to a humdrumR data class you may use the data's fields as arguments. If no field names are specified, the first selectedField is used as x.

Usage

## Default S3 method:
kern(
  x,
  ...,
  generic = FALSE,
  simple = FALSE,
  octave.relative = FALSE,
  Key = NULL,
  transposeArgs = list(),
  parseArgs = list(),
  gamutArgs = list(),
  inPlace = FALSE
)

humData |> select(Token) |> kern() 
humData |> kern(simple = TRUE)
humData |> kern(Token, Key = Key)

kern(
  x,
  ...,
  generic = FALSE,
  simple = FALSE,
  octave.relative = FALSE,
  Key = NULL,
  transposeArgs = list(),
  parseArgs = list(),
  gamutArgs = list(),
  inPlace = FALSE
)

Arguments

x	Input data to parse as pitch information. The x argument can be any (atomic) vector, or a tonalInterval, or NULL.
...	Arguments passed to the pitch deparser. There are also two hidden (advanced) arguments you can specify: memoize and deparse (see the details below).
generic	Should "specific" pitch information (accidentals and qualites) be discarded? Defaults to FALSE. Must be a singleton logical value: an on/off switch.
simple	Should "compound" pitch information (octave/contour) be discarded? Defaults to FALSE. Must be a singleton logical value: an on/off switch.
Key	The input Key used by the parser, deparser, and transposer. Defaults to NULL. Must be a diatonicSet or something coercable to diatonicSet; must be either length 1 or length(x)
transposeArgs	An optional list of arguments passed to a special transpose() call. Defaults to an empty list(). Must be a list of named arguments to transpose().
parseArgs	An optional list of arguments passed to the pitch parser. Defaults to an empty list(). Must be a list of named arguments to the pitch parser.
inPlace	Should non-pitch information be retained in the output string. Defaults to FALSE. Must be a singleton logical value: an on/off switch. This argument only has an effect if the input (the x argument) is character strings, and there is extra, non-pitch information in the input strings "besides" the pitch information. If so, and inPlace = TRUE, the output will be placed into an output string beside the original non-pitch information. If inPlace = FALSE, only the pitch output information will be returned (details below).

Details

The pitch part of ⁠**kern⁠ tokens breakdown tonal pitch information as so:

• Steps

  • 1: "C" or "c"

  • 2: "D" or "d"

  • 3: "E" or "e"

  • 4: "F" or "f"

  • 5: "G" or "g"

  • 6: "A" or "a"

  • 7: "B" or "b"

• Accidentals

  • Flat: "-"

  • Sharp: "#"

• Octave

  • Octave is indicated through the case of the step characters, as well as repetition of the step character. Uppercase letters are used for octaves below ; lowercase letters for the middle-C octave and higher. The octave, and the octave below it get one character each, with higher and lower octaves repeating that character. For example, using C# as the step value, and relative to the octave:

    • -3: "CCC#"

    • -2: "CC#"

    • -1: "C#"

    • 0: "c#"

    • +1: "cc#"

    • +2: "ccc#"

    • +3: "cccc#"

Tokens are ordered Step/Octave + Accidentals, with no separator.

Like all humdrumR pitch functions, the ways that kern parses and deparses tokens can be modified to accomodate variations of the standard ⁠**kern⁠ pitch representation.

Basic pitch arguments

Each pitch function has a few standard arguments which control details of the output. The most important are the generic and simple arguments, which allow you to control what type of pitch information is returned.

Generic vs Specific

If generic = TRUE, specific pitch information (accidentals or qualities) is omitted from the output. As an alternative way of controlling the same functionality, you can use the specific argument, where specific == !generic.

In the case of atonal functions, the "generic" version of that pitch is output: for example, semits('c#', generic = TRUE) will return 0, because the "generic" version of C# is C, which corresponds to 0. However, note that the generic version of a pitch follows the key, so semits('c#', generic = TRUE, Key = 'A:') will return 1!

Simple vs Compound

If simple = TRUE, compound pitch information (octave and contour) is omitted from the output. As an alternative way of controlling the same functionality, you can use the compound argument ,where compound == !simple.

There is actually more than one way you might want to divide compound intervals up into simple and octave parts. For example, you might like to call an output -M2 (descending major 2nd) OR +m7 (ascending minor 7th in the octave below). This functionality can be controlled with the octave.round argument: see the pitch deparsing documentation.

Key

The Key argument must be a diatonicSet, or something that can be parsed into one. The Key argument is passed to the parser, deparser, and transpose—unless an alternate Key is passed to transposeArgs or parseArgs. Various deparsing options use the Key argument; for example, use of implicitSpecies (see advanced parsing section) is dependent on the Key.

