CompMusic_MTG praat scritpts (2013) | praat script archive
This page is a collection of the Praat Scripts I developed at CompMusic Project,Music Technology Group,Universitat Pompeu Fabra, Barcelona, Spain, in 2013. CompMusic is an audio signal processing and Music Information Retrieval (MIR) research project aimed at exploring domain-specific knowledge and computational tools to better process world music.It is funded by the European research council(ERC), PI:Xavier Serra.
Tools Classification
The praat scripts are developed in the beginning of this project for the preprocessing and annotation of Beijing Opera dataset. In the first phase, we have developed a set of automatic annotation and preprocessing tools to experiment on a selection of 20 arias. These tools are combined with manual works to produce a core data set of 20 arias fully segmented and annotated. This mainly involved contributions from me and Rafael Caro Repetto. Mainly due to our research question to compare linguistic tone contours with music melodic contours, we develop tools of the two kinds,plus misc. tools:
- text processing and speech data processing tools
- melodic pitch contour processing tools
- others
File format
Most files are in source code format, unless otherwise noted(some are in .txt format for easy viewing in browser). Clicking them would download them instead of view them in the browser.
Most files are praat scripts. Java or other source code files will be noted.
Most files are praat scripts. Java or other source code files will be noted.
SCRIPT FILES
Text Annotation & Tone Cluster in PKO Lyrics
- [PKO annotation script(txt)][algorithm chart]
- [sound cluster script(txt)][algorithm chart]
- [documentations(variable & algorithm)]
- [documentations2 (functionality)]
- [Algorithm flowchart]
- sample lyrics.txt
Pitch-synth
Utility for Pitch Track Data Processing, Evaluation & Synthesis (batch processing, w/updates) in Praat
- how does this work?
- documentation
- log+updates
- algorithm
- pitch-synth.zip (contains all source code and system files needed for synthesis,except for the pitch track files and original sound files)
1.Pure tone synthesis from pitch tracks(vocal_pitch,w/updates)
- v1.1[MELODIA extracted pitch track of Jingju in pure tone synthesis.txt(simple version on short test file and small pitch track)][pitch track][synthesized pitch track+original track (wav file)]
- v1.2[MELODIA extracted pitch track pure tone synthesis.txt] (multiple files, read from MBID pitch track names, with duration normalization) | documentation | sample pitch track
- v1.3[synth12.praat] update 08/10/13
- v2.1[MELODIA extracted pitch track of Jingju in original extraction.txt]
- v2.2[MELODIA extracted pitch track in original extraction.txt](pair with 1.2,multiple files, read from MBID pitch track names, with duration normalization) | documentation | sample pitch track
- v2.3[synth-ori12.praat]update 08/24/13
- v3.1[shell-pitch.praat](praat shell script for synthesis & extraction & combining into stereo)
misc.
other scripts
4.Misc.(query yaml,filename-MBID processing,etc)
8.Others
- v4.1[query the trackInfo yaml file from praat script1.txt (search by the aria)]
- v4.2[query yaml2.txt (search by name of the opera and by the aria, returning more refined results)][sample yaml used]
- Query_yaml Documentation
- system files:MBID-WORK(full database);MBID-NO-WORK;MBID-PINYIN(20 ARIAS),mbidsel.txt,mbidsel_test.txt
- [convert stereo to mono.praat(large files batch processing)]
- [clear abnormal time jump.praat]
- [intensity-segmentation.praat(segment all wav files in a directory by intensity and store in a textgrid file, if that textgrid file does not already exist.)]
- n2p conversion (java tool written by me)
8.Others
- pitch_plot.py (by Ken Chen)