MML has often been used as a language for describing music in strings, for example, in the BASIC language. The mml_parser is a minimalistic Music Macro Language parser library written in pure C intended for resource-constrained platforms, especially microcontrollers and other embedded systems.
Add “#” or “+” after the note indicates Sharp. ex: "C#" "C+." Add “-” after the note indicates Flat. ex: "C-"
Length of the tone can be specified in two ways. One is to specify a length for each note. Add a number after the note in 1, 2, 4, 8, 16, 32 or 64. ex: "C8 C16 C#4."
The other is to use L . The L sets default length. If the note without length, the number which is indicated by the “L” is used. The number followed after the “L” can be in 1, 2, 4, 8, 16, 32 or 64. ex: "L4 A" means “A” with length 4.
In addition, dot is supported. For example, length of “C4.” is 4 + 8. Length of “C16..” is 4 + 8 + 16.
Rest is represented by “R”. Length of the rest is following after the “R” as the same as the note length. If no length is specified, the length specified by the ”L" is used.
Chord is supported. If some notes are enclosed in parentheses by [ and ], they are interpreted as a chord. ex: "[CEG]" is a chord of Do, Mi and Sol. Chord's length can be put after ] . ex: "[CEG]4" is a chord with 4 length.
Note: Max notes in a chord is defined as MAX_CHORD_NOTES in mml_parser.h.
Tuplet is supported. If some notes and Chord are enclosed in parentheses by { and }, they are interpreted as a tuplet. ex: "{C E G [CEG]}" is a tuplet with C, E, G and chord of CEG. Tuplet's length can be put after “}”, and the length is divided equally among each note. ex: in "{C E G [CEG]}4" case, C, E, G and chord CEG has each a quarter of the L4 length.
Octarve is controlled by “O”, “>” or “<”. When “O” is used, the O is followed by a number indicating the octave. When “>” is used, the value of the new octave is the current octave plus one. When “<” is used, the value of the new octave is the current octave minus one. ex: "CDEFGAB > C R C < BAGFEDC", "O4 CDEFGAB O5 C R C O4 BAGFEDC"
Tempo is indicated as “T” and numter following after the “T”. Tempo number decide a speed of the score. This value is used for culculating sample number for the note (or rest). ex: "T120"
Volume can be controlled by “V”. And numter following after the “V”. ex: "V4"
Tempo 120, Voulume 10, Octave 4, Default length is 4.
"T120 V10 O4 L4 C. D8 E. C8 E C E2 D. E8 {FF} {ED} F2"
mml_parser is providing just 2 functions.
Initialize an instance of mml parser.
#include <audioutils/mml_parser.h> int init_mml(FAR struct music_macro_lang_s *mml, int fs, int tempo, int octave, int length);
The function initializes struct music_macro_lang_s instance provided as 1st argument. The argument fs is a sampling frequency of target audio output system, and this value is used for calculating sample number in case of a tempo and length of note. tempo, octave and length specify initial values for tempo, octave, and length, respectively.
On success, init_mml() returns 0. On error, it returns an negative value.
Currently no error is happened.
Parse MML from given string.
#include <audioutils/mml_parser.h> int parse_mml(FAR struct music_macro_lang_s *mml, FAR char **score, FAR struct mml_result_s *result);
parse_mml() parses the first MML of the string given by the argument score and gives the result in the return value and the argument result. The result is an instance of mml_result_s, which contains note_idx, length, and chord_notes as members. The meaning of the value of each member depends on the return value.
On error, a nevative value is returned. On success, following values can be returned. And those values are defined in mml_parser.h.
| Return values | Description |
|---|---|
| MML_TYPE_EOF | This means that it have reached the end of the string. The content of the result has no meaning. |
| MML_TYPE_NOTE | This indicates that some note has been parsed. The scale of the note is stored in note_idx[0]. The length of the note is given by the length member as the number of samples. In the case of tuplet, this return value is returned at the time each note is parsed. In other words, a tuplet is parsed as a single note. |
| MML_TYPE_REST | This indicates the rest has been parsed. The length of it is given by the length member as the number of samples. |
| MML_TYPE_TEMPO | This indicates "T" is parsed. length member in result has the value of the tempo. But tempo value is kept in mml instance for calculating sample number for each notes. So basically, no need to handle this return value in your code. |
| MML_TYPE_LENGTH | This indicates "L" is parsed. length member in result has the value of the parsed length. But current length value is kept in mml instance. So basically, no need to handle this return value in your code. |
| MML_TYPE_OCTAVE | This indicates "O", ">", or "<" is parsed. length member in result has the value of the octave. But the octave is encoded in note_idx in MML_TYPE_NOTE case. So basically, no need to handle this return value in your code. |
| MML_TYPE_TUPLETSTART | This indicates tuplet is just started. And total length of the tuplet is stored in length of result members. |
| MML_TYPE_TUPLETDONE | This indicates the tuplet is just finished. |
| MML_TYPE_VOLUME | This indicates "V" is parsed. length member in result has the value of the parsed volume. |
| MML_TYPE_TONE | T.B.D. |
| MML_TYPE_CHORD | This indicates a chord is parsed. Chord has some notes, and how many notes is stored in chord_notes member of result. And each notes are stored in note_idx[]. Length of the chord is stored in length member of result. |
The value of note_idx[] is encoding octave, like
| Octave | Note | node_idx value |
|---|---|---|
| O0 | C | 0 |
| O0 | C# | 1 |
| O0 | D | 2 |
| O0 | D# | 3 |
| O0 | E | 4 |
| O0 | F | 5 |
| O0 | F# | 6 |
| O0 | G | 7 |
| O0 | G# | 8 |
| O0 | A | 9 |
| O0 | A# | 10 |
| O0 | B | 11 |
| O1 | C | 12 |
And so on.
So for example, G# at Octave 4 is encoded as 56.
Following error code can be received as return value.
| Error code | Description |
|---|---|
| MML_TYPE_NOTE_ERROR | |
| MML_TYPE_REST_ERROR | |
| MML_TYPE_TEMPO_ERROR | |
| MML_TYPE_LENGTH_ERROR | |
| MML_TYPE_OCTAVE_ERROR | |
| MML_TYPE_VOLUME_ERROR | |
| MML_TYPE_TUPLET_ERROR | |
| MML_TYPE_TONE_ERROR | |
| MML_TYPE_CHORD_ERROR | |
| MML_TYPE_ILLIGAL_COMPOSITION | |
| MML_TYPE_ILLIGAL_TOOMANY_NOTES | |
| MML_TYPE_ILLIGAL_TOOFEW_NOTES | |
| MML_TYPE_ILLIGAL_DOUBLE_TUPLET |
Please see examples/mml_parser
There are plenty. Report them on GitHub, or - even better - open a pull request. Please write unit tests for any new functions you add - it's fun!
mml_parser was written by Takayoshi Koizumi <takayoshi.koizumi@gmail.com>