BRIEF DOCUMENTATION OF SOUND CLUSTER SCRIPT
#############################################
1.will sequentially process all sound files selected at one time
2.process and extract all melody segments in tone 1, tone 2, tone 3, and tone 4 words in all sound files selected to be processed
3.concatenate sound segment extractions from each tone category into one big sound file, with each segment separated by 0.2 of silence (silence is created in accordance to the sampling frequency and number of channels of the original sound file)
4.copy tags of word identity for each segment in a generated textgrid for the clustered sound;
5.copy tags of rhythmic information for each segment(banshi)
6.copy tags of artist information
7.copy tags of position information (as in the position of the segment in a 7-word or 10-word phrase) and automatically detect the types of position for each segment. position types marked include: UB(Unit Boundary, usually marks the smallest semantic unit, such as in a 7-word phrase, there may be three units of 2+2+3 words, separated by a short instrument accompaniment, in the non-fast rhythmic/tempo type); P1B (in a two-phrase stanza, A,B, P1B is the last word of A, usually also followed by accompaniment); P2B(in a two-phrase stanza A,B, P2B is the last word of B). 
8.This script let the user choose from two modes: in mode 1, it will extract sound segments and concatenate them into sound clusters, and copy tags of 4,5,6, above; in mode 2, it allows additional copying of tag information if the sound clusters and accompanying tags are already generated.currently this include the tags of position information. 
9.it includes a functionality to determine whether current word on the word tier is the last word of the phrase on the phrase tier(sometimes the last interval on the word tier is not a word, but'ac')



##############################################
data extraction & research plans
##############################################
1.need a descriptor of pitch relations and pitch contours to describe its similarity or dissimilarity with the linguistic tone contour, then find out the consistencies and deviations for single word-to-melody relationship
2.need to extract pairwise relations (also depending on the pitch contour/relation descriptor) for melody and for linguistic tones
3.segment more arias 
4.check the factors of position, banshi, changqiang, etc on the relationship/ranking of musical and linguistic constraints;
5.when we segment and process more data, some descriptive statistics showing some of these factors;
6.extract some other relevant information of interest, such as the duration of each tone-note; 
7.the quesiton of tone-carrying information in melody; onset; contour;






##############################################

#####################################################################
brief documentation for annotation script's functionalities
#####################################################################
annotation script
************************************************
0.automatically insert tiers to be annotated(or, can detect when those tiers are already inserted)
1.annotate lyrics tier
2.annotate word tier
3.can detect if word tier is already annotated
4.can detect if lyrics tier is already annotated
5.can find discrepancies between the number of word sung and the number of word in the lyrics if they are different,and detect the differences for each phrase, thus guide human to note these differences and correct them before the script set the text for each word
6.include a functionality to segment the whole passage of lyrics into phrases
7.include a functionality to segment lines of phrases into words
8. if number of empty interval for each phrase match up, but  total number of empty interval does not match up with total number of words in lyrics, then there must be some extra empty spaces outside of these phrases, such as an empty interval where it supposed to say 'ac' as copied from the lyrics phrase tier. 
9.is able to process multiple files at one time
*************************************************


sound cluster script
*************************************************
1.clusters sound segments of the same tone type across all arias selected to be processed into a big sound file, each segments are separated by 0.2s of silence.
2.copies annotation information about each syllable into the sound cluster, including word, and banshi(tone is not necessary).


*************************************************
#######################################################################