From: JMP100%DE0HRZ1A@UK.AC.RL.EARN 27-MAY-1987 12:28 To: ARCHIVE Subj: english code description Via: UK.AC.RL.EARN; Wed, 27 May 87 12:26:44 BST Received: From DE0HRZ1A (JMP100) By UK.AC.RL.IB (MAILER) ; Wed, 27 May 87 12:26:39 BST Date: Wed, 27 May 87 13:21:12 GMT From: JMP100%DE0HRZ1A@UK.AC.RL.EARN To: ARCHIVE@UK.AC.OXFORD.VAX3 Subject: english code description DB: Encoding Music with ESAC p. 1 Universit{t Essen - Gesamthochschule Prof. Dr. Helmut Schaffrath Cooperation: Barbara Jesser Translation: Christoph Lepping Essen, May 14, 1987 (= update) 1 ENCODING MUSIC WITH THE "Essen-Associative-Code" (ESAC) _______________________________________________________________ 1.1 DESIGN OF THE DATABASE OF MELODIES __________________________________________ Various concepts have been developed which make it possible to transcribe musical notation into codes that are accepted by comput- ers. The most popular systems of transcriptions are presented in Steinbeck (1982). Those concepts might be considered to be appro- priate or even ideal for our purposes. But if you know how one- part (folk song) melodies are codified or if you think of the nota- tion based on numbers which is a Tonic-Do method - a method employed in the whole of South-East and East Asia - you will rightly ask: Why not use this associative rather than utterly abstract code? Notation is a code. So if it were possible to transform the nota- ___________________ tion employed in the field at issue in such a way that on the one hand the highest measure of unambiguity and on the other hand all associative features are retained then every musician could play those characters by sight which the computer is supposed to read. Furthermore such a code is useful because it makes analyses and encoding of Chinese folksongs drawn up at Essen University directly readable for the respective colleague in the People's Republic. Likewise this meets the requirements of our data-base system which already contains information of about more than 3.000 pieces of music (ETNO; as of 3/84). In this data-base blocks of codes are listed meter-wise (or phrase-wise when the music is metrically free) in a thesaurus where they are available for every analytic query. But no-one will search motives, particles etc. if the music (= "words") has to be transcribed into one of those codes which are far too abstract. So anyone who makes use of our system will - in a query based on few distinct criterions (not nesessarily a coded tune) - find two things: on the one hand the information he was looking for and on the other hand an easily accessible and readable code (i.e. it can easily be transformed into normal notation.). So the consultation of different publications becomes superfluous. The melodies should be encoded in a way that allows the highest DB: Encoding Music with ESAC p. 2 possible amount of visual similarity to the Asian numerical nota- tion on the other hand uncompromising unambiguity shall provide the prerequisites for automatic re-coding. This certainly implies the possibility to print or let plot the melodies in "common" notation if needed. The keynote always lies within the one-accented octave (comp. key-line 260). It's conceivable that an additional program may contain the instruction to admit that position of the octave only, in which the smallest number of added lines are produced. Design of the Database of Melodies LIAO _______________________________________ At the moment the database of melodies LIAO is compatible with the data-base of sound recordings ETNO. The data are being recorded in files (record length 72, filetype CARDIN). The first three figures of every line are reserved for the running line-number (correspond- ing to the paragraphs in ETNO). The appendix of this manual contains information concerning the order of the paragraphs of a document. (3.1 - 3.3) DB: Encoding Music with ESAC p. 3 1.2 PITCHES _______________ 1.2.1 PITCH-SYMBOLS The pitches are defined by their respective intervals based on the keynote: Prime = 1 Minor Second = 2b Major Second = 2 Minor Third = 3b Major Third = 3 Fourth = 4 Augmented Fourth = 4# Diminished Fifth = 5b Fifth = 5 Minor Sixth = 6b Major Sixth = 6 Minor Seventh = 7b Major Seventh = 7 Rest = 0 Further symbols can be used to represent augmented intervals (3#, 4b). Compare 1.2.2. Alterations. 