Lesson files

You can extend these exercises by writing lesson files: chord, chord-voicing, id-by-name, dictation, sing-chord and harmonic-progression-dictation.

If you want to write lesson files, my advise is to look at the example files in example-lesson-files/ and some of the files in lesson-files/ and use this document and the music format spec for reference. To use the lesson files, you have to save them in $HOME/lessonfiles/. If you think it is difficult to write lesson files using this documentation, you can vent your frustration to the email list solfege-devel@lists.sf.net or on the wiki at www.solfege.org/SolfegeWiki.

File encoding

New in version 2.0 is that Solfege by default expect the content of lesson files to be in UTF-8 encoding. gedit is a nice little editor that let you edit unicode files.

If you don't like unicode, you can tell Solfege that the file has another encoding by inserting a special comment line as the first line of the file. The following example set the charset to ISO 8859-1, a charset commonly used in many west-european language:

# -*- coding: iso-8859-1 -*-

Russians might want to use koi8-r:

# -*- coding: koi8-r -*-

The program use the python libs to convert to unicode, so it should understand almost any encoding you can think of. If you see some characters are missing, for example when the name of questions are displayed on buttons, then most likely you have done something wrong with the encoding.

Lesson file contents

A lesson file consist of a header block and one or more question blocks:

header {
  ASSIGNMENT
  ASSIGNMENT
  ...
}
question {
  ASSIGNMENT
  ...
}

Header block

The header block can be placed anywhere in the file, but by convention it should be the first block in the file. If more that one header block is found, for example when using the include function, only the first header block is used, the rest are ignored.

Definitions to put in the header block:

version = STRING

Tell the oldest version of solfege the lessonfile is known to work with. For example version = "1.1.1" This variable is not required, but it should be used because it can (but don't guarantee to) help avoid trouble if the lesson file format changes in the future.

title = STRING

Short oneline description. You should add this to all lesson files.

description = STRING

Long description that can contain html tags. This is optional. Use this if you have to describe how the lesson file should be used.

content = LIST

This variable tell the program what exercises should try to use this lesson file. This variable is requried. Possible values are: chord, chord-voicing, dictation, id-by-name, harmonic-progression-dictation and sing-chord.

Some lessonfiles can be useful for more than one exercise. For example, a file with chords can also be used by the id-by-name exercise, and should have: content = chord, id-by-name

musicformat = normal | chord | satb | cmdline

This variable must be set, unless the musicformat for the file is normal. FIXME. musicformat is completely ignored by singanswer because this exercise use the new music object constructor. The musicformat header variable will probably be deprecated.

normal

The default value. Can be used in these exercises: id-by-name, dictation and harmonic-progression-dictation

chord

Any music "MUSIC CODE" will be interpreted as:

\staff{ < MUSIC CODE > }

or:

\staff\transpose NOTENAME{ < MUSIC CODE > }

chord format is necessary if lesson files written for the chord exercise should be used by for example the id-by-name exercise. Can be used in these exercises: chord, singanswer, chord-voicing and id-by-name.

satb

This musicformat is only used to enter chords. Think of it as "sopran | alt | tenor | bass". The music "c | bes g | e | G C" will be interpreted as:

\staff{ \stemUp < c > }
\addvoice{ \stemDown < bes g > }
\staff{\clef bass; \stemUp < e > }
\addvoice{ \stemDown < G C > }

Can be used in these exercises: chord, chord-voicing, id-by-name and sing-chord.

cmdline

Run an external program to ask the qustion, for example

music = "aplay soundfile.wav"

Should only be used by the id-by-name exercise.

random_transpose = [key|semitones|accidentals], INTEGER, INTEGER | no | yes

In some exercises the program can transpose the music to create variation. The variable is ignored in exercises that does not support transposing.

random_transpose = no

This is the default value. No transposition will be done.

random_transpose = yes

The exercise will do random transposition. What kind of transposition depends on the exercise, but you get a ok result from this.

random_transpose = accidentals, INTEGER, INTEGER

Transpose the music by random. But make sure the number of flats or sharps are within the range specified by INTEGER, INTEGER. To explain how this works, think of flats as negative numbers and sharps as positive numbers. The code

random_transpose = accidentals, -2, 3

will give you questions with 1 or 2 flats, 0 accidentals or 1, 2 or 3 sharps. The code

random_transpose = accidentals, 2, 3

will create questions with two or three sharps.

Please notice that the music in the lessonfile has to be in c major or a minor for this to work correctly.

random_transpose = key, INTEGER1, INTEGER2

Transpose the music INTEGER1 steps down or INTEGER2 steps steps up the circle of fifth. (And by my own definition, up is more sharps and down is more flats ... :-) )

random_transpose = semitones, INTEGER1, INTEGER2

Transpose the music at most INTEGER1 steps down or INTEGER2 steps down. The code

random_transpose = semitones, -2, 6

will transpose the questions no more than 2 semitones down and 6 semitones up. This is the same as the old random_transpose = INTEGER, INTEGER syntax.

labelformat = normal | progression

(Default: normal)

questions_selectable = yes | no

Let the user select what questions to be asked from the lesson file. Works for id-by-name and is ignored by the other exercises.

have_repeat_slowly_button = yes | no

Set to yes if you want the exercise to have a "Repeat slowly" button. Works for id-by-name and is ignored by the other exercises.

have_repeat_arpeggio = yes | no

Set to yes if you want the exercise to have a "Repeat arpeggio" button. Works for singanswer and is ignored by the other exercises.

have_music_displayer = yes | no

Set to yes if you want the question to have a music displayer. The program will display the answer in the music displayer when the user gives up or answers correctly. Works for id-by-name and is ignored by the other exercises.

enable_right_click = yes | no

Default value: yes. Set to no for lesson files where it does not make sense to right click buttons. This variable is recognised by the exercises id-by-name, chord-voicing and chord.

filldir = horiz | vertic

Tell the direction the buttons are filled. Default value is horiz. Works for id-by-name and is ignored by the other exercises.

fillnum = INTEGER

Tell how many buttons there are in each row or column. The default value is 1. Works for id-by-name and is ignored by the other exercises.

