TimeView

TimeView is a cross-platform desktop application for viewing and editing Waveforms, Time-Value data, and Segmentation data. These data can easily be analyzed or manipulated using a library of built-in processors; for example, a linear filter can operate on a waveform, or an activity detector can create a segmentation from a waveform. Processors can be easily customized or created created from scratch.

This is a very early preview, and is not suitable for general usage yet.

Features

Cross-platform, verified to run on macOS, Linux, and Windows
Flexible arrangement of any number of panels, which contain any number of superimposed views (e.g. waveforms, spectrograms, feature trajectories, segmentations)
Views can easily be moved between panels
Views can be linked so that modifications in one panel are reflected in other panels
Customizable Rendering of views (e.g. frame_size for spectrogram)
On-the-fly Spectrogram rendering automatically adjusts frame-rate and FFT-size to calculate information for each available pixel without interpolation
Editable segmentation (insertion, deletion, modification of boundaries; modification of labels)
Basic processing plug-ins are provided (e.g. activity detection, F0-analysis)
Processing plug-ins are easily customizable or extendable using python
API allows accessing processing plugins for batch file processing or pre-configuring the GUI (examples are provided)
EDF-file-format support
A dataset-manager allows grouping of files into datasets, for quick access to often-used files
Command Line Interface support, for easy chaining with other tools

An introductory video is available

Installation

Pip

To install TimeView via pip one needs a python 3.6 environment setup in advance

pip install timeview

To execute run

timeview

Pipenv

TimeView supports installation via pipenv

pipenv --python=3.6 install timeview

To execute run

pipenv run timeview

Contributing

TimeView welcomes contributors. Currently the primary use case for TimeView is audio files of speech, however there is no reason we any other signal data cannot be used here. This is the authors first major project that is public facing and deployable via pip, so there are lots of opportunities to help us make TimeView better.

Some immediate goals of the project include

Expanding the test suite (right now it verifies the application starts and that's about it)
Incorporate mkl-fft library for faster spectrogram generation
Create documentation and provide a consistent API so other developers can integrate their own processors or renderers
Create new renderers, allowing for modification of pitch curves
Incorporate audio playback
Make better use of multithreading
Incorporate type hinting, such that we can use mypy
Make application launch through the GUI of all major operating systems

To contribute, follow the following steps

Fork the repo!
Create a feature branch (git checkout -b feature/fooBar)
Commit changes (git commit -am 'added feature fooBar!)
Push to the branch (git push origin feature/fooBar)
Create a Pull Request!

Testing

To setup a development environment for TimeView, set up an empty python environment

git clone https://github.com/{username}/timeview.git timeview

# installing dependencies to run and execute tests
pip install -e "timeview[test]"

After installing the test dependencies, you can run the tests.

python timeview/setup.py test

Support

Report issues to the GitHub issue tracker

Authors

Dr. Alex Kain - Professor at Oregon Health and Science University - Originally envisioned this project, had lots of existing code to make it happen
Ogi Moore - Student at Oregon Health and Science University - Hired at BioSpeech to take this project from an idea into something usable

Acknowledgments

BioSpeech for funding the initial work and allowing to open source
Dr. Alex Kain for giving a student trying to make a career transition to a software role an opportunity
Library maintainers of PyQtGraph, SciPy, NumPy, and other dependencies if it was not for their effort this project would never have existed

j9ac9k/timeview