3b is used as a symbol for the Minor Third even when the context proves undoubtedly that the mode is minor. So '3' always means 'Major Third'. The end of the song is marked by BLANK and two diagonal bars //. 1.2.2 ALTERATIONES Alterations are indicated by one of the two symbols '#' or 'b' behind the respective tone. These symbols are available after a conversion of the line-separator with the help of 'SET LINEND ON %' in the Editor-Mode. 1.2.3 POSITIONS OF THE OCTAVE A - (minus) in front of the number indicates the position of the octave under the keynote, whereas the position of the octave above the keynote is indicated by a + (plus). Likewise two Minus or two Plus can be set if necessary. DB: Encoding Music with ESAC p. 4 1.3 RHYTHM ______________ 1.3.1 METRICAL KEY The absolute durations of tones unequivocally emerge from the coded relative durations (1.3.2) plus the indication of the smallest rhythmic note value in the key which is represented by its denomi- nator only (i.e. 04 means a crotchet, 16 means a semiquaver). 1.3.2 DURATIONS The smallest note value of the piece is represented by the pitch number without any additional symbols. By adding one (or more) ________________________________ underlines ( _ ) the respective time value is doubled each time. An example: Let the smallest time value be 08 = a quaver. According to the rule given above the various note values of the keynote are transcribed as follows: quaver = 1 crotchet = 1_ minim = 1__ semibreve = 1___ etc. Dottings are indicated in analogy to "common" notation. The dot is ________ set behind the duration to be extended. An example: Dotted crotchet = 1_. Dotted minim = 1__. etc. Binds connecting notes within a bar as well as in adjacent bars are _____ marked by ~ which replaces the pitch-symbol (comp. ex. 5 on one of the following pages). 1.3.3 PERCUSSION AND SPRECHGESANG Analogous to conventional notation 'x' is used as a 'pitch-symbol' in rhythmic notation in which the pitch itself is of no relevance. The notation of duration follows the rules given in 1.3.2. DB: Encoding Music with ESAC p. 5 1.3.4 TIME SIGNATURES AND CLASSIFICATION Time signatures are given in the traditional manner as 4/4 or 6/8 _______________ etc. Free meter is marked by FREI (comp. ex. 5). Likewise the composition of complicated meters is indicated in accordance with traditional notation; so '3+2' for instance specifies the informa- tion 'five-four meter'. The number of changes of meter in a song is optional; the maximum number of different meters, however, is limited to three. Two BLANKs in the score seperate the bars visu- ally. No distinction is made between songs that are metrically free (caesurae) or not (bar lines). In any case it must be taken into consideration that the tones between two bar lines (=BLANKs) make up one word in the thesaurus. In meters classified as 'FREI' these words may be very long. They too are seperated from each other by BLANKs (comp. ex. 5). In this field all the problems are not solved yet! Repetitions are written down in full. This simplifies the program- ___________ ming of formal analyses. 1.3.5 RESTS Analoguous to the smallest rhythmic value, rests - just like the digits are treated as 0. For example: A crotchet rest following 08 (=quaver) as the smallest metrical basic unit: 0_ (See the follow- ing pages for further examples, comp. ex. 1 - 3) 1.3.6 TRIPLETS The group of notes providing for a triplet (or a sextuplet) is put into brackets. In the following examples the semiquaver is the smallest note value: Quaver triplet = (1_3_5_) Semiquaver triplet = (135) etc. 1.3.7 PHRASING AND EXPRESSION MARKS are not encoded unless notes of the same pitch are bound. In this case the symbol ~ replaces the pitch number (comp. 1.3.2.). DB: Encoding Music with ESAC p. 6 According to these definitions the rhythm of the incipit of a song which is given as an example of an encoded song further down (Es waren zwei Koenigskinder) would run as follows (metrical key 08 = quaver): Incipit: -5_ 3_333_2_ 2_.11_ Rhythmic Values: C C QQC C. C. QC Key: C = Crotchet, C. = Dotted Crotchet, Q = Quaver. So the unequivocalness of this visual notation permits an automatic transcription into any other kind of rhythm-code if required. 