Skip to main content

Prosody stylization

Project description

CoPaSul - Contour-based, parametric, and superpositional intonation stylization

Author

  • Uwe Reichel, Research Institute for Linguistics, Hungarian Academy of Sciences, Budapest

Purpose: Prosodic Feature extraction

  • Intonation
    • Global f0 register
    • Local f0 shapes
    • Prosodic boundaries
  • Energy
  • Rhythm
  • Voice quality

Dependencies

  • Python: >= 3.8
  • Packages: see requirements.txt
  • tested for Linux and Python 3.8, and 3.10

Installation

From PyPI

  • set up a virtual environment venv_copasul, activate it, and install copasul. For Linux this works e.g. as follows:
$ virtualenv --python="/usr/bin/python3" venv_copasul
$ source venv_copasul/bin/activate
(venv_copasul) $ pip install copasul

From GitHub

$ git clone git@github.com:reichelu/copasul.git
$ cd copasul/
$ virtualenv --python="/usr/bin/python3" venv_copasul
$ source venv_copasul/bin/activate
(venv_copasul) $ pip install -r requirements.txt

Usage

Required files

  • audio (wav), f0, pulse, and annotation files (Textgrid), see on GitHub: tests/minex/input
  • a config file (json), see on GitHub tests/minex/config/minex.json
    • see on Github docs/copasul_commented_config.json.txt and the article for further details

Extract f0 and pulse by means of Praat

# add this line if you use the code from GitHub
# sys.path.append(PROJECT_ROOT)

from copasul import praat_utils

# Praat pitch extractor
piex = praat_utils.PraatPitchExtractor(
    hopsize=0.01, minfreq=75, maxfreq=600)

# process mono files
piex.process_mono_files(
    input_dir=INPUT_AUDIO_DIR,
    output_dir=OUTPUT_PITCH_DIR,
    input_ext="wav",
    output_ext="f0"
)

# ... in order to process stereo files, use:
# piex.process_stereo_files()

# Praat pulse extractor
puex = praat_utils.PraatPulseExtractor(
    hopsize=0.01, minfreq=75, maxfreq=400)

# process mono files
puex.process_mono_files(
    input_dir=INPUT_AUDIO_DIR,
    output_dir=OUTPUT_PULSE_DIR,
    input_ext="wav",
    output_ext="pulse"
)

# ... in order to process stereo files, use:
# puex.process_stereo_files()

Call Copasul from terminal (if cloned from GitHub)

  • see on example script
  • python scripts/run_copasul.py [-c myConfigFile.json]
  • PROJECT_ROOT: directory where GitHub project has been cloned
(venv_copasul) $ cd PROJECT_ROOT/scripts/
(venv_copasul) $ python run_copasul.py -c ../tests/minex/config/minex.json
  • if you use ../tests/minex/config/minex.json as config file, CoPaSul
    • processes the files in ../tests/minex/input/, and
    • outputs feature tables to ../tests/minex/output/test.FEATURESET.{csv, R}
    • outputs the CoPaSul output dict to ../tests/minex/output/test.pickle
    • this output dict can be used for further (warm start) processing
  • if you use your own config file, make sure, that all directories in fsys:*:dir are either absolute paths or relative paths relative to your config file

Example integration into python code

import json
import pickle
import sys

# add this line if you use the code from GitHub
# sys.path.append(PROJECT_ROOT)

from copasul import copasul

# feature extractor
fex = copasul.Copasul()

# processing based on config file
config_file = MYCONFIGFILE.json
copa = fex.process(config=config_file)

# processing based on config dict
with open(config_file, 'r') as h:
    config_dict = json.load(h)
copa = fex.process(config=config_dict)

# warm start: continue processing
#   - implicit loading from config
config_dict["navigate"]["from_scratch"] = False
config_dict["navigate"]["overwrite_config"] = True
#     ... change further navigation values to your needs
copa_further_processed = fex.process(config=config_dict)

#   - explicit loading from file
copa_file = MYCOPASULOUTPUT.pickle
config_dict["navigate"]["overwrite_config"] = True
#     ... change further navigation values to your needs
with open(copa_file, "rb") as h:
   copa = pickle.load(h)
copa_further_processed = fex.process(config=config_dict, copa=copa)

Material

Reference

Reichel, U.D. (2017). CoPaSul Manual: Contour-based, parametric, and superpositional intonation stylization, arXiv:1612.04765.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

copasul-1.5.3.tar.gz (134.0 kB view details)

Uploaded Source

Built Distribution

copasul-1.5.3-py3-none-any.whl (131.0 kB view details)

Uploaded Python 3

File details

Details for the file copasul-1.5.3.tar.gz.

File metadata

  • Download URL: copasul-1.5.3.tar.gz
  • Upload date:
  • Size: 134.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.8.10

File hashes

Hashes for copasul-1.5.3.tar.gz
Algorithm Hash digest
SHA256 08e045beb632308115c16eaf891b525ab9e95f1446c5996c017c7c7cd2508a97
MD5 a94020ce508bafc94c9481317ac0df6d
BLAKE2b-256 62b2f4b518c99d609c76f2fa927ee530c6510ab5cdd13a8721b9320f149501ca

See more details on using hashes here.

File details

Details for the file copasul-1.5.3-py3-none-any.whl.

File metadata

  • Download URL: copasul-1.5.3-py3-none-any.whl
  • Upload date:
  • Size: 131.0 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.8.10

File hashes

Hashes for copasul-1.5.3-py3-none-any.whl
Algorithm Hash digest
SHA256 dd611751de061b2cf955dc154007836ace694929c1181a4729f504374fdb2d36
MD5 18a02cbbfda38395655b516246469e44
BLAKE2b-256 14a576c6ebb90f31bcbc45584647b53f9b326343ee50a379d0ce11855c5aa66f

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page