Question block

name = "some short name"

Questions written for id-by-name and chord exercise need a name.

music = "\staff{c'2 g' e4 e c1}"

The music in the question. Note that there is a shortcut. Instead of:

question {
   tempo = 160/4
   name = "Lisa gikk til skolen"
   music = "\staff\relative c'{c d e f g2 g2}"  
}

you can write:

question {
   tempo = 160/4
   name = "Lisa gikk til skolen"
   "\staff\relative c'{c d e f g2 g2}"  
}

tempo = INTEGER / INTEGER

beats per minute / notelen per beat, for example 120/4.

instrument = STRING, INTEGER, instrument = STRING, INTEGER, STRING, INTEGER, STRING, INTEGER

Example:

instrument = "cello", 100

The instrument name should be one of those defined in the file soundcard/__init__.py. If you don't enter a complete name, solfege will try to find a matching instrument name. The integer is the velocity of the tones, it should be in the range 0-127. For lesson files where it makes sense, it is possible to specify three set of instruments. The following example will play bass for the lowest tone, piano in the middle and clarinet on the top tone:

instrument = "bass", 100, "acoustic grand", 100, "clarinet", 100

Solfege will try to lookup the midi instrument in this order:

  • inside the question block

  • lesson file global variable

  • instrument set in preferences window

Question block definitions for the chord exercise

In lesson files that is written to work with chord exercise, two more variables can used:

inversion

0 = root position, 1 = first inversion etc.

toptone

1, 3, 5, 7

Question block definitions for the dictation exercise

By default, the dictation exercise will show the first column of music, and then the user should write the rest. But if the first column is not good enough, if there for example are only rests on the first beat, these two variables can tell the program how much music to display:

clue_end = nn/dd

clue_end=1/4 will display all music in the first quarter note.

clue_music = MUSIC

Write the exact music you want to display. If for example you want to display the whoe bar in the upper staff, and only one note in the lower staff.

(Use only one if these variables in a question. Using both are undefined.)

Question block definitions for the sing-chord exercise

Questions for this exercise need to have the key variable set if the key signature is anything else than c major (or a minor). Example:

question { "c''|e'|g|c" }
question { "a'|e'|c'|a" }
question { key="d \major" "a'|fis'|d'|d"}
question { key="f \minor" "as'|f'|c'|f"}

Global variables

Global variables can save you a few key strokes.

s = "\score\relative c'{ %s }

question {
  # instead of music = "\score\relative c'{ c d e f g2 g2 } :
  music = s % "c d e f g2 g"
}

Also, the two variables tempo and instrument can at top level to set the defaul values for a whole file. You can override this for singel questions by putting another difinition inside the question block.

Comments

Everything after # on a line is ignored

Grammar

Disclaimer: I don't know any of the standard ways of describing grammar, like BNF. So this is just an attemt to describe the file format to get people started writing lesson files.

Lessonfiles consists of assignment statements and blocks containing assignment statements.

Functions

_

takes a string as its only argument. Use this if you want Solfege to translate the string for you. Example:

description = _("This is a short description")

include

includes another file in to this file. Example:

include("singchord-1")

Types

string

is quoted using the " character, like this: "this is a string". Use tripple quotes for strings that contain line breaks:

description = """<h1>Long desription<h1>
               This lessonfile need very much descriptions.
               bla bla bla"""

If the string need to contain the " character, you have to use tripple quotes:

description = """Try <a href="solfege:practise/melodic-interval">this
</a> for a simpler interval exercise."""

NB: All strings have to be unicode strings. If you get error messages like this one:

In line 21 of input: does not recognise this string '<' as a valid token.'
(line 20): 
question {
(line 21): question {
(line 22):   name = _("Ionia�)

then you must check the encoding of your file. See the section about file encoding. You can change the encoding of a file using the iconv program:

iconv -f YOUR_ENCODING -t utf8 your.file
integer

is an integer is an integer...

tempotype (in lack of a better name)

is entered as bpm/beatlen. An example:

tempo = 120/4

will set the tempo to 120 beats per minute, each beat being a quarter note.

Operators

Operators can only be used on strings

+

is used for joining strings

%

is similar to what you find in python, but it is very limited. It only know about %s. One example:

"\staff\relative c'{%s}" % "c d e"

evaluates to

"\staff\relative c'{c d e}"

Reserved words

OUTDATED. FIXME. To be put on the right side of assignments. The current parser does not complain if you redefines the identifiers, so don't do that!

chord, dictation, id-by-name, sing-chord

Possible values for content

horiz, vertic

Used for table filling in id-by-name exercise

normal, progression

labelformat values.

Identifiers

Identifiers match the regular expression "[a-zA-Z]+[\w_-]*".

Block

There are two block types, header and question.

blocktype {
  CONTENTS
}