1.4 KEY ___________ 1.4.1 KEYNOTE Only the keynote is encoded. It is represented by uppercase char- acters and by the symbols '#' and 'b' as the following example shows: Keynote F sharp = F#; Keynote E flat = Eb etc. The incon- sistent use of the letters H for B and Bb for B flat is a compro- mise between the German and the English names: we hope this will help to avoid misunderstandings. Normally the keynote lies within the one-accented octave. If in an example a different position of the octave occurs this will be entered in the comment (line 281-284). 1.4.2 MODE The coding of the pitches should allow of a fairly precise analysis of the mode. With the help of automatic procedures of analysis it is ascertained and entered into line 285, provided the notes avail- able form a scale from which an unequivocal identification of the mode can be drawn. But nonetheless it is expedient to analyse the mode 'by hand' if modulations, alterations and incomplete scales occur. Such remarks can be entered in the lines 281-284. DB: Encoding Music with ESAC p. 7 1.4.3 ORNAMENTS Ornaments (and glissandi) are not represented in the transcription but their occurrences are entered into the paragraph 'Comment' just like all pitches diverging from our notation. 1.5 DIVISION INTO PHRASES _____________________________ Folk songs in German language are divided into phrases to facili- tate the procedure of analysis. According to Steinbeck *) a phrase is defined as "... a musical unity of sense which is defined by a metrical quality in relation to an adjacent unity; usually it is delimited by a musical caesura...". It is decisive that "...the line, separating melodic units from each other, is set and defined precisely by syntax and content of the text...". In case of doubt the shorter melodic unit should be preferred to the longer one. _______ The division into phrases is to be seen as an additional pre-coding which has no influence whatsoever upon the unambiguity of the coded melodies. -------------------- *) W. Steinbeck, Struktur und Aehnlichkeit, Kassel 1982; p. 42 DB: Encoding Music with ESAC p. 8 1.6. A SELECTION OF CODED SONGS __________________________________ The following examples are taken from the data-base LIAO. Each of them deals with one (or more) specific problems. FIGURE 1 (different durations, dotting, rest) ES WAREN ZWEI KOENIGSKINDER E0084 08 G 4/4 -5_ 3_333_2_ 2_.11_ -5_ 4_444353 2__0_ -5_ 5_565444 3__54 32 1_11322_ 1__ 2__ 3_33542_ 1__0_ // FIGURE 2 (different positions of the octave, minor) ICH HAB DIE NACHT GETRAEUMET E0207 16 E 6/8 -5_ 1__2_3b_.21_ 5__.4__ 4_ 3b_.21_-7__1_ 2__.0__ 1_ 5__5_3b_.21_ +1__.6b__ 4_ 3b_.21_2_.1-7 1__.0__ // FIGURE 3 (triplets, modulation) WENN I MORGENS FRUEH AUFSTEH Z0258 08 D 3/4 113142 5_+135_ 2_5432 1_313_ 1_3131 5_+135_ 3_5432 1_0_ (543b) 2_0_(6b54) 3b_0_ (543b) 4_0_(6b54) 5_0_ 7b7b 7b+3b7b_5_ 3b_0_ 53b -7b7b67b6b4 3b_0__ // Modulation in bar 9 FIGURE 4 (change of meter, altered notes) MAEDELE RUCK AN MEINE GRUENE SEITE E0525 16 F 4/4 3/4 113_ 5_5_5_5_4#_5_6_5_ 5_+1_3_ 3_2_3_ 5_.44_4_3_4_ 6__5__ 5_5_ 4_4_2__6_6_ 5_5_3__ 5_5_ 4_4_2__6_6_ 5_5_3__ 113_ 5_5_5_5_4#_5_6_5_ 5_+1_3_ DB: Encoding Music with ESAC p. 9 3_2_3_ 5_.44_3_4_2_ 2__1__ // FIGURE 5 (meter FREI) ES WOLLT EIN HERR AUSREITEN D0001 08 G FREI -5-5-51_21-7b_123b 22121-5-5_-5_-6-71_ 3b_244543b_221121-7b_ -5-5-51_3b_2_1_-71__ // FIGURE 6 (Pinyin; bound notes) Ta4i ya2ng chu1 la2i xi3 ya2ng ya2ng Auf geht die Sonne C0013 08 E 2/4 121-7b 1_20 -7b121 ~-7b-50 -4-5-7b-5 11-5_ -4_-50 -7b-51-7b ~-51_ ~__ // Keytone lies within the two-accented octave; ornamentation Recording available; classification: 00151 DB: Encoding Music with ESAC p. 10 1.7 PROVISIONAL LIST OF REFERENCES CONSULTED (ABOUT 4.300 SONGS) ____________________________________________________________________ The songs coded up to now (most of them were taken from Steinbeck et al.) come from the following sources: QUELLE DOKUMENTENNAME ________________________________________________________ Boehme, Vaterl{ndische Lieder ***BOEHME Sammlung Fink ***FINK Sammlung Kretzschmer-Zuccalmaglio ***ZUCCAL Boehme Kinderlieder ***KINDER Sammlung Erk-Boehme ***ERK Lieder aus dem Deutschen Volksliedarchiv ***DVA Chinesische Lieder ***CHINA DB: Encoding Music with ESAC p. 11 2 TRANSFORMATIONAL RULES FOR CHINESE TEXTS ________________________________________________ Pinyin is a romanized form of Chinese words. It gives us a very clear idea of how the standard language (putonghua) is to be pro- nounced. But Pinyin can't convey the meaning of the words as extensively as the characters do. A stress which indicates the pitch is put on the vowel bearing the tone: - / ~ ` . Some syllables have no definite pitch. Since screen and printer don't permit of putting all of these marks on every vowel without difficulty we number the tones in the order described above. The assignment of numbers follows the interna- tional practice. The numbers 1-4 follow the respective vowel as ______ can be seen further down. E.g.: depending on the tone the word hua has (among others) the following meanings: hua1 = flower hua2 = China hua4 = language We hope this system will allow the user to find identical words much more easily. But only a proficient user of the language will be able to cope with their ambiguities although in most cases the meaning can be inferred from the context, i.e. from additional Ger- man information about subject and function of the respective song stored in the document. Today it would already be possible to store the numbers of Chinese characters rather than the words as realised on various plot-pro- grams (in SAS on IBM for instance); these programs also allow an easy reproduction of the numbers. We, however, will try to write a little program which is supposed to 'convince' the printer of the respective accent on the vowel so that - at least occasionally - we can produce perfect Pinyin-expressions DB: Encoding Music with ESAC p. 12 3 APPENDIX ________________ The appendix describes paragraphs and formats important for encod- ing ESAC-melodies for the database LIAO. 3.1 ORDER OF THE PARAGRAPHS _____________________________ - ***DOKUMENTENNAME: Label of the source containing not more than twelve uppercase characters. - The lines 140ff contain the title of the song and optionally its text-incipit. - The lines 170ff contain information about the region. - The lines 180ff contain ethnic information. - The lines 250ff correspond to the paragraph RHYTHM-ME in ETNO. They contain the shortest version of the rhythm incipit (i.e. diminished if possible). - The lines 260ff correspond to the paragraph SKALA-MEL in ETNO. In variance to that data-base the lines 286 resp. 287-289 are labelled as 'Sequence of Cadence-Tones' and 'Sequence of stressed tones'. Apart from that, line 260 is reserved as a for- matted field for the purpose of later analyses. It is designed as follows: * 6 figures for the shelf-mark. This field is subdivided as follows: * 1 figure for the reference (uppercase character is the ini- tial letter of the name of the reference, comp. 1.6.). * 4 figures for the number of the song (right-justified; zeroes are entered into the unused figures). * 1 figure for the marking of variants (with an uppercase char- acter - this figure remains empty if not needed). * 1 BLANK * 2 figures for the smallest note value. * 1 BLANK * 2 figures for the keytone (right-justified). * 1 BLANK * 11 figures for the time signatures (indications of the meter). Up to three changes of meter - separated by the insertion of a BLANK - can be entered here. (Only 3 figures per meter are available; so in songs with changes of time the DB: Encoding Music with ESAC p. 13 meter 'FREI' must be inserted without a BLANK). * 15 figures for the structure of bars in complicated meters. They are enclosed by the %-symbol which functions as a sepa- rator. This field remains empty when the meter is simple. When songs contain changes of meter it is essential to give the correct position where the change occurs. In the follow- ing example the indication of the three-four time is omitted (that's why two %-symbols are entered one behind the other); the structure of the bar in the five-four meter is given in the second place; the third figure may ignored altogether: 260X0000 16 G 3/4 5/4 4/4%%2+3% - The coded melody is to be found in the next line (261-280) from column four onwards. Not more than 72 characters are allowed. For structural reasons the lines may start with a BLANK. - Verbal information is given in the lines 281ff: line 281-284 are reserved for the coder; lines 285-289 povide room for the results of the analysis-programs; line 285 is reserved for the analysis of the mode while the sequences of the cadence-tones and the sequences of the stressed tones can be entered into the following lines. - The results ascertained by the formal analysis-programs can be entered into line 290ff. - line 310ff; information about genre and function (-> paragraph FUNKTION). DB: Encoding Music with ESAC p. 14 3.2 DESIGN OF AN INPUT FORM _____________________________ ***DOKUMENT (name of the document) 003 FF INTERVALLE (ascending) (inserted AUTOMATICALLY) 004 FF INTERVALLE (descending) " 005 FF TONLAENGEN " 140 CUT (title, text; not more than . 30 lines) . . 170 REGION (not more than 10 lines) . 180 ETHNIE " . 250 RHYTHM-ME (rhythm incipit . inserted AUTOMATICALLY) 260 SKALA-MEL (first line FF) 261 (until line 280 262 space reserved for the . coded tune) . 281 (4 lines for analyses 282 and commentaries by . the coder) . 285 (Mode. AUTOMATICALLY) 286 KADENZTON (cadence tones.AUTOM) 287 AKZENTTON (3 lines for stressed . tones. AUTOMATICALLY) . 290 FORM (not more then 10 lines, . formal analysis. . AUTOMATICALLY) . 310 FUNKTION (not more than 10 . lines) 330 BEMERKUNG (not more than 30 lines . not used at the moment) DB: Encoding Music with ESAC p. 15 3.3 COMPARISON OF DOCUMENTS _____________________________ Input: ***BOEHME 140MARIE SASS TRAURIG IM GARTEN 170Europa, Mitteleuropa, Deutschland, Wetzlar, Schwalbach 260B0486 16 Bb 6/8 261-5_ -3_.-3-3_-6_-6_-6_ -5_-5__0__ 262-5_ -3_.-3-3_-5_-5_-5_ -4__.0__ 263-5_ -5_.-5-5_-5__-5_ -6_-6__0__ 264-6_ -4_.-4-4_-6__-6_ -5__.0__ 265-5_ -3__-5_3__2_ 2_1__0__ 2661_ 1__1_-7__1_ 2__.0__ 2671_ -7__-6_-5__-4_ -7_-7__0__ 268-6_ -5__-6_-4__-5_ -3__.0__ // 290NN 310Liebes - Lied, Leid, Klage Database - Output: BOEHME DOCUMENT= 1 OF 1 NUMBER OF LINES = 19 UNNM P/C* 4602090402020200000200000000021057 1604190704000000000000000000033D12 06483300000606000005066 CUT MARIE SASS TRAURIG IM GARTEN REGION Europa, Mitteleuropa, Deutschland, Wetzlar, Schwalbach RHYTHM-ME *_ *_.**_*_*_*_ *_*__0__ *_ *_.**_*_*_*_ *__.0__ SKALA-MEL B0486 16 Bb 6/8 -5_ -3_.-3-3_-6_-6_-6_ -5_-5__0__ -5_ -3_.-3-3_-5_-5_-5_ -4__.0__ -5_ -5_.-5-5_-5__-5_ -6_-6__0__ -6_ -4_.-4-4_-6__-6_ -5__.0__ -5_ -3__-5_3__2_ 2_1__0__ 1_ 1__1_-7__1_ 2__.0__ 1_ -7__-6_-5__-4_ -7_-7__0__ -6_ -5__-6_-4__-5_ -3__.0__ // HEPTATONIK IONISCH KADENZTON -5-4-6-512-7-3 AKZENTTON -3-6-5 -3-5-4 -5-5-6 -4-6-5 -332 1-72 -7-5-7 -5-4-3 FORM NN FUNKTION Liebes - Lied, Leid, Klage DB: Encoding Music with ESAC p. 16 Table of Contents _________________ 1 ENCODING MUSIC WITH THE "Essen-Associative-Code" (ESAC). . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 DESIGN OF THE DATABASE OF MELODIES . . . . . . . . . . . 1 1.2 PITCHES . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.2.1 Pitch-Symbols . . . . . . . . . . . . . . . . . . . . . . 3 1.2.2 Alterationes . . . . . . . . . . . . . . . . . . . . . . 3 1.2.3 Positions of the Octave . . . . . . . . . . . . . . . . . 3 1.3 RHYTHM . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3.1 Metrical Key . . . . . . . . . . . . . . . . . . . . . . 4 1.3.2 Durations . . . . . . . . . . . . . . . . . . . . . . . . 4 1.3.3 Percussion and Sprechgesang . . . . . . . . . . . . . . . 4 1.3.4 Time Signatures and Classification . . . . . . . . . . . 5 1.3.5 Rests . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3.6 Triplets . . . . . . . . . . . . . . . . . . . . . . . . 5 1.3.7 Phrasing and Expression Marks . . . . . . . . . . . . . . 5 1.4 KEY . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.4.1 Keynote . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.4.2 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.4.3 Ornaments . . . . . . . . . . . . . . . . . . . . . . . . 7 1.5 DIVISION INTO PHRASES . . . . . . . . . . . . . . . . . . 7 1.6. A SELECTION OF CODED SONGS . . . . . . . . . . . . . . . 8 1.7 PROVISIONAL LIST OF REFERENCES CONSULTED (ABOUT 4.300 Songs). . . . . . . . . . . . . . . . . . . . . . 10 2 TRANSFORMATIONAL RULES FOR CHINESE TEXTS . . . . . . . 11 3 APPENDIX . . . . . . . . . . . . . . . . . . . . . . . 12 3.1 ORDER OF THE PARAGRAPHS . . . . . . . . . . . . . . . . . 12 3.2 DESIGN OF AN INPUT FORM . . . . . . . . . . . . . . . . . 14 3.3 COMPARISON OF DOCUMENTS . . . . . . . . . . . . . . . . . 15