If you use any pitch function within a special call to withinHumdrum, humdrumR will automatically pass the Key field from the humdrum data to the function—this means, that in most cases, you don't need to explicitly do anything with the Key argument! (If you want this not to happen, you need to explicitly specify your own Key argument, or Key = NULL.)

Parse arguments

The parseArgs argument must be a list() of (named) arguments which are passed to the input parser. For example, if our input representation uses "X" to represent double sharps, we could specify kern('CX5', parseArgs = list(doublesharp = 'X')) and get the correct result ("cc##"). As a convenient shorthand, or "syntactic sugar," you can specify parseArgs in an alternate way: Simply input parse(args...) as unnamed argument to any pitch function. For example, we can get the exact same result as before by typing kern('CX5', parse(doublesharp = 'X')).

Transpose arguments

The transposeArgs argument must be a list() of (named) arguments which are passed to an internal call to transpose(), allowing us to easily transpose pitch information. For example, we could type kern(c('C', 'D', 'E'), transposeArgs = list(by = 'M9')) can get the output c('d', 'e', 'f#'). The possible transpose args are:

• by (tonalInterval, length == 1 | length == (x))

• from (diatonicSet, length == 1 | length == (x))

• to (diatonicSet, length == 1 | length == (x))

• real (logical, length == 1) Should transposition be real or tonal?

• relative (logical, length == 1) Should key-wise transposition be based on relative or parallel keys?

As a convenient shorthand, or "syntactic sugar," you can specify transposeArgs in an alternate way: Simply input transpose(args...) as unnamed argument to any pitch function. For example, we can get the exact same result as before by typing kern(c('C', 'D', 'E'), transpose(by = 'M9')).

Transposing by interval

As when calling transpose() directly, the by argument can be anything coercable to a tonalInterval, and the output will be transposed by that amount. If real = FALSE, tonal transposition (within the Key) will be performed. For more details on transposition behavior, check out the transpose() docs.

Transposing by key

Another way of transposing is by specifying an input ("from") key and an output ("to") key. By default, the Key argument is passed to transpose as both from and to, so nothing actually happens. Thus, if you specify either a from key or to key, transposition will happen to/from that key to Key. Of course, if you specify from and to the transposition will happen between the keys you specify.

If you use any pitch function within a special call to withinHumdrum, humdrumR will automatically pass the Key field from the humdrum data to the function. If you specify a to key, the Key field will be passed as the transpose from key, and as a result, all the pitches in the input will be transposed from whatever keys they are in to your target (to) key!

The real and relative arguments give you special control of how key-wise transposition works, so check out the transpose() docs for more details!

In-place parsing

In humdrum data, character strings are often encoded with multiple pieces of musical information right besides each other: for example, ⁠**kern⁠ data might include tokens like ⁠"4.ee-[⁠. The humdrumR parser (tonalInterval) will automatically "pull out" pitch information from within strings, if it can find any using the appropriate known regular expressions. For example, pitch('4.ee-[') returns Eb5. However, all the pitch functions (like pitch() and kern()) have an option to keep the "extra" information and return the result "in place"—i.e., embedded right where it was found in the input string. This is controlled with the inPlace argument, which is FALSE by default. So, pitch('4.ee-[', inPlace = TRUE) will return 4.Eb5[—keeping the "4." and the "[". (This obviously only works if the input is a string, not a numeric!) Note that inPlace = TRUE will force functions like semits, which normally return numeric values, to return character strings if their input is a character string.

Pitch-Gamut Levels

The table() will automatically generate factor levels for pitch data using the gamut() function. This is makes sure tabulated data sorted in a logical order, and includes missing pitches. The simple/complex and generic/specific arguments are automatically passed to gamut(); additional arguments can be passed to gamut using gamutArgs = list(...), or with the syntactic sugar gamut(...). (Read the gamut() docs for an explanation of gamut generation.) This feature be used to control table layout of pitch data, as well as to assure consistent tables when grouping data.

When inPlace = TRUE no special tabulation will occur.

See Also

To better understand how this function works, read about the family of pitch functions, or how pitches are parsed and deparsed.

Other absolute pitch functions: helmholtz(), lilypond(), pitch(), solfg(), tonh()

Other pitch functions: accidental(), bhatk(), degree(), freq(), helmholtz(), interval(), lilypond(), octave(), pc(), pitch(), quality(), semits(), solfa(), solfg(), step(), tonh()

Examples

exampleToken <- c('4GG', '4G', '4E', '4F#', '4G', '4D', '4E')
results <- kern(exampleToken)
results

exampleHumdrum <- readHumdrum(humdrumRroot, "HumdrumData/BeethovenVariations/B075_00_05_a.krn")
results <- with(exampleHumdrum[[,3:4]], kern(Token))
results

Computational-Cognitive-Musicology-Lab/humdrumR documentation built on Oct. 22, 2024, 9:28 